TW200304298A - Software, systems and methods for managing a distributed network - Google Patents

Abstract

A system and method for managing network bandwidth consumption. The system may include an agent module loadable on a networked computer and configured to aid in managing bandwidth consumption within a network, The agent module is configured to obtain an allocation of network bandwidth usable by the networked computer, and is further configured to sub-allocate such allocation among multiple bandwidth-consuming components associated with the networked computer. The system may further include multiple such agent modules loadable on plural networked computers, and a control module configured to interact with each of the agent modules to dynamically manage bandwidth usage by the networked computers.

Description

200304298 发明 Description of the invention: Cross-reference to related applications The present invention claims the priority of the United States Proposal No. 60/3 5 7, 7 3 1 filed on February 8th, 2002. The case is hereby incorporated by reference. TECHNICAL FIELD This invention relates generally to distributed network systems, and more particularly to software, systems, systems, and methods for managing resources of a distributed network. The previous technology A, and the private network system have shifted to a superior type, and have continued to be two in size, capacity and complexity = the long system has been accompanied by the corresponding increase in the need for information technology. Increase productivity at the end-of-business level, operation and consultant / user support In order to achieve increased interoperability in network systems with increased complexity, round-trip control protocols, Internet protocols, and other standardized communication protocols have been widely adopted. Placement: Although many of these agreements have been effective in achieving mutual operations, <,,, and 'for networks that use these agreements, the general configuration has been accompanied by the corresponding management solution. Development of wide resources = The two networks provide very few ways to manage network cases' and typically provide known A's best efforts to all network traffic. The best-effort system of the Internet Protocol network: This default behavior :, its; = in the case of excessive network congestion, the network nodes are only indiscriminate: due to the best-effort service, do not Set up a mechanism to avoid the introduction of the envelope of the 200304298 packet, and the network flow iS data, j *,, and 枓 are not classified to ensure the reliable transmission of more important data. In addition, users do not have resources about network conditions or incomplete performance. This lack of management often results in duplicate, unsuccessful network requests, user failures, and reduced productivity. The problems related to the management of network resources are more serious because of the rapid increase in demand for these resources. New applications for the Internet connection of Dao and Dao are developed at a rapid pace. These applications have dramatically changed performance conditions. Multimedia applications, for example, have unpredictable sensitivity to jitter, loss, and delay, while other applications have fairly appropriate frequency conditions. Two: The problem of the Γ step is that many frequency-intensive applications are used in entertainment bars. Or other lower priority work. The result of the lack of effective management and the competition for change is the increase in the use of network resources and the loss of productivity. When the network department cannot distinguish between unimportant jobs or requests for work, the network resources are usually implemented in a way that deals with business: pay. Bandwidth may be wasted or consumed by better performance. Consumers may place a high :: load on the network and experience unsatisfactory network performance. The same solution has been adopted, but there is only a limited degree of success in solving these network problems. For example, in order to reduce congestion, road controllers usually add more bandwidth to the congested soil. The solution of this solution is temporary, and the network makes the provided chain pure and fast and grows from the plug-in + π to form a new base. This is usually the case where the main cause of the congestion has not been resolved. Generally, it is expected that the existing resources of Huidi Government Management will be 200304298, rather than "over-provisioning," that is, only providing more resources to reduce sacrifice. The broad and conceptual classification of individual government solutions can be classified I want to make a graph of f in a decentralized network environment. I know that the concept is that his system knows more about the application or work performed on a network-connected device, and vice versa. However, it can take more effective use of network resources = ν. For example, if the network management software system becomes aware that a particular user is executing a lower priority application, the network can be blocked Or restrict users to access network resources. If the management software system becomes aware that, under a given nature, the network's individual group contains a high percentage of external customers, bandwidth preferences and priorities can be Modify to ensure that the customers have a positive experience with the network. Under this concept, increasing application and network awareness is a desired goal #, however, most of the vendors of applications ignore this Considering and tending to pay no attention to the internal structure of the network, but to pay attention to enhancing the function of the application. Quality of Service (Q0s) and policy-based management technology represent the network, applications, and use To bridge the gap between the two, to more efficiently manage the use of network resources. The quality of service is a request that allows applications that the network knows to request and receive according to performance specifications such as bandwidth, jitter, delay, and loss. A term for a technology of predictable service level. Known methods of service quality include not allowing certain forms of packets, slowing down transmission rates, establishing distinguishable service levels for certain forms of packets, with a priority value Marking packets and different enqueuing methods. In a decentralized environment with scarce resources, the service product 200304298 and technical titles must introduce inequality into the network by giving better treatment to certain network communications. System. 'Policy-based network management uses policies or rules to define how network resources are used. For example, a policy consists of a condition and an action. An example of a policy could be, if the source of the Internet Protocol source address of the data is a list, (conditions) contained in a disallowed address, then Block or not access or allow packets (actions). Based on the _ management technology-a material determines when and how the inequality introduced by the service mouth and mouth method should be applied. Based on policy The management solution typically requires that the communication data of the network be classified before and before. The classification process can be generated at different levels of data extraction, and can be used to pass through the network link according to the network device. Hierarchical communication protocols for road communications. There are two types of protocol hierarchies that dominate the field. The first is the open system, first interconnect (0SI) reference model, as shown in Figure 丄. The layers of this open system interconnection model are: application layer (Figure 7), presentation layer (Layer 6) 'meeting instrument layer (Layer 5), transport layer (Layer 4), and network layer (Layer 3 Layer), data link layer (layer 2), and physical layer (layer 2). The first major model is the basis for the transmission control protocol / Internet protocol. The layers are the application layer, the transport layer, the network layer, the data link layer, and the corpuscle layer. Drawing. The transmission control protocol / Internet protocol layer corresponds in function to the layers of the open system interconnection, but does not have a presentation layer and a conference layer. In both models, data is processed and changes its form as it passes through the layers sequentially. 200304298 Known methods for the quality of political table-based management solutions and services typically categorize data by monitoring data flows at and below the transport layer. For example, a common multi-parameter classifier is a 'well known' five-tuple, which contains the Internet Protocol source address, Internet protocol destination address, and Internet Protocol, transmission control protocol / user data protocol source port and transmission control protocol / destination port. These parameters are all obtained in the transport and network layers of the models. The vast majority of existing policy-based and service-quality solutions are implemented by monitoring and classifying network activities at the layers of these agreements. However, the higher the protocol level, the more accurate and specific the available data and classifiers. Because the traditional policy-based and service quality method of f is not to use a classifier at a level higher than the transport layer, it cannot use more detailed usable data at a higher level and adopt a policy-based approach. Technology or quality of service. Therefore, these traditional systems make the network more application-aware, and vice versa limited by its capabilities. In addition, the known system for managing network resources is an inefficient solution to the problem of bandwidth management. Bandwidth is usually consumed by lower-priority tasks. The cost is a commercially important application. In a system that does not provide bandwidth allocation in the order of priority ', ..., the bandwidth allocation is static and does not adjust dynamically in response to changing network conditions. SUMMARY OF THE INVENTION A system for network bandwidth consumption and a network-contained power system Therefore, the present invention provides a method for management, which can include the use of an agent module that can be used on the brain. The agent module is constructed to allocate the network bandwidth that can be used by the computer connected to the network, and is further constructed into a plurality of bandwidth consumption elements related to the computer connected to the network. This time the allocation is made. The system can further include a plurality of agent modules and a control module loaded on it to interact with each other to dynamically manage the bandwidth used by the computer connected to the network . Embodiments The present invention provides a system and method for managing a server and a server, such as those shown in a distributed network 100. This software is Μ β + I. 糸 糸 Pregnancy, the system and methods are to increase the satisfaction of the user, to minimize the failure of your field knife and customers / make use of the network, This is to ensure that network resources are used in one way and the other and the last way. For the industry or other purposes, the system and method are as follows: your agent and a control module: / a major software component, a ^ Μ% ^ ,,,,, and also a control point. This agent can be configured in the entire decentralized network. Two: two interactions to manage network resources. In order to manage each other's parent in 6, several agents can be configured to connect customers intelligently. Feed 1 The agent configured by Cheng Yuehui is to monitor, divide: and set to the main network. Take action on the network events of the device. H Regarding its related connected network and / or by one or more The proxy agent is typically a central agency that can interact to control and control the system. § Agents and control points operate and congestion status, select network events, and track network resource choices for the network. The optimization goal of the request, dynamic 200304298 &amp; bandwidth usage, and share the status of the network with customers, users and computer technicians. Figures' δ Decentralized Network 丄 0 can include a regional network Road 12 and a plurality of remote roads connected by a public network 16 such as the Internet. The side LAN and remote networks can be connected to the public network with a network infrastructure device such as a router. = Network i 2 is typically bundled Server 2 0 and customers such as Lei Nao 9 which are interconnected by the network key 2 4 can make the network 12 available to customers. In addition, 'Area = 3 any number and type of devices, including file server Customer electricity 电 月 _ ^ 'mail server, global web server, database, cry, ^^ distance access device 1 storage device, list machine and such as routing, 士 ## 局 土 ㈣, hub and relay Network foundation, port configuration device. Long-distance network] 4 network-connected devices. 'Similarly includes any number and kind of things consistently' Any form of real is shown in Figure 2 of the network The device can be connected to a handheld computer, a general purpose computer, a laptop, a computer, a wireless computing device, and a wireless computing device. It can be used for mobile phones, pagers, popular computers, and other special devices. The device is a general purpose computer, a device connected to the 4th floor, a computer cart antimony / 1 ′ ,,, and at least some of the components shown in Figure 3 Owner—_-a piece of digital data processing ^ No. The computer system 40 includes a processor 4 for 伹 伹 贝 # 2. Miscellaneous instruction set computing microprocessor, one thinks that one complex one, and one very long instruction Ann π s s 7 set leaves different microprocessors, processors or any processor, one implementation instruction, one micro-control 7 other processors ν The combined processor of the controller device. 12 200304298 Referring again to Figure 3, it should be noted that the processor is passed between the processing unit H and other components in the computer system; . Those skilled in the art will understand that the bus 2 is considered to be a single bus or a plurality of buses. A memory system is connected to the bus 44 and includes a random access memory element 4 7 (referred to as a main memory), which is stored by the processor 4 2 during the line Information or other intermediate information. The memory 4 6 also contains a read-only memory and "other static: storage media 4 8 'which is connected to a bus that stores information and instructions for the processor 4 2. A basic input / output system containing a basic input / output system that helps to transfer information between components such as the computer system during the initial period is stored in the read-only memory 48. A data storage device 50 is also connected to the bus 44 and stores information and instructions. The data storage device may be a hard disk drive, a floppy disk drive, a read-only optical drive device, a flash memory element, or any other storage device. In the computer system shown in the figure, a network interface 5 2 is also connected to the bus 4 4. The network interface is operated to connect the computer system to a network (the computer system 40 has not been dried out, and may also include a display device connected to the bus 4 4 to control 5 4 ° 1¾ display | A display device is connected to the computer system and is operative to connect the display device interface to the computer system. The display device controller 54 can, for example, be a monochrome display conversion card 'a color graphics conversion card' Or other display device controller. The display device (not shown) may be a television set 13 200304298 computer monitor, a flat display or other display device. The display device is received and processed by the display device controller 54. The information and materials / materials from the computer 4 2 and _ show the information and data to the user of the computer system 40. y ...

An input device 56 containing numbers and other keys is typically connected to the bus 4 4 'for communicating information and commands with the processor 4 2. Alternatively, the input device 5 6 is not directly connected to the bus bar 4 4, but is connected to the computer system through an infrared coding signal, and is transmitted from the input device to an infrared receiver (not shown) of the computer system. . ^ Lose

The input device may also be a remote control unit with a key for selecting characters or two selections on the display device. P

Different computing devices connected to the networks of Figure 2 typically communicate with each other through network links using communication software using different communication protocols. The communication software for each device connected to the network typically includes a number of protocol layers through which data is transferred sequentially as the data is exchanged between the devices via a fixed network link. The figure shows: a layered protocol for open systems interconnection and a layered model based on the transport control agreement / Internet protocol. These two models are the field of main control network and communication software. As can be seen in the figure, the open system interconnection model system has 7 layers, which includes an application layer (Figure 7), a presentation layer (6th, 4th (5th), and 4th) a transport layer (4th layer). ), Network layer (Layer 3), data link layer (Layer 2) and physical layer (Layer 2) 0 ° H Transmission Control Protocol / Internet Protocol Group model includes application layer, transport layer, network Road layer, data link layer, and physical layer. 14 200304298 Each layer plays a different role on the network communication path: Two: Yes: Some protocol layers are located in a data transmission. 'Between an application on the specific connected network device and the network link, and the: application. Caution light and mountain, Θ is taken close to the 05 Beco system by a Application running on a computer 4 When transferring through the network to the implementation of the% formula, an application on a computer &quot; Beijing is transferred through the protocol layers of the first computer Cross the network key to the second, and then move up through the: second layer. These two protocols are shown in the two In the model, one of the devices in the application should act as a filial piety. Your bebe and the connection network are executed at the "set 1: Γ connection 'and used to provide-a transport layer is responsible for providing one to access the The network. Between the routers. The road servo! And the area name is connection-oriented or no-connection: ::: Coordination. The transportation function can make the land system and the end-to-end point realized. The network The road layer code is not related to Xunting, but it is related to Beike under the Bubu-yang_Da and routing * materials to point f and = transport layer and network layer. The solution through this network Key Society = 4 U work month b 'and the lowest level is the problem. The key, ..., and the physical and electrical properties of the original bit transmission are described in any of the 0h above. Department: Should: Use the wide variety of nets attached to the Tong A Association of Exhibit 1 15 200304298 road soil, or can be applied to any other model with 1 layer. In addition, this method can be applied to Any form of t; fe Μ B -rm, in the form of, Kushiro topology, and can be applied to the body and wireless connection network. The present invention is intended for Management continuation ― 踗 之 二 欠, if the province is not in the same way as in Figure 2 of an enterprise network, the original software, and the soft knife of the ancient knife _ ”/. This can use two interactions. A human body is made of 7L pieces, one is crying AM m U and one control point is completed. The two intersecting software components can be adapted to one or the other, such as the juice shown in FIG. 3 # The computing device installed on the 4th computing device is connected to the external device following m Dingci, and the mouth is blocked. As shown in Figure 3, the computing device is set to open. For example, η ^ ^ 4%, multiple agents器 7 υ and or more control points 7 2 # 可 Lv μ 丄 ， 々 软件 模 袓 ## It is unreasonable to load the agent and control point into a basket such as customer Ye Zeng F w μ 2 And the server connected to the Internet, you can use the +, +, free-standing, and Kushiro 7 4. If it will be described in detail, the system of the agent and control system can be changed and constructed to implement ', bundle' blood vision and analysis of network events;轵 捗, 丄 L 古 谪 合 耔 1 1 contending documents and adopting Tian Yidong according to these events to provide valuable resources. Internet connection: ^ · Finance finally ensure that the network resources are connected with the main users, and most

For effective use.饤 I use the above software, system, and method to construct the sisters and daughters τ to use a third software element to construct, the second to do 6 pictures to ... Brother-software components are referred to below with reference to 1 Figures 3 to 1b are illustrated in detail. A typical 、, 忒, 、, σ structure device is a platform benefit application, which provides a…, user interface for the application history, in order to apply, the taxi is always used for these control points and these agents, The device system can be adapted to other managements. In addition, the structure is a management platform provided by other vendors. 'Surface connection, including 16 200304298 As shown in Figure 4, the / L control points 7 2 are typically combined with a plurality of agents 7 0, and the surname of the surname ... & Is within a field 7 &lt; of the specific path point. The control point coordinates and controls the distributed agent to move within its domain. In addition, the control points can monitor the status of the network poor source, and share the information of the management and support season with the agents. Jiu Wenba

These control points 72 and G π A Gori σσ 7 〇 can be flexibly arranged in different structures. By way of example, the "Siu Niu 1" direction = mother-an agent can combine a main control point and if necessary, ensure that one or more backup control points are maintained. Such a structure is not as shown in FIG. 4. Among them, the control point 7 2 in the dotted line is the main far 6 ± ny / pecker // and is connected to the server device 20. The δH control point is used as a backup link for the proxy state. In addition, the exemplary system of convergence can be constructed so that a system of points and agreements control and control the actions of a single, early, or multiple domains. Alternatively, a domain can be controlled and coordinated by a combination of multiple control points. In addition, an agent can be constructed with built-in control point functions, and thus may not require a correlation Control point entities. Typically, the 4b agent, θ, .. is a device that respects the resources of the network and the devices that combine with the control points and transmit the information. This ^ &amp; In order to respond to the monitored network status and the viewport of the agent I, the remote control can change the behavior of a specific agent to provide the desired network service. These control points and agents can be loaded on many types of devices, including-general purpose computers, servers, routers 'hub, palm brain' mobile computers and any computer with a processor on the entity. 17 200304298 Memory other devices connected to the network. Agents and control points can reside on separate devices, or on the same device at the same time.

Figure 5 shows an example of how the different components of the described software, system, and method can actually be interconnected with a network link. The components are all connected to the network link 90 by the layered communication protocol software 92. The components communicate with each other through the communication software and network keys. As those skilled in the art will understand, the network key 90 can be a physical or wireless link, or a series of links containing physical and wireless segments. More specifically, the system shown includes an agent 70 integrated with a client computing device 22, which includes an application 98. The other agent is a server computing device 20.6. These agents monitor the actions of the computing device to which they are connected and communicate with the control point 72. Structural equipment! Q 6 is to communicate with all other components and other management systems to achieve the operation of a few components and monitor the status of the network.

These system policies that define how network resources are used can be centrally defined and trimmed to the most efficient way to achieve the main goal. The defined policy is picked up by the control point, which then transmits the different components and parameters related to the policies to the agents in its domain, under a very basic layer-under the people '-a policy Contains rules about how network resources are used, and the rules contain conditions and actions that are taken when conditions are met. The agents and control points monitor the network and the devices connected to the network 'to decide when to apply different rules and whether the conditions that accompany the rules are met. Once the agent and / or control point 18 200304298 is required by the U.S.A. Operation System, its, ^ ^ policy of the system. Turning it around to implement it, 5 ’successful companies usually try their best to provide good customer service. The second key corporate goal is to be able to translate into many policies that define network resources. An example of such a policy is that the effectiveness of the application of… 'I has dropped to a threshold or restricted access to non-enterprise important, preventive, and important applications. Another example would be a technology that uses the quality of the service to provide an e-commerce application. Yet another fan… the service to the computer is from one client to two: =: whenever a connected network is added, the network bandwidth allocated to that connected network is increased sadly. In addition, the bandwidth used for different applications can be limited when customers frequently use network resources. The control point 72 will receive these policies and provide policy information to the agent 70. The agent 70 and the control point 72 communicate with each other and monitor the network to determine how many customers are accessing the network & which computers should be accessed and what applications those customers are accessing. 1: When the condition is measured, the agent and control point will interact with each other, = reallocate bandwidth 'to provide a specific service level, block or limit customer actions, and so on. Another policy-based example would be the best specification for defining a specific form of network resource or service level for network work ^ Some specific policies will guide the management entities to decide whether the specific The work is allowed, and if it is allowed, the management system of Yanhai interacts, and Cui Baoyuan's expectation level of Beiyuan is provided to complete the work. False 19 200304298 If the best resource is unavailable, the applicable policy is able to further indicate that the work of the request was blocked, and the user of the request was provided with a detailed description of why the request was rejected Information message. Or, ‘these policies can indicate that the user is offered different options, such as performing the requested work with the next best resources, or waiting to implement the work until a later time. For example, continuous media applications such as Internet Protocol phones have certain bandwidth conditions for best performance and are particularly sensitive to network jitter and delays. A policy can be written to indicate a desired service level 'including bandwidth conditions and threshold levels for jitter and delay for client computers attempting to implement Internet Protocol telephony applications. These policies will further guide the agents and control modules to attempt to provide the specified service level. Confidentiality checks can also be included to ensure that the particular user or client computer is allowed to execute the application. When the specified service level cannot be provided, the user of the request can be provided-a message indicating that the resources used for the request are unavailable. The user can also be offered different options, including performing at a sub-optimal service level, placing a traditional telephone call, waiting to perform the job until a later time, and so on. The software, systems, and methods of the present invention can be used to implement a wide range of system policies. These policy rules and conditions can be based on any number of parameters' including Internet Protocol ^ source address, Internet protocol destination address' source port 'destination', protocol, application identity, user identity 'device identity, User Route List (虮 20 200304298) 'Accessible Bandwidth' Application Round Temple 'Server Round Identity, Router Identity, -Time in the Day, Network Congestion, Network / Network The number of users, available domain bandwidth, and resource status are listed above. The actions taken when these policy conditions are met are sufficient to block network access, adjust service levels and / or the bandwidth allocation of devices used to connect to the network, and block specific user routing tables. Qingchang 'avoids network requests from overload or performance sources, ^ directs network requests to alternative resources and collects network statistics. Some of the parameters listed above can be considered "customer numbers, because they are normally calculated by an agent that monitors client devices on a single connected network. These include Internet Protocol source addresses, Internet Protocol destination address' source port, destination port, protocol, application: copies, user identity, device identity, user routing table. Other parameters, such as application profile, server profile, gateway identity, Router ^ Share | Time of day, network congestion, network load, number of network users, domain bandwidth and resource status obtained by 2 can be considered "system parameters", because it is about shared resources, total Network conditions or the calculation of data that requires multiple agent modules. However, there is no accurate distinction between customer parameters and system tea numbers. Some parameters, the time in the scholars, can be Considered one or both of a customer parameter or a system parameter. Policy-based network management, service quality, and the agents and control points Other functions are based on the availability of real-time funding on the network1. As will be discussed, some of the described examples and implementations are referred to 21 200304298 Improvements to known quality management solutions for policy-based services Because the enhanced capabilities of "Hai" are able to obtain detailed information about the network conditions and activities of the devices connected to the network. Because these agent module embodiments can be connected to their related agricultural homes in a specific way. The communication software, the policy parameters and conditions described by Huduoming are acceptable. In addition, as suggested in the above example, 'the management bandwidth for core applications and ensuring the availability of the system are: Increasingly important considerations. Certain embodiments described herein provide improved control of f-source consumption and dynamic allocation of bandwidth in response to changing network conditions. The flexibility described herein is based on The ability to configure policy-based, service-based quality management solutions with detailed information about road conditions has many important advantages. These advantages are Including: reducing failures on using the network, reducing calls for help from Internet technicians, 'increasing productivity, reducing business costs associated with managing and maintaining corporate networks, and increasing customer / user loyalty and satisfaction Degrees. Finally

These systems and methods ensure that network resources are used in a manner consistent with the primary goals and objectives. The implementation of the quality of policy-based services between the application layer and the transport layer has another advantage. This allows the provision of encryption and other confidential implementations using Virtual Private Networking (VPN) or Internet Protocol Privacy (IPSec) protocols. Please refer to Figs. 6 to 9, the embodiment of the illustrated agent module will be described in more detail. These agent modules can monitor the status, activities, and related information of customers, servers, widely used computing devices, or other computing devices 22 200304298, and send this information to one or more control points. Demonstrate the implementation of system policies; and provide information to the network user: a manager about network conditions. Figures 6 to 8 are conceptual diagrams of the notion of connecting to the Internet 'and show how the agent software can be combined with the layers used by the devices used for network communication: Internet connection Of the device. Stem version :: As can be seen in Figure 6, the agent 70 is set in the application program stack and is used to provide endpoint-to-endpoint data transmission such as the communication protocol and the transport protocol layer 124 of the communication protocol layer. between. Typically, the agent module described herein can use w /, or transmission control protocol / reverse layer communication software for the network = one =: The type points are displayed as "placed in 1FV for wide use." Therefore, the agent 70 is used for shore ", that is, in a data path, between a ::: and a ship agreement layer. However, those skilled in the art will be able to understand the τ ^ ^ The embodiment attached to the opener module can be based on a non-based agreement ^, a control protocol / Internet protocol, and a layering of results. The network device of the communication software is used together with a protocol layer that provides transportation functions, that is, provides end-to-end * data transmission.丨 The greed from the point of knowledge to the point of knowledge is because the position monitors the network circulation data at the location within the data path, which is enough or at the level of the ship. "Cannot be attached to the transport layer ... Obtained information # and take W. At the higher level, the stack allows the network to become more detailed. The stacking attached to the higher level is more accurate than when monitoring occurs at the transport and lower levels 23 200304298 may "know the application". The agent module can be set at a different point between the application program 22 and the transport layer worker 2. 4. Specifically, as shown in Figures 7 and 8, the agent 7 〇 Can be combined with a client computer so that it is adjacent to an Application Programming Interface (API) suitable for providing a standardized interface for application programs 1 2 2 to access a regional operating system and communication stack 9 2 In Figure 7, the agent 70 is adjacent to a winsock application interface worker 28 and is located between the application 丄 22 and the window socket interface. Figure 8 shows an alternative Structure where The processor 70 is attached to an envelope target again, such as the application interface i 2 8, but it is downstream of the envelope interface. With one of the above two structures, the agent is set in the application and The communication stack 9 2 is between the transport layers 丄 2 4 and is suitable for directly monitoring the data received or transmitted by the window envelope interface. As shown in FIG. 8, the agent 7 〇 can be Constructs a lower communication layer attached to the communication stack 92. This allows the agent to calculate the protocol layer generated under the transport layer i 2 4 by providing: a correction mechanism ^ data compression and encryption, and Accurately monitor the amount of data circulated on the network. For example, if data compression or encryption occurs within the transport layer 丄 24, monitoring at a point below the transport layer will generate information about the calculation. Inaccurate measurement of the network circulation data. Attachment to the lower layer by the agent 70 allows material compression, encryption, or qualitative or quantitative effects of other processing of the network circulation data to be generated from the lower protocol layer. Case, Lu 24 200304298 The circulation data is accurately measured.: Generation: An example of the device module is shown in Figure 9. As shown in the figure, it can include a redirector module 丄 3 〇, a / Fantong shell material control module 3 2, a manager module = domain name system feeder module 136 'a pop-up application module 13 8 fen-a message intermediate module 140 system service module 142 rows = The pop-up application 144. The redirection module 13 is to intercept the window cover application program interface group 1 implemented by the application such as the application shown on the network-connected devices of the client computers shown in Figures 2 and 3 3 〇. Then these chants are handed over to one or other of the other agent elements for processing. As given with reference to Figures 6 to 8 = modulo: 130 is positioned to allow the agent to execute on the application on the top and the stack of the communication according to the agent and control—passed by the envelope call can be called Refusal 'to change or be approved by the agent 70 ^ the data control module 132 is to implement the quality of the service and the system administration ... -Communication data control module 1 3 2 series = 1 = The method between the application on the agent device and the network link :: Shellfish flow, and the quality of service is implemented. The circulation of assets is controlled, and the value translation, including frequency treasure ... Internet service standards', can be used for specifications such as output volume, jitter, delay, and data loss. For the specific network service level, the circulation data control module can maintain a single row or multiple rows. When the data system 25 200304298 is transmitted from the shoulder customer to the network, or from the network to the customer, the redirection module i 3 〇 intercepts the data, and the circulation data controls: Group 1 3 2 It means that the individual units of data are placed in the appropriate queue. These control points can be constructed to provide periodic circulation data control commands. ', It can include the quality parameters and service specifications of the above services. In response, the circulation data control module 1 3 2 controls to the queues: = 彳 or the transmission from these to provide the specific and more specific information. The outbound circulation data rate can be Borrow a few such as the transmission queue! 9 Q Use complex and controlled. When one is used: the predecessor-based pass is used or the program is initiated by a proxy 7 ===, a priority level is based on the use of :::: 控: = For priority information for designation = Users roundabout, application round temple and network resource round? ⑼ = is the priority information provided to these agents. The pilot transmission line 1 3 2 a can be constructed to release the data and be transmitted to the network during the period. Using the parameters specified by a control point: out = data control command, the circulation data module "denotes: how much data can be listed by these transmissions in a specific period. In the case of cattle, 'if it is specified The average circulation is οοκbits / second, and the release period is 1 millisecond. The total amount of data that can be released during -M period is i 〇0: Γ The number of pieces of data to be transmitted. The priority is determined by each 26 200304298 "?" For example, suppose that there are only two columns with the data for the value of ★ μ σ Yau Cheng 1… first and 2, then if k is first pushed to be Q2's superior game $ 棼 # 妯 八 柘Two of them will be allowed to transmit 6 &quot; 6% released during the period of =: =. … Only allowed =: 3 3.3 3% released during the distribution period. If "1 = 2 priority of being shot by the second match, Q1 will be allowed to transmit the total 5 of the ^ released by the knife for transfer to the network link. If waiting for the poor material system If it is encapsulated to be more than one yuan, then the detailed accumulation is inexorable == between 4 times. When enough credit is accumulated, the waiting message is released for forwarding to the network. Similarly, in order to control the network The rate at which the circulation data is received, the circulation material &amp; manufacturing module 1 3 2 can be constructed to maintain a plurality of individual reception queues, such as the reception queue &quot; 2b. In addition to the above methods, different other methods It can be used to control the rate at which the network circulates data transmitted and received by the queues. In addition, the behavior of the transmission and reception symbols can be controlled in different ways to control jitter, delay, loss, and network Road link response time. The transmission and #receiving columns are also constructed to detect network conditions such as congestion and slow response applications or servers. For example, for each application, Transmitted packets or Other data units can be time stamped when transmitted from a transmission queue. When the corresponding packet is received for a specific application, the reception and transmission time can be compared to detect different targets Source's network congestion and / or slow 27 200304298 response time. This information can be reported to these control points and shared with other agents in the domain. The response time is compared by comparing the transmit and receive times. Other obtained performance information can also be used to edit and maintain statistical data about different network resources. Using this detection and reporting mechanism, a control point can be constructed to control the module i 3 by instructing the circulation of data. Close the low-priority conversations and one of the agents in each field reports severe network congestion to block additional conversations and reduce the network load. In combination, as described below, the application module 1 pops up 3 8 can provide a message to the messenger explaining why the dialogue was closed. &Amp; In addition to closing existing dialogues, the control point can be constructed Instruct the agents to block any further conversations. This action can also be done by a user message, in response to an attempt to send a new application to the network process. When the load on the network is reduced, the Control 2 will send a message to the agents that allow dialogue. ^ In addition to identifying congestion and slow response time, the flow data control module, '3 2 can be more generalized to help identify downward or performance failures. 2 ， Road. When—a link to the target system: the material control module 132 is notified and then issued—an executable 2 :: ;; the pop-up application module for solid root cause analysis ... The device 70 provides the control point with a message identifying the Z shell / prime and its status. When Dakki Baku, Kiki fails, the pop-up application module i ° W is constructed to provide a message to the operator ' , Marked as the unusable information tone # ώ1 Dagger 3 — An automatic routing option for the starting source. Enabling automatic link 28 200304298 caused the agent to retry the unusable resource periodically. If desired, this feature can be disabled to allow the control point to perform tasks that determine when the resource becomes available again. As will be described later, the system of the present invention can be constructed so that the control modules also perform the task of monitoring unusable resources to minimize unnecessary network circulation data. The different proxy components described below also monitor the network status and resource usage, α is used to edit statistics. An additional function of the circulation data control module ^ 3 2 is expected to help implement these functions by providing f-messages to other surrogates: access to resources ’, including access to resource performance and frequency. As suggested in the above discussion of the circulation data control module 132, the _application module 1 3 8 stores and is responsible for launching different small application models such as applications 1 4 4 to implement different Operation and strengthen the function of the narrative system. # 出 应用 系 Detecting and diagnosing the network status of resources such as downwards' Provide specific information to users and about the wrong state of Wu and the status of the network film , Report or operation support system as interface link, various: policy manager, service level manager and network and system management platform. Pop-up applications can be customized by customers to add features to existing products, in order to meet specific requirements, and integrate the software, systems, and methods with technologies provided by other vendors. The manager module 1 3 4 interacts with different other agent modules to 'maintain and provide network statistics, and provide __ agents for building the system 29 200304298 Throw π 4, 〇 Thousands of faces. Regarding the structure of the agent, the mouth of the philosophical fork, and the 134 series and the structure are prepared as an interface connection (shown in Figures 5 and 13 to 16), ^ ^ ^ 〇 construct different agent parameters. The Lu Li state branch group 1 3 4 also uses ^., Β ^ n to call the information obtained from the storage of local reports and statistical information that will be transmitted to the control points. The management mode:? According to other agent modules, the management group 1 3 4 series maintains local information, which is about the access server, area, / %%, first server, gateway, routing benefits, Switches, applications and other resources. Hiccup

^ ^ ^ ^ »I li H -r is transmitted on request based on the Bei Xun system: and can be used for network planning or to dynamically change the behavior of agents. In addition, the battalion of the battalion ~ li ... and 1 3 4 storage system policies and / or policy elements, and to inform I Katian to implement and implement these policies, provide policy information to different agent components . The manager module! 34. Also includes interface links supporting the described software and systems and standardized network management protocols and platforms.

— The area name system server module 136 provides a constructable address for the agent to determine the effectiveness of the service. Yanhai area name system server module 1 3 6 can contain a local cache of the area name system server information, and can be constructed to first use the local cache to determine an address request. If the request cannot be decided in this case, the request is made to a = control point, the control point determines the address with its own cache, assuming that the address is in the control point cache The user is allowed access / address. If the request cannot be determined by the control point cache, the control point connected by j will submit the request to a regional name system server for decision . If the address still cannot be determined at that point, 30 200304298, the control point sends a message to the agent, and then submits the request directly to its own area name system server. m dozen,

The area name system server module i 3 6 is also a monitoring bit and shares the content of these requests with the management module 134. The requests are = local ^ ride 'and are ultimately provided to the control, which dynamically maintains an updated list of the most popular regional name system servers. In addition, today's regional name system server module "6 is suitable for the control point 7 J ::: to: T if necessary 'to guide the address decision request and other network routes to the goal of replacement. The message intermediate module 140 generates and maintains links to interact with the agent-or control points. The different agent components use message intermediate modules to communicate with each other and communicate with a connected control point :: The message intermediate module 40 includes a message generator and message sending = private sequence 'to generate and send messages to the Some control points. The message generator private sequence contains the function of generating shells, which generates control point messages by receiving the message contents as parameters and encoding the contents in a standard network format. The message generator program also includes a member function to decode the remote messages received from the control point and return the contents in a format that cannot be used by different components of the agent. After being encoded by the generating program, the control point message is added to: a transmission queue 'and retrieved by the message sender function for transmission to the control point. The message retrieved from the queue is transmitted to the active control point of the agent. In addition, the material can be read and constructed using a 31 200304298 sequence numbering mechanism or other error detection and recovery methods to ensure the transmission of the message. Messages and communications from an agent to a control point are implemented by using a single-send addressing mechanism. Communication from the control point to one agent or multiple agents can be implemented by using a single send or a developed addressing mechanism. When constructed for multiple transmission operations, the control point and agent can be configured to transform into a single transmission ^ to allow communication with devices that do not support Internet Protocol multiple transmission. ~ ,, mouth 丨 Little Fox Wener, then the middle of the message-Ju 40 is to monitor the state of the switch and the connection when a connection fails: a backup control point. If the active connection is not active, the network circulation data is transmitted in a transparent manner. 5HAI System Service Module 1 4 2 #Description · τ k i is a different support function to it. First, the system service model outlines the outline of the temple server, the name of the area ... 42 series, the quasi peach link and other information ... The outline of the system feeder, control〗

^ / i £ m ^ ^ ^ Μ 糸, first service module 1 4 2 I

Very, traceability timer service for debugging. Used by the agent: The application program interface Bing Bing 4 2 can also be constructed to form a library of "Qu", so that the device combined with F w + # &quot; ° and 5 Hai operating system And the other components of the "King of Generation" 3 Hai device as an interface link. Now, please refer to Figures 〇 to 丄 2, and give a special description of ift mountain one work system 7 2 and its functions and call for ~ description. If it is not in the figure i 〇, the communication data module 1 β n, a 7 working point can contain a sub-domain name Fengsi / server profile module 1 62, an e

Into the name of the server profile module E υ 4 '— a gateway profile die island 32 1 6 6, a manager module 0, a # Ψ _, 1 6 8, a message intermediate module 7 _ 1 7 2 and a pop-up application 丄 7 4. The M1 of the control point circulation asset boat μ1, 1,6, and 16 is a technology that implements government-based products and services by coordinating these agents. As part of this function, the county's circulation data module 160 has obtained the distribution data from the 4 capital oblique --- agents according to the eighth degree. The bandwidth is allocated, and the calculated bandwidth is allocated to the agent, and the idle bandwidth is used to implement, and the bandwidth between the agents in its domain is dynamically allocated. See the example and $. The control point γ 2 can be constructed to recalculate the bandwidth allocation for 5 seconds. During each cycle, between re-entrants, these agents limit the amount of frequency used by the devices to which they are combined, and monitor "the amount of bandwidth used. At the end of the period, each unavailable parent 1U agent reported the bandwidth usage and "his allocation data to the control point used in the re-bandwidth." During the redistribution period, the circulation data control module 丄 60 is based on the priority data reported by the agents, and divides all the available bandwidth of subsequent cycles in the agents in the domain. . The result is a frequency $ cB, which is constructed specifically for each individual agent, which corresponds to the fair occupancy of the agent with the available bandwidth. The priorities and constructed bandwidth are a function of system policy, and can include application identity, user identity, device identity, source address, destination address, source port, destination port according to a wide range of parameters. , Agreement, user routing table, time of day, network load, number of network users, and other related parties; 胄 b enough parameters to be transmitted to the control point or the 33 network resources obtained by the control point . It can be counted by the client provided to the control point, the field and characteristics are compared with the Xi, the guide module i 3 〇…… fSun &amp; ^ greatly strengthened. Bandwidth allocation is achieved under the triggering conditions of the high-level crane in the stack, and it is clear that the political flute is only &amp; This can greatly enhance the flexibility and effectiveness of the software, systems and methods of the present invention. The priority reported by ❹agent n: ㈣may contain superior information related to the execution of the "It order-a plurality of applications on the device connected to the network.'In this case 'the relevant agent may be built Constitute @effective application priority ", which is a function for the individual application priority 2. For example, 'If device A is running two applications and device B is running a single application, the effective application priority of device A will be twice the effective application priority of χ B. It is assumed that the individual priorities of all three applications are the same. The reported priority data for a device running multiple applications can be further subdivided by weighting the reported priority according to the relative degree of activity of each application. Therefore, in the previous example, if one of the applications executed on device A is stationary or idle, the application's contribution to the effective priority of device A will not be counted, so that Device A and Device B have almost the same effective priority. To determine the effective application priority in order to use this weighting method ’The relative extent of an application ’s relative activity can be measured based on bandwidth usage, transmitted packets, or any other condition that indicates activity. In addition to priority data, each agent can be constructed to report the amount of bandwidth υβ used by its associated device during the previous cycle of 34 200304298, as described above. The data for the previous period can also be used for the allocated bandwidth AB of each device. The circulation data control module 16 can compare the constructed bandwidth CB, allocated bandwidth AB or used bandwidth UB, or any combination of these three parameters for each device to determine the following: The allocation of one cycle. In order to total the three parameters, UB is the number of devices connected to the network in the previous cycle, ab is the maximum number of devices allowed to be used, and CB is the number of devices that can be used during the next cycle. "Fair occupancy using bandwidth." Both the used bandwidth UB and the allocated bandwidth Αβ can be greater than, equal to, or less than the constructed bandwidth ... This transferable example occurs when there is Many network-connected devices use their constructed possession 帛 CB. In order to use the available bandwidth also, these unused quantities are allocated: devices with a frequency of 1 external frequency, the results are some The device is allocated to the excess share of the fair share CB AB. Although AB can be exceeded: CB, the used bandwidth UB generally cannot exceed the allocated bandwidth because the agent circulation data control module is implemented The allocation algorithm can be used for each generation of the St bandwidth CB, the allocated bandwidth AB, or the new 50 allocation that is used. However, the system has some general, Standing pass :! Then, for example, when the bandwidth is taken out by the device ... it is expected that the allocation that is least affected by the downward adjustment is reduced first. Therefore, weeping, s times, and day setting are first +%, and 4 control mode Group 1 6 Q can be constructed without any reduction> The relevant generation 5 ^ The reporting bandwidth of the benefit uses UB to allocate to customers or other devices under the allocated number 35 200304298 ΐ AB. It can be assumed Yes, if the allocation of these devices is reduced, these devices will not be affected. · In general, the circulation data control module 160 should not reduce any other allocations until all unused allocations or The distribution portion has been reduced. Then, the circulation information control module 丄 60 can be constructed to adjust continuously according to certain other conditions, or reduce the distribution of particularly high. The circulation information control module 1 6 0 also It can be constructed so that when the bandwidth becomes available, the most recent available bandwidth is provided according to the generalized preference. For example, the circulation data control module i 6 〇 can be constructed to provide first The remaining bandwidth is an agent with a low allocation amount and is requesting additional bandwidth. After these requests are satisfied, the remaining bandwidth can be allocated according to priority or other conditions. Section 1 A, 1 1 B, 1 1 c, 1 1 D shows examples of different methods that can be implemented by the circulation data control module 160 to dynamically allocate bandwidth. Figure 1 1 A shows the circulation data control module Group i 6 〇 is a procedure necessary to decide whether any adjustment of the bandwidth allocation AB is necessary. The bandwidth Aβ to be adjusted for certain agents is adjusted under at least the following circumstances. First, 'as shown in In steps S4 and S10, if the sum of all allocated bandwidths ABtotal exceeds the sum of the constructed bandwidths cBt0tal ..., some of the allocated bandwidths are modified. This situation can occur when, for some reason, in a previous cycle, a certain P-knife of the available total bandwidth is insufficient for use, or because it has been reserved for other purposes. In such an environment, it is important that during the next cycle, it reduces / some allocation AB to prevent the total allocation from exceeding the available total of 36 200304298. Secondly, if there are AB &lt; CB and any agent driving AB, then the allocation for these agents n is modified as shown in steps% and 51. The allocation to any such agent is typically increased. In this case, an agent has an allocation AB that is smaller than its constructed bandwidth cb, that is, its existing allocation is less than the fair share of the bandwidth that will be used in the next cycle rate. In addition 1 the reported use for the previous cycle UB is assigned to the implementation Aβ

The implementation is allocated at AB, and thus it can be ensured that if its :: increases, more bandwidth will be consumed by the relevant device. Furthermore, if there is any agent reporting bandwidth using UB which is less than its assigned AB, as determined at step 38, then for this-the assigned AB of an agent is reduced for the next cycle to Free up the bandwidth used. Steps S4 46 and S8 can be implemented in any suitable manner. In summary, if one or more of the following three conditions are =

(1) ABt〇tal> CBt〇ta (2) AB <CB and UB4AB for any agent, or (3) +, 0, AB for any agent, these three steps ensure some The bandwidth allocation is modified, and == is increased or decreased. If none of the three conditions are straight, then, i.e. increase, then the allocations AB from the previous cycle are not adjusted. At steps s ^ and 丄 υ, if the right requires ‘the circulation data control module 16’ is modified to allocate knife milk ΑB. After the necessary modifications have been implemented, the control point transmits this μ ‘new allocation of the agents for implementation during the next cycle. Main Figure 11 shows the redistribution of bandwidth to ensure the next period 37 200304298. The total allocation AB does not exceed the total available bandwidth. At step S18, the circulation data control module 16 has confirmed that ==. The sum of the allocated AB from one cycle is more than the available bandwidth of the next cycle, that is, ABtotal & gt CBtotal. In this case, certain allocations of the AB system must be reduced. As shown in steps s 2 0 and s 2 2, the circulation data control module 16 0 can be constructed to first reduce the report that the bandwidth usage level for a particular agent is lower than its allocated amount, that is, Assignment of UB &AB; These agents do not use 其 of their allocation and are therefore unaffected when the unused portion of the allocation is removed. At step S2 Q, the circulation information I is modeled, and 160 has no such agent. At step S22, the allocations for some or all of these agents are reduced. These reductions can be asymptotic, or the entire unused portion of the allocation can be removed at once. After any and all unused allocations have been removed, further reductions may be required to properly reduce the overall allocation. As in step V 2 S4, the further reduction is implemented for agents with an existing allocation AB &amp; Xin that the bandwidth CB of the construction is large, that is, AB & CB. Yes = In step S22, in step S22, the allocation is reduced due to unused frequencies: In step S24, the bandwidth is exceeded by the existing allocation that has occurred in the next period J over the calculated "fair" The occupancy rate is removed from the device. As shown in step S26, the reductions captured in steps S22 and S24 can be applied until the allocation of the total is J or equal to the total of the next cycle The available bandwidth is up to 38 200304298. The music charter shows a method for increasing the allocation of certain agents. As described with reference to Figure 1 A, where for any agent 'ABCCBa UB With AB, the allocation of such agents should be increased. The existence of this situation has been determined at step s 40. In order to provide additional bandwidth for these agents, the allocation to some other agents Typically, it needs to be reduced. Steps $ 2 0 and S22 'similar to Figure 丄 B and S22' unused bandwidth are first identified and removed (steps and S44). Again, this control point can be constructed Change unused points Partially removed rate. If the reported data is not against the library's bandwidth: the bandwidth of the circulation data control module 16 can be constructed to enjoy the CB, if not in step S46. In step S44 The bandwidth restored in S46 and S46 is then provided to the agent requesting additional bandwidth. The quantity method can be used to provide the restored bandwidth. For example, in order to give the report t the allocation AB and its The difference between the constructed parts = the agent with a large difference. Or, preferences can be based on the identity of the application, user identity, priority information, other customers, or any other suitable criteria. Fei Kung No. 1 1D picture shows a for re- The allocation method is not used. In step S 60, Gan and β,, b U, it has been determined that some allocation elements are used by the individual agent, that is, for the reason in step S62 These agent allocations AB are reduced. Due to the reduction and modification of the reference ° σ ^^ 2, the rate of the adjustment can be described by 11β 以 κ through 39 200304298 structural changes to the control points. ... for example, It may be expected that only a part of the unused bandwidth is removed during the redistribution cycle. Or, the unused bandwidth during the redistribution cycle ^ 'month' is entirely removed and reassigned. At 11th! In step S 64, the quantity to be restored is provided as necessary. The restored bandwidth can be used, please eliminate as shown; the overall allocation of ABtotal and the available frequency in Figure 11B Forgiveness: the difference 'or as shown in item 丄 丄 c, increase the allocation of agents that are requesting additional frequency and have a relatively low allocation.

Sufficient removed bandwidth ’is allocated to the agent requesting additional bandwidth, that is, UR £ =: ^ ^ w • Αβ’ is even used for such an agent

The distribution AB is quite high ', for example, when AB &gt; CB. Because of the methods described in sections uB and 1 1 C B1, the recovered bandwidth can be allocated using many methods and according to any suitable conditions. As indicated above, the circulation data control module 丄 60 can be constructed to constitute the rate at which the above-mentioned distribution adjustment is implemented. For example, suppose a particular device is assigned "κ bits, seconds (AB) and reported in the previous

Cycle 6 2 K bits / 5H, # M W. ^ / Shift use case. Then the circulation data control module 160 can not determine how much additional bandwidth the dream department ^ (客 # # 疋 ”茨 忒 置 will use. Therefore, 'if the distribution system is a large office AΓ7, A 丨 i.丁 A〒 田; If you add t, for example, twice, a substantial portion of the increase will become unused. You can use this system for less / less time. However, because the device uses an approximately equal The ball-necked prince A n steals the amount of AB allocated by Kou Hai, and it can be assumed that if the distribution system increases, the nightmare magic edge setting will be used more. Therefore, it is usually preferred Provide small increments to do plus., Taxis & soil said Riga. The amount of some incremental adjustments and the rate of implementation can be broadcasted in the form of η μ ▲, also known as The construction is as described with reference to Fig. 40 200304298 in Fig. 16. If the device consumes this additional amount, then if additional bandwidth is available, a continuous increase can be provided. In addition, the bandwidth allocation and Calculate the transmission and reception rates that are available for the connected device Separately implemented. In other words, the methods referring to Figure 丄 1 A _ 1 1 D can be used to calculate a transmission allocation and an individual reception allocation for a specific device. Alternatively, the calculations can be combined, To produce an overall bandwidth allocation.

The temple server corridor module 1 62, the area name is the server profile module 1 6 4, the gateway profile module 丄 6 6 and the manager module: 1 6 8 are: these agents Interact with each other to monitor the status of network resources more clearly. Figure 1 of Figure 2 provides an exemplary example of how the control points and agents can be constructed to look at the state of the shell source on the network. The monitored resource can be a ^ 1 J service state, an area name car system server, a router, a "gateway state, a switch, an application, etc. In step S 100, _ find In U υ, a poor state change system has been exemplified by a server system ^

The communication information has been greatly increased. One win-from Tian Yi's machine, a special application system cannot be used, or a special gateway device. . The effect has been reduced to the predetermined threshold specified in a policy. ',,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, (,,,,,,,,,, (,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,)), (&) On the network, related to a specific resource &lt; If q ft% is too far off, there is an execution produced by the bomb {j 接 接, and 1 3 8 and ... The execution method (Figure 9) is to analyze the situation and report the source of the shell. The status of this identity is so sad that it is connected to the agent; 41 200304298 The source of the request for delivery of the application ... The system policy can include the use of different network resources: Haijian :: t :: ticket 'and can Further &quot; &quot; specifies that when the resource efficiency system is shifted: :: this 弹出 a certain amount of pop-up application analysis system is implemented. In addition, the network control point can be similarly configured to issue a pop-up application to analyze the information to: =; :: =: 'The control point reports the asset: ... Baiyuan's further customer request, as shown in step s Step S 1; [〇. For the target resource or ..., give these alternatives with caution, / S # Phase 7 will be based on whether the resources of one generation: column: make two. "The system dynamically stores updated alternatives for the agent to cry. The 4th generation resource system can be used. Then, the instructions: ° σδδ may include transparent redirection of the command line ::::: The source command, as shown in the step of the server, the second point is to know that a server maps another server that has already left, and the client request that leads to the left server can only be made available to one generation of resources.使用者 The user said that under the attempt of receiving, Γ provides a processor to pop up the library 22: step S110. The signaling function is replaced by the controller. In addition, the 'pop-up application function can be used by other vendors Provided pipe = status: =: control point and * Lu Buqian $ 0 In addition, information about the status of the resource or 42 200304298 network status can be provided by e-mail or paged to an Internet technician ... Referring to Figure 2, the control point can be constructed as a task responsible for tracking the state of the resource to determine when it becomes available again: as shown in step S112. A slow polling technique Is used: take unneeded streams on 4 networks The information is not used. In the meantime, the agents provide error messages based on the resources provided by the control point and "redirect the request to these resources. If not in step S1, 1 6. Once the control point confirms that the resource can be used again, the control point shares the information and the agents, and disables the instructions provided in steps S108 and si10, such as Shown in step s 118. The method of tracking and monitoring the status of resources has important advantages. First of all, it is to reduce the unnecessary and frustrated access to unusable resources, and repeatedly attempt to implement—a health work, -A user's request was re-quoted to enable the request to be successfully served, or the user was provided with information about why the attempt was unsuccessful. Because obtaining this information, users are less likely to generate There is a waste of network circulation data with repeated access attempts within a short period of time. In addition, the network circulation data is also reduced by having only one entrance through which communication is a control point, and the responsibility for monitoring cannot be compromised. Tasks of the resources used. In addition to helping these resource monitoring functions, the server profile module 1 &amp; is a tragic update of the servers received by the agent within its domain. List. The server statistics can be taken out by using 2003 200304298 using the structure equipment, or with different 1 cents, and the existing management platform. This server statistics can be used for network planning, or can Used to become a different system policy 'for the dynamic implementation of this agent and control point. For example, -Haiji &amp; manufacturing points and agents can be constructed to be loaded feeders or other For the source, turn to the circulation information. In addition to the second name of the area, the server of the local name system module 1 6 4 helps to monitor and track the resource, and also implements certain specific functions. This area name system, feeder, and corridor module 1 is to maintain a local area name system for efficient ^ α sheep's home address determination. As summarized by referring to the agent area name system module worker 36, the agents and control points are used interactively to determine the address request, and can be referred to by the agents and or control Click on the local area name system information to determine the address. The profile module 162, 哕 ρ ^ service time wheel 4 £ Domain name, system server profile module i 6 4 also maintains statistical data for use in network compliance dynamic system policies. Structure ^ Up = Function: The manager module 168 is a maintenance control point. Xi Servo-II is equipped with the poor information to the agents in the domain. Similar to = :, area name system, and closed circuit module, the manager module worker also helps to collect and maintain the uniformity of network resources == group ... is taken out from the centralized policy storage; 2: Save the policy information for use by these control points and agents in the poor yoke system policy. Niu 72 2 also includes a synchronization interface (not shown) for synchronizing information between multiple controls in the same domain. 44 200304298 Message middle module 1 7 〇 # 〇μ χ n Λ υ implements different functions, so that the control point communicates with these agents. Since σ worker

The cheek is similar to the agent's message intermediate module! 4 0, the message intermediate module 丄 7 Q Wu Hesan, female ^ system package 3 Λ interest generation state and message sending private "order, to generate and translate"> Χ7 6 玍 and get alum to these control points. The news self-program contains member functions, is devoted to ecology, and generates control points by receiving flood information content as a bundle number and editing it in a standard network format. message. Continued The coded message is transmitted ° Xuan Wang η Bujie and other network entities code the function of the received message ^ A 1 body is used for solution. Xi is also included in the message generator program, "Hai IS Yunyi private sequence is transmitted from ..y. R L and ensures a reliable transmission through a retry mechanism and 铒 祆 detection and recovery methods. 〃 现 Refer to Figures 13 to 1 β ^^^ 16 Figures, these agents and control points are available

To use the structure device 106 and # 捿 β J β μ / * 结构. Typically, the structural device 106 is a virtual field 4 of the Pingtang Yiguan Pass, and the m application 'provides a graphical user interface for centrally managed applications. This agent was crying for $. In order to construct the peptide control-Deliyi knot agent Cry 7 and the agent 'the structure device is the agent module 70 and the manager module device and control point γ 2 the f ~ structure to go, Zhan sister ^ It is used for interface connection of Bin group 1 6 8. Or-the configuration device 106 can be connected to the /, and then, the control point is connected to the interface _ = ,, and 168 as the interface. The manager module 0 of the agent 0 0 Figure 13 shows a main structure screen 8 of the structure device 10. As shown in the figure, the main warrior _ τ A structure screen 1 8 8 can be used in砚 Look at different managed object texts for agents and other network entities and resources. I use daggers. For example, in the main structure 45 200304298 labor screen 1 8 8 the screen frame on the left side 19 0 is a scalable representation of the control points that can be used to render the network.

When a specific control point is selected in the main structure Fluorescent 188: different settings for the control points can be constructed. For example, the name of the crafting point can be edited, agents and other entities can be added to the control point's domain, and the control point can be designated as a second agent or a second agent group. link. In addition, the system can be specified in the control point's domain for the user's total available bandwidth for transmission and reception, as shown in Figure 14. This bandwidth specification will affect the constructed bandwidth ⑶ and the allocated bandwidth M which are deleted in the method of controlling point circulation data module i 6 〇 shown in FIG. 11. The structure device 106 also provides structures related to different settings of users, applications and resources related to a specific control point. For example, users can be grouped for common collection processing. Forbidden users can be specified for specific users or user groups. . Priority can also be assigned to users or

As mentioned above, the priority data plays a role in determining the bandwidth allocation for these agents and their related devices. In addition, the best and smallest levels of performance can be established for use in applications or other tasks using network poor sources. Referring to the Internet phone example above again, continued this broadcast: ▲ # 古 ㈣U 冓 The minimum threshold that a device can be used to specify a network-connected device that is used to run the Internet Protocol phone application Efficiency level. The performance level can be specified based on performance parameters such as bandwidth, yield, jitter, delay, and loss of service quality. The agent module associated with the company 46 200304298, and then will monitor the Internet Protocol Thunder; * ra α · ^ Α application related to the road circulation data to ensure that the performance is the most efficient ], Above 5 °° limit. If the minimum level is not met, the controls and agents interact with the allocatable resources and provide the specified minimum service level. Similarly, an optimal service level can be specified for different, Kushiro applications and jobs. More generally, the structural equipment 106 can be constructed to manage system policies by providing for providing the following functions: Maintenance and storage system policies are managed on a decentralized network from other locations. The removal of system policies, such as a dedicated policy servo_, Ming &gt; According to Figure 16 it will discuss different other control points

Structure of agent parameters. As shown in the figure, structural equipment workers 06 can be used during the period when construction resource allocation is implemented. For example, the default period for recalculating bandwidth allocation is 5◦Q◦msec, which is 5 seconds. Outside 'As mentioned above, the rate of resource allocation can be specified to prevent over-compensation and unnecessary adjustments to the allocation, as well as inefficient new structures. To be clear, the percentage of bandwidth that was removed and reallocated from a customer facility and over-allocated can be specified by the structural equipment, as shown in Figure 16. In addition, the agent provides feedback to the control of the network status or the rate of activity of its related devices. The rate can be built: The structure described in W: The method is implemented. The first layer can contain-or more two =: 渚 如 7 2. Because the control points are the control and coordination agents 47 200304298 above, the control points can be referred to as "upstream," or "elements of the ancient," which are relative to the agent models they control. The following: the agents that form the second layer of the system can be lower than M ,, 4 ,, i, above or below, the elements that are relative to the eight that control As far as control points are concerned, the solution is that the systems and methods described here are quite capable and proportional, and can be applied to networks of widely varying sizes. In some settings, the ratio Transformation is through the implementation of the expansion hierarchy to three or

Even more reached. In fact, a very large structured model can be extended. The two-layer example described above is built on the upstream and / or downstream side. Such a model can be used to deliver consistent application performance in a very large and complex "broadway" structure. An upstream component can Control or multiple downstream elements. An upstream element can also be constructed as some downstream element that binds the upstream body. Similarly, a downstream element can be constructed as an upstream element that controls its downstream entities. Figure 17 Shows an example of a multi-layer implementation. In this example, 'the layer 1 component is upstream relative to the 3 components in layer 2. Because the component of layer 2 is below and is connected by the component of layer 1 'Controlled' so the element of layer 2 is the downstream element relative to the element of layer 1. As shown in the figure, the elements in layers 2, 3 and 4 can be constructed to serve as both upstream and downstream elements. Here The agent modules and control points described may be implemented at any level within a multi-layered environment such as shown in Figure 17. For example, 48, the elements of layer 4 may include The control can be constructed in the τ control modules shown in the figure, "2, module γ." The agent executed on the layer 5 element of the square is based on the layer structure, the first-to-many structure. Implementation of Move 2 diagram: This structure is typically tied to a heart sentence that can be controlled immediately below it-an upstream element migratory element can be controlled at the next two = downstream element 'and "" Fixed downstream element blood: multiple elements in the downstream layer. However, a. ^ 〇〇 _ ^ 'is reported to a single upstream element

W control by morning swim element. Library # Μ θ.... It should be understood that many different ^ structures are possible. The structure r T eve is small. There are any number of layers &amp; elements or envy of these layers A group of components within. Among the multi-sounding Sheshan M Xixi brothers such as those mentioned above, it is usually expected to provide a mechanism for the definition of policies and allocation of bandwidth and the management of other sources in China and China. Figure 18 shows an example of a network connection using this mechanism. In this example, ^ \ ^ The same knife office is connected to f-a centralized data center 2 Q 〇 through the network 002. The system can be constructed

Cheng Zhineng—a manager means a wide range of corporate policies in the central servo state of a corporate policy feeder. In such an environment, the four control points described herein can be implemented in combination with a controlled location policy server 206, as shown in the exemplary branch office 2 at the bottom of Figure 18. 0 8. Typically, a given location policy server 2006 is retrieved by its corporate policy server 2004 for its branch office policies. It may contain a control module 7 2 of the location policy server 2 06 2003 49298 and then assign the relevant policy to the relevant agent module 7 in its domain. Then the 'out of &amp; policy is assigned to different agents controlled by the control point module. The policy definition and distribution mechanism can be easily changed and scaled to manage a widely changing corporate structure. It is particularly important that this allocation policy can be used to facilitate the bandwidth management techniques described herein. In addition to the bandwidth management features previously described, bandwidth management can be affected using layered implementations such as those described above. Figure 9 shows an exemplary system 300 for managing bandwidth on a network link 300. As shown in the figure, the system 300 can include different components called a bandwidth manager, which can be described in the control modules (for example, the control point 7 2) and the agent module ( For example, implemented within agent module 7 0). For example, the system shown includes a controlled position bandwidth manager 3 04. The controlled position bandwidth manager 3 04 can be controlled as previously described, point 7 2 or in a controlled The controlled location policy server 206 is implemented as shown in FIG. In any case, the controlled location bandwidth manager 3 04 is typically implemented as a software program running on a computing device that is blocked such as being connected to a network link 3 02 A server (not shown). The controlled location bandwidth manager 3 04 is typically communicated with agent software (eg, 'Agent module 7 0') running on a computing device connected to a network link 3 Q 2,卩 Manage the bandwidth used by these devices. For example, as shown in the figure, the plurality of unique devices 2 and 2 which are also referred to as agent computers can be interconnected through a network link 3 ◦ 2. Each agent ② The computer can be executed on:-an agent bandwidth manager 3 〇 0 50 200304298 on the frequency band or multiple application bandwidth manager 308, and-or multiple envelopes, bandwidth manager 3 1 〇 上. The frequency managers that are loaded on the agent computer are typically children of the agent module 70 as described above. At a given connection to the network link.

Inside, it typically has an agent bandwidth manager 3 06, a three-volume application bandwidth manager 3 0 8 ′, and a variable number: seal = official manager 3U, which is based on the used Depending on the application and envelope: For example, 'when the bandwidth managers are attached to an agent module 70, the agent 11 module is typically changed to execute for each one; application 32 on the electricity monthly 22 ◦ Issue-an application bandwidth manager

Issued by each envelope opened by each application 3 2: The envelope bandwidth manager 310. As indicated in the figure, a given device can be executing multiple applications 3 2 Q, and a given ⑻ ^ 2 〇 can have multiple open envelopes 3 2 2. Computing device 2: It is through the transmitting and receiving control part 3 2 4 and 3 2 6 * Transmitting and receiving: the data on the network link 3 02. The transmission and reception control parts ^ 2 ^ and 3 2 6 may respectively include transmission and reception symbols 3 2 8 and 3 3 ('which will be described in detail below. In the example shown, the controlled position Bandwidth control 4 branches account for ο u "The overall bandwidth on the network link 3 〇2, an agent frequency ^ ^ 3 3 0 6 is managed by its related agent computer 22 2 Shoulder width, a given application bandwidth manager 3 08 manages the bandwidth used by its related applications 3 2 0 and an envelope bandwidth manager 3 51 200304298 1 0 system manages the envelopes related by it 3 2 2 The bandwidth used. As shown by the arrows interconnecting the different bandwidth managers, the managers can be configured with a hierarchical control structure, i, a y Ν J 'y (solid

An upstream element interacts with and / or controls one or more downstream elements below-or multiple downstream elements below. In the example shown ,: The controlled position bandwidth manager 3 0 4 is managed and used on the network link 3 02 by interacting with the next generation = the bandwidth bandwidth manager 3 0 6 Bandwidth to manage the bandwidth used by the particular downstream computing device 2 2 under the controlled location bandwidth manager ^ 0 4, application 3 2 envelope 3 2 2. It should be understood that the controlled location bandwidth controller 300 can manage any form of interconnection used between computers. For example, in Figure 2, a given long-distance network; 4 can include a server computer that executes a server computer that is interconnected with a number of client computers through a LAN section. Controlled position bandwidth manager 3 04, where each client computer is loaded with an agent module 70. In this case, the control computer can control the portion of the LAN used by the customer's bandwidth (for example, an Ethernet-based LAN) and through the router 18 to the Yagong network 1 of 6 links. Referring again to FIG. 19, a given agent bandwidth manager 3 06 manages the bandwidth used by its associated computing device 22 by interacting with the application bandwidth manager 3 08 below. To manage the bandwidth used by the particular downstream application 3 2 0 and the envelope 3 2 2 located below the agent bandwidth manager 3 06. Similarly, a given application bandwidth 52 200304298 manager 308 manages the bandwidth used by the application bandwidth manager 2 by its role. Reports such as bandwidth consumption are good for control. The bandwidth used by the application 3 2 0 that interacts with the lower envelope bandwidth manager 3 1 0 to manage the specific downstream envelope 3 2 below the physical 3 08. As explained below, the downstream components are typically With some data to the upstream components located above, there is a link in the display, φ, &amp; + »The medium bandwidth can be adopted by existing in a specific

One of the bands in the hierarchy is divided into A, A and B, and the allocation is artificially managed for allocation between downstream components. In the case where there are a plurality of downstream components, the sub-surface allocation typically involves Jiang Ba-, which is divided into parts for each of these downstream components with a knife. The position bandwidth manager 3 4 controlled by the article 5 can present an allocation corresponding to the available bandwidth on the network key 3 002: the available bandwidth can be specified according to the designation The system policy of the parameters of the network link 302 is constructed or determined or constructed by other methods. The available bandwidth on the link 302 can be allocated by the controlled location bandwidth manager 304 times to become one or more agents. Depending on the number of agents Yidianyue a 2 2 below the location bandwidth manager 3 04. Each agent allocation represents the amount of bandwidth allocated to individual agent computers. For example, if 0 0 agent computers 2 2 are controlled by the controlled position bandwidth manager 3 0 4, then the position bandwidth manager 3 0 4 controlled by § 20 is typically assigned twice. The available link bandwidth becomes 100 individualized allocations for the individual agent computers located below. The individualized allocation will be provided to the individual agent bandwidth manager 3 06 at each agent computer 22 for implementation and further sub-allocation.

Similarly, at each agent computer, the combined agent frequently sees that the agent 3 06 can assign its agent to become an individual 3 2 0 of each application running on the agent computer. Apply assignments. A given application allocation represents the number of bandwidth allocated to the corresponding application. 1 These application allocations are typically provided to the individual application bandwidth management unit 308 for implementation and further sub-allocations. Finally, at each application 320, the combined application bandwidth manager 308 series can be stabbed with its application allocation to become an individual envelope allocation for the envelopes 3 2 2 that are open to the application. The individual envelope allocation is typically provided to the individual envelope bandwidth manager 3 1 0.

Typically, at least some of the bandwidth managers are constructed to pass through to full, to facilitate the different sub-distributions described above. In fact, data can be provided upstream from each of the envelope bandwidth managers 3! 0 to the library bandwidth manager 3 0 8 above it, and by the application bandwidth manager 3 &quot; each A group is provided to the agent bandwidth manager 3 06 above it, and each group of the agent bandwidth manager 3 06 is provided to the controlled position bandwidth management above the basin器 3Q4. This information provided upstream can be used to calculate sub-distributions, which are then sent back downstream for implementation, and / or further sub-grading ^ man-to-know allocation. Generally, the data that is expected to be transmitted upstream is about the activity of the envelope to efficiently adjust the allocation of m to take out the bandwidth from where it is not needed and provide it to where it is needed more. In fact, the interaction between different bandwidth managers is typically implemented in 2003 2003298 to efficiently distribute network bandwidth among the different bandwidth consuming components. As mentioned above, the consumable component may include an agent computer 2 2 ′, an application 3 2 0 executed on the computers, and opening a cover 3 2 2 for the applications. One of the systems described here that can efficiently allocate bandwidth is the use of mobile bandwidth allocation to high priority use 'and the shift of bandwidth allocation away from relatively low priority use. Another allocation criterion can be adopted which involves providing further allocation based on the consumption of past allocations. For example, if the past allocation was only partially consumed, or there is a tendency to reduce usage, the allocation to a specific bandwidth manager (for example, from an agent bandwidth manager 306 to an application bandwidth The application allocation provided by the official server 308 can be reduced. The above-mentioned parent interactions need not be generated in any particular order, and can be generated periodically or aperiodically, and / or have different rates for different layers. For example, the controlled location bandwidth manager 3 0 4 was handed over to the computer 2 2 during the fixed period, or only when a changed condition in the network was detected. Agent assignment. Application allocation and envelope allocation can be propagated to downstream periodically, but under the same rate. In any case, it is often advantageous to update these allocations and sub-allocations repeatedly and dynamically, which will be described in more detail below. Now, bandwidth management for a representative envelope will be referred to in Section 2 In the description, the figure 20 shows an agent computing device 2 2 connected to another computer (not shown) via a network link 3 02. The computing device 2 2 executes an application program 3 2 0, the application program 55 200304298: = H = road link 3 〇2 is stacked via a layered protocol (: ,, and '' as described with reference to the example above) Group 70 can be set in the application agent. Benefit :: Interval: r "The flow control can be achieved = Lv Zhi, such as referring to the 6th, the 7th and the month is definitely set in the protocol stack management group 7 0 typical one The envelope object is used as an interface and is connected to or in the application 32: and: Second: the network circulation data is intercepted, which is hooked at one point == Γ92 in the transportation layer ... between 7 and 4 not less than 2 The agent module 7 of FIG. 0 may include 舲, +, and + gan with reference to FIG. 9 μ 'a re-guider 1 3 〇f ＣＯＭ, special processor module interception or Fishing two pictures) can be used to allow the generation of harvesters to hook the envelope data stream of application 3 2 0 and the net. For the representative envelope displayed between JUZ, the network link 30? Ximen ^ ^ 0 2 4 and receiving control section? W | knife 3 receiving control section 3 2 6 ^ 6 'The transmission control section 3 2 4 and connecting., 〇 It can include the transmission queue 3 2 8 and the makeup queue 3 3 0. The data flowing through Xiha,, ^ 0, and receiving 〇〇R ^ μ appropriate control section 3 2 4 and receiving control section 3 2 6 can New% .. knife _ broken monitoring and through the envelope bandwidth manager 3 1 0 of the agent module 70 and 0 u, fen s 5,. 1 text control. As mentioned above, the envelope bandwidth Guanlikou d 1 〇 can Panhui A Ning Qili crying _ /, the upstream frequency implemented in the agent module 70 (A is not in Figure 20) interaction. To be clear, envelope allocation can be provided to Envelope Manager 3 1 0 Division 2 'Member Width Manager 3 1 0, the envelope frequency controls the transmission control part 3 2 4 and the reception control 56 200304298 part 3 2 6' to ensure that the envelope bandwidth manager Activities) to further facilitate the further allocation of these allocations are implemented. In addition, 'as shown in the figure = 1 〇 can provide feedback (such as the envelope bandwidth controller. This feedback can be used envelope allocation can be detailed This interval is provided at a fixed interval of 31. Under the concept of "individual individual", the allocation is typically related to-the transmission and reception allocation of ⑴) is in the form of multiple bytes, B / 4, the success of the meal and / or can be received Bit New

form. Here, during the month, the data is transmitted through the transmission 4 and the reception control section 326, and the envelope bandwidth manager 3

: Depending on = network link 302 or received by network link 302. Assuming that for-given allocation in a given period does not exceed, the envelope bandwidth manager 31〇 can Only passively monitor the circulation information: However ,:-once used-the distribution for a certain period is used up, then further transmission and reception are prevented 'until a new distribution is obtained (for example, in a subsequent Period). As shown, queues 328 and 330 can be provided to retain the overflow until the envelope is filled with a new distribution as the point in time at which point the queued data can be transmitted and / Or receive. When an allocation is not exceeded, the queues 3 2 8 and 3 3 0 are not used, so that the data transmitted or received is transmitted through the transmission control P knife 3 2 4 and the reception control part 3 2 6 without In these implementations, the queue needs to be triggered only when the envelope is used for a given period. Regarding the agent device described herein, the term envelope should be widely regarded as 57 200304298, which means that it is intended to be used above the ship agreement layer (for example, in layers 6, 7, and f 〇 layer 1 2 4) Internet conversations or links implemented. In the case of the second, these dialogues or connections are characterized by at least one or two, and the flow control is implemented by the force of °°. Therefore, it should be understood that the,, and are not limited to any particular standard within the upper layers of the agreement stack. In fact, the generations described here are fast-food; [fruitful, savvy, and savvy ... one hundred Internet conversations implemented by applications that do not use the widely popular Winsock standard.

‘Another Township Photo 2 丄 Picture’ The agent computer 2 2 is shown; 2: Ding Ying: 3 2 0a ’3 2 0 b and 3 2 0. . Apply the 3 2 0 "wooden 1 ° envelope (Wi_ck) standard and consequently apply the lower layers (eg, 3 * 〇 and 3 4), which are stacked through the window envelope to the user 1 to 8 ) And communicate with the network link 3 Q 2. Therefore; 〇 ... as shown in the figure relative to the application 128 settings to catch; Features. In contrast, 'Applications 3 2 0 b and 3 2 0 c are not related to this window •'

Standard together # Use. / 々 # An example of using such an application is to adopt a

Native Input Output System (NetBiQS) standard. For these non-windows I alternative agent modules 7 0b can be provided for: category :: module 7 0a and the previously described implementation of the agent module 7 0 b The system can hook the network f and A above the transport layer. The system does not require a window envelope, although its flow control can be different. Different parameters can be accessed to strengthen monitoring and policy; The &amp; Management's comparative agreement layer is still active. 58 200304298 Returning to the discussion of bandwidth management, many different methods can be used to provide downstream bandwidth allocation. Many of these methods use a priority mechanism, where priority can be specified through network policies. Priority can be specified according to different parameters, including Internet Protocol source address, Internet protocol destination address, source port, destination port, protocol, application identity, user identity, device identity 'user Route list (User Route List, URL), available device frequency I application queue, temple server corridor, gateway identity, router body

Injury time, network congestion, network load, number of network users, available domain bandwidth and resource status. Other parameters are possible. As mentioned above, the position of the agent module relative to the stack of the protocol allows access to many different parameters. The ability to describe bandwidth management under such changing conditions allows for increased control over bandwidth allocation, and by increasing efficiency. In a typical implementation, the above-specified priority can be converted to a priority level or expressed according to the priority level, the priority

The level of power is used in particular for a given web 4 "conversation, linking, or wording-an envelope data stream. In fact, many of the dispensers described below are based on envelope priority for individual envelopes In this system based on political load, the envelope priority is typically constructed to constitute a value derived from the different parameters of the above-mentioned women. For example, it is used for a The priority of the construction given to the envelope of = can be determined by the combination of user identity, health, application form of T, and the source address of the requested data. Τ '59 200304298 Again refer to the first Figure 9 will now describe an exemplary method for sub-allocation of bandwidth. First, sub-allocation for an application becomes one or all envelope allocations, as can be implemented by the application bandwidth manager 308. "Suppose for a given application 3 2 0 there are k envelopes 3 2 2 open. For a given envelope 3 (where 3 = 1 to 10), a valid envelope priority ESP (s) can be calculated as: m where 'SP (S) is the construct used for the envelope Priority (eg ^ ^; cons.lasU (s) ^ ^ t during the N allocation cycles, the bandwidth consumed by the envelope; aHoOastJCs) is allocated to the last N allocation cycles The bandwidth of the envelope. &Amp; The result can be multiplied by a convenient factor to avoid processing the number of fractions. J. Later, the envelope allocation for a given envelope, salloc (s) can be Calculated as ··· SALLOC (s) = APALLOC ~ ^ L · (2) lLESP {s) ▲ The towel, apappocm is currently being used by the application's bandwidth manager 308. Application assignments assigned in different envelopes; 1 Esp (S) is the effective priority of the envelope. Therefore, it should be understood that the allocation is derived from the effective envelope priority, a weighted average. 60 200304298 Therefore, in the above example, the envelope allocation will tend to be higher for envelope systems with higher priority and higher for envelopes that consume a larger percentage of the previous allocation. In the above example, the calculation will be based on a plurality of past allocation cycles (N cycles). A single cycle calculation can be used to make the calculation of the cycle smoothly and stably adjusted. Typically, the above-mentioned consumption data depends on, although it is used. In some cases, a plurality of leaves different between different envelopes are based on the bandwidth consumption data received by the envelope bandwidth manager 3 1 0 below. Delta Hai implemented by the bandwidth manager 308. Now it is described that the allocation of the agent becomes one or more application allocations, if it can be implemented by the application's bandwidth manager 3 0 'assuming it has] · applications 3 2 0 are executed on a given computer 2 2 on. First, a valid application priority can be calculated for each application. Using the above application as an example, the effective application priority EApp can be calculated as follows: EAPP = yS £ ^ Jast ^ N (s) x SP (s) r L 77-: ~ —-— x factor fo \ alloc last N (s) ^ 〇) Again, a factor can be used to avoid processing scores, or to facilitate processing. A similar calculation is performed for all other applications 3 2 0 located below the agent's bandwidth manager. Then, similar to the envelope sub-assignment described above, the individual application allocation can produce a clear tongue with a weighted average of 'for a given application ap (where ap = 1 to] ·)' an application allocation APALLOC (ap ) Can be calculated as 61 200304298 APALLOC (ap) = AGALLOC x ίΕΑΡΡ (αρ) (4) ap =] where AGALLOC is being allocated by the agent's bandwidth manager 3 〇6 in this different application And the EApp (ap) is the effective priority of the application for which the allocation is being calculated. Typically, these calculations in the morning are performed by the agent bandwidth management manager located above based on the bandwidth consumption data received by the application bandwidth manager 308 located below. In the above example, the allocation of applications will tend to be higher for applications with more envelopes open, and higher for envelopes that consume a larger percentage of previous envelope allocations. Now the description of the link bandwidth (for example, the overall bandwidth available on the network link 302) becomes one or more agent allocations, such as the controlled location bandwidth manager The implementation of 3 04 is assumed to have i agent computers 2 2 interconnected with the network key 3 02. As with the above secondary allocation, a valid priority can be calculated, and then, Lupin II can be allocated based on a weighted average of the valid priorities. The explicit agent uses the above-mentioned agent as an example. The effective agent priority EAGP can be calculated as follows by adding her effective application priority to Ding + Jingyou plus t below: ΕΑ〇Ρ ^ ΣΕΑΡΡ (αρ) (

ap ^ \ \ U J 62 200304298 A similar calculation can be implemented for all other agent computers located below the controlled location bandwidth manager 3 0 4 to obtain the priority of the valid agent. In some cases, it is expected that the &amp; · Gang value will be used to judge the most recently calculated value to obtain the valid priority 'to smooth the adjustment and frequency between all the agent computers. Wide redistribution. In some implementations, for example, it has been proven to calculate the effective agent priority by mixing the new: value (derived from the above equation) and the most recent value at a 60-4.0 ratio System is advantageous. Any other desired weighting system can be used, and / or other methods can be used to smooth the transition of allocations. In any case, any allocation to the agent computer can be implemented with a weighted average: AGALLOC (ag) = BWCLx ^^ GP (ag) Y, EAGP {ag) (6)% = ι

Among them, AGALLOC (ag) is the agent allocation for a given agent electricity ton Ug = 1 to i); it is the available bandwidth to be allocated among all the agent computers; and E (referred to as the agent for the financial benefit of the given agent m brain priority in the above example 'separate envelope activity is typically transmitted upstream (for example, from an envelope bandwidth manager to- Used bandwidth manager at the top, to _ agent bandwidth manager at the top and to a controlled position bandwidth manager at the top), and there is one in-step in the entire system Distribution and sub-distribution effects. Consumption at a particular envelope can affect the further distribution of the envelope and other envelopes used in the same 63 200304298. In addition, because the consumption data is passed upstream and used In other distributions and sub-distributions, the same envelope may affect the further allocation of all envelopes, applications and agent computers in the system. In some embodiments +, the effect of the feedback System can significantly increase the efficiency of bandwidth management.

Those familiar with this technology will understand that distribution and sub-distribution can be implemented for transmitted data and received data, respectively. In addition, the upstream and downstream feedback can be generated at any desired frequency or at irregular intervals. Within the agent computer 22, the sub-allocation is typically implemented at a fixed interval, that is, it is expected that the envelope allocation will be calculated at a frequency greater than the application allocation. As an example, the agent can be updated every 2 seconds for allocation, the application system is updated every half a second, and the envelope allocation system is updated every 1000 milliseconds. When all the envelopes involved in Tian are open, have a non-zero allocation, and are at least somewhat active (for example, consume certain bandwidth), the above exemplary equations are mainly used under steady state operation. Typically, some provide

And modifications are implemented under initial conditions, and are implemented under conditions where the envelopes are idle and / or have envelope assignments that are negligible or zero. For example, in the case where an agent computer may have started to communicate with the network application. In this case, the valid application priority will be all zeros, because there is no envelope activity at all, resulting in an undefined result of equation (4). A direct average can be used to resolve this October situation, in which case the available bandwidth is evenly distributed among the applications. Even distribution can also be used in the case of undefined results in equations (2) and (6 64 304298). The other bandwidth is Γ square private type (1) ′ The application bandwidth manager and / or it is constructed with ΓΓ to provide excessive bandwidth. This can be implemented in advance: because the envelope set by Γ maintains a valid envelope of zero, d does not get a positive allocation (for example, by applying the weighted berber of equation [2), ^ allocation to the becoming active Leisure can also be used to provide an internal non-zero valid envelope Jr to allow these enveloping envelopes to achieve,, D, and the right to claim, and thereby obtain the response that is being addressed by the above. A percentage of the allocation is applied to the edge of the distraction blade. Regardless of the specific method used, the rabbits can broadcast one plus-the method, the bandwidth can be between a specific water supply element or the envelope of the element that is located in the state of the element or under the physical distribution ^ / 牛Be assigned. The envelope state can be included in a given control module, a given number of opened envelopes within a given range, or used for-given response) The above equations (3) and (4) are partly based on k The number of envelopes opened in each of the eighth-order applications is a dry case of 4 '. All other matters are the same, in these equations, one and right. On the next day, 1 J-claw, high right m &amp; open application will have a more effective priority, and will receive a larger percentage of the 4-band configuration bandwidth. In the text, it should be understood that the envelope-like energy __ specified by ^ ruthenium-like resentment can also include the envelope's quilt 杻 4 first. As mentioned above, the priority assigned from the Jinshunshun can be deduced from an array of parameters that are almost ... restricted, and can be sighed by the Internet policy. Each of the above equations is defined according to the priority of the envelope, and χ provides the number of distributions and sub-distributions that are being applied, according to the priority of the envelope. All other matters are the same. Yes, an application with a higher envelope &amp; no * or an application or computer with a higher priority of 65 200304298 will receive a larger allocation in several of the above examples. The envelope state can also include the consumption of past allocations, as should be understood from the exemplary equations described above. The above examples provide several examples, which are configured by the envelope, the application and the computer mobile neck wide knife that consume the less allocated portion in the past, which are relative to the envelope, the application and In terms of computers.

The implementation of the above-mentioned anomalous system and method is to enable the bandwidth resources to be &amp; to a higher priority or important work. The bandwidth can be used by a relatively low cost entity and / or consumed by a relatively idle entity compared to its counterpart. The different embodiments described above can be implemented to provide a coarser or finer meat to the 糸 associated meat _ 4 main-π watts and the control of the bandwidth of the 7G pieces in the system # 1 special Yes, the above-mentioned bandwidth management 11 machine shots are constructed to constitute: to: the envelopes and / or cross-protocols of multiple dialogs implemented by an application. For example, one gives ^ ^ ^, 疋The application can implement multiple functions with changing priority levels.

^^^, The above-mentioned increased metric allows the authors of the system to specify the relative bandwidth preferences ... I prioritize the work to be performed by a given application. Fine-grained control over bandwidth usage is possible due to envelope priority

After being designated, provide envelope priority ’,! ·, $, ^ ~ && sigh. As in the above J ^, sleeve priority (such as the envelopes in the equations (N1 ^ and C3); narrative presumption) is typically issued only when the envelope is determined, and the date is opened on the relevant computer. . Although the policy parameters of the iron ± 仃 not network push pen cover are activated when the cover is opened to maintain static 66 200304298, some important tasks of the network work are being implemented by the cover, Xiao., DD mode, and Instead, it can be constructed to dynamically change the priority of the envelope. The task of extending some important tasks can use any desired set of policy parameters. ^ And it is defined through the above-mentioned policy mechanisms. In testing a pre-defined work, the designated priority used for the envelope is modified ^ such as by vetoing the designated value with a higher priority). The “modified” land is adopted to ensure that the level of a desired service or resource (for example, a desired allocated bandwidth) is desirable for the scheduled work. Alternatively, the predefined work may be Work for a lower priority% φ 榣, so that the envelope priority ▼ is dynamically downgraded to ensure that the bandwidth can be used by more important work. Any of the above embodiments or method implementations can be changed to provide Such a dynamic modification of priority. The measurement control (such as the embodiment of the agent module described herein) under the uncle's easy level is to enable the system to detect the beginning of important transactions. When such a transaction is detected, 'the system is able to communicate with the link of an upstream file to ensure that a sufficient amount of bandwidth is retained to complete the transaction within an acceptable time. For example, In other words, a dynamic increase in envelop priority will typically result in not only the envelop but also applications related to the envelop and the applications being executed thereon (for example, through the demonstration above). The application of the well-known formula) is used to calculate the increased allocation. Or, for low-priority transactions, the bandwidth can be dynamically reduced, and it can also be passed upstream to affect the allocation at different levels. For example In other words, the above-mentioned exemplary system can be constructed so that the bandwidth priority of important tasks is only assigned to a data system that is requested by a given web server in response to a request. In this example, the network policy access to a specific / on a customer's computer-a device that Liu Lan attempted to belong to-a: page, the envelope priority is blocked, to ensure that it is important, Unimportant graphics are provided at a service level. Other materials (such as ^. 9 贝 Λ) can be assigned a lower priority. More specific information, where to enter in Tongdang, a protocol for public awareness. In loading a single web page, it may involve several "acquisition," orders issued by the guest browser ^ 4 ^, and all the information that ^ ,, and so on. The agent module described herein can be described as follows. The agent module is based on the policy of monitoring its related computers, and is configured to find the predefined high priority of searching the information on the webpage. Order of obtaining P knife. When such an acquisition order is issued by the detective side, the 4-pack level priority will be dynamically adjusted for that part of the webpage data. Another example showing the benefits of increased metrics is the use of multimedia applications to send sound, video, and other data over a wide area network. In the case of high network congestion, it is typically expected to assign a higher priority to the audio data stream. The increased measure of the narrated system allows the priority to be applied through an appropriate network policy and guarantees that the video stream does not adversely affect the transmission of the audio data in a substantial manner. The systems described herein may further be provided with a layered communication architecture to facilitate communication between components. A typical implementation involves the use of a communication reservoir to represent a communication separation layer to different communication elements. For a given communication element, the separation layer is structured to hide the details of the communication mechanisms and / or the location of other communication elements. This system allows these components to be transparent to the transport structure used below. The transport agency can be a transfer control agreement, a user agreement, or any other transport agreement. The transport agreement can be replaced by a new agreement without affecting the components and their operation. The components are "transmitted objects, and other information" in the component interface. The communication layer can transform these objects into two

The list is exhausted to a format that provides information that will be transmitted and / or edited through any media. These operations are typically performed so that they are transparent to the communication elements. You can respond to these communication layers: increase the network efficiency through multiple × several communication streams to become a single link. Although the implementation of the embodiments and methods of the present invention has been particularly indescribable, 'However' Those skilled in the art of m will understand that, without departing from the spirit and scope defined in the scope of the appended patent application, Huxi Chemical Technology can be completed. These narratives should be understood to include all novel and non-obvious combinations of elements described herein. And the scope of the claimed patent in this case is presented as any combination of novel and non-obvious components. Where the scope of a patent application states "a ,, or" a first, element, or equivalent thereof, such a patent scope should be understood to include a combination of one or more such elements, rather than a request or Exclude two or more such elements. Brief Description of the Drawings Figure 1 is a conceptual diagram of the open system interconnection and transmission control protocol / route protocol model; Figure 2 is a distributed network in which software, systems, and methods described herein can be used Diagram of the road system; 69 200304298 Figure 3 is a schematic diagram of a computing device that can be used in the decentralized network of Figure 2; Figure 4 is a display that can be used in a Figure 2 is a block diagram of an exemplary agent module module of a decentralized network resource; Figure 5 is a display showing the different software that can be used in conjunction with the described software, Block diagram of exemplary components, including two generation modules, a control module and a structural device;

Figure 6 is a block diagram showing an exemplary configuration of a layered protocol stack of a computing device = an agent module; 疋. Figure 7 is a layered display of a computing device Agreement: a block diagram of another exemplary configuration of an agent module; Figure 8 is a demonstration of two of the agent modules of a layered agreement about a computing device A block diagram of the configuration; Figure 9 is a demonstration of an embodiment of the agent module, a block diagram of the file part; an embodiment of the control module is used to divide among multiple computer demonstrations.

Fig. 10 is a block diagram showing a part of a sexual component; Fig. 1 A is a flowchart showing a method of displaying a bandwidth; A picture used in a plurality of computers is a 1 B Flow chart showing the method of assigning another bandwidth; Figure 11C in multiple computers is a flow chart showing another method of using bandwidth allocation; 70 200304298 The first ID picture is one display and another A flowchart of a method for allocating bandwidth among multiple computers; Figure 12 is a flowchart showing a method for monitoring the status of network resources; ~ Figure 13 is a flowchart that can be used Diagram of the main structural compartment of the structural equipment linked to the software, systems and methods described here; Figure 14 is a diagram of another structural compartment of the structural equipment shown in Figure 13; ° Fig. 15 is a diagram of another structural compartment of the structural equipment shown in Fig. 13; Fig. 16 is a diagram of another structural compartment of the structural equipment shown in Fig. 13 Figure 17 is a schematic illustration of the tube that can be used with the tube described herein. Management software, systems, and methods—a multi-layered control structure; Figure 18 shows an exemplary network structure that allows a centralized definition of network policies and transmits those policies to the appropriate decentralized locations. Figure 19 shows an exemplary multi-layer system and method for providing a rough idea through the use of bandwidth by individual bandwidth consuming components. Figure 20 shows a package that contains the components to manage an individual application envelope or transaction. An exemplary computing device of an embodiment of an agent software module that consumes bandwidth; and FIG. 21 shows an example of an alternative / livelihood device carrying an alternative embodiment of the agent module. The agent The device module is constructed to monitor and manage 71 200304298 applications executed on the computing device and network communication data from the applications executed on the computing device. 〔Description of component symbols〕 10 Distributed network 12 Local network 14 Long-distance network 16 Public network 1 8 Router

2 0 server 2 2 client computer 2 4 network link 4 0 computer system 4 2 processor 4 4 bus 4 6 memory 4 7 random access memory components

4 8 Read-only memory and / or other static storage media 4 9 Basic input / output system 5 0 Data storage device 5 2 Network interface 5 4 Display device controller 5 6 Input device 7 0 Agent 7 2 Control point 72 200304298 7 4 Decentralized Network 7 6 Domain 9 0 Network Bonding 9 2 Protocol Stack 9 8 Application 106 Structural Device 12 2 Application 124 Transport Layer

128 Application program interface 13 0 Redirector module 132 Circulation data control module 13 2b Receive queue 13 4 Manager module 13 6 Area name system server module 13 8 Pop-up application module 14 0 Message middle module

14 2 System service module 14 4 Pop-up application 16 0 Circulation material module 16 2 Server profile module 16 4 Area name system server profile module 16 6 Gateway profile module 16 8 Manager module 17 0 Message middle module 73 200304298 1 7 2 Pop-up application interface 1 7 4 Pop-up application 1 8 8 Main structure screen 1 9 0 Screen frame 2 0 0 Centralized data center 2 0 2 Network 2 0 4 Enterprise policy server 2 0 6 Controlled Location Policy Server 2 0 8 Branch Office 3 0 0 System 3 0 2 Network Link 3 0 4 Controlled Location Bandwidth Manager 3 0 6 Agent Bandwidth Manager 3 0 8 Application Bandwidth Manager 3 1 0 Envelope Bandwidth Manager 3 2 0 Application 3 2 2 Envelope 3 2 4 Transmission Control Section 3 2 6 Reception Control Section 3 2 8 Transmission Queue 3 3 0 Reception Queue 3 4 0 Transport Layer 3 4 2 layer

74

Claims (1)

^ UUJU4298 Patent application scope: · ―A system for dynamically allocating bandwidth on a network link X The network link is interconnected with a plurality of computers, which includes one control = and a plurality of Agent computers, # 中 'each—each agent computer can be used with one or more envelopes or multiple applications. The system includes a control group, and the remote control module is suitable for the control. Running on a computer; and a plurality of agent modules, each of which is adapted to execute on one of the agent computers, wherein each agent module is suitable for monitoring The envelope status on its related agent computer, and repeatedly communicates with the 5H control module according to such a cover state. /, ZhongHai &amp; manufacturing module is suitable for connecting according to these agent modules. Coming from the flood, and for the agent computer, the wide agent allocation on the network link is determined, so that the agent allocation for an agent computer is relatively based on the envelope state of such an agent computer. Force _ other generations The state of the envelope of the processor computer is determined, and each of the agent modules includes a plurality of layers of the bandwidth management element, and the plurality of layers are suitable for the agent allocation of the associated agent computer. . 2 · If the system for dynamically allocating bandwidth on a network link in item 1 of the patent application scope, wherein the control module is suitable for the relative situation between the envelope states of the agent computers, The agent assignments are updated dynamically and repeatedly with 75 200304298. 3. As in the system of patent application No. 2 for the system for sadly allocating bandwidth on a network link, the control module is suitable for updating agent allocation, so that the consumption of agent allocation in the past is low Percentage of agent computers and δ, agent allocation is reduced. 4. For the system for dynamically allocating bandwidth on a network link in item 1 of the scope of the patent application, where 'each' of each agent module is suitable for executing a plurality of corresponding agent computer systems In the case of application, The next time it has been assigned by the next generation, it becomes one of these applications. 5. The system for dynamically allocating bandwidth on a network link is stated in the patent scope item 4, wherein the agent module is suitable for the situation between the envelope states of the applications, and Update the application allocations repeatedly and repeatedly. 6 · If the system of patent application No. 4 is used to dynamically allocate the bandwidth on a network link, the direct φ, benefit rte ^, shanger, nine soldiers, the application allocation of a parent is based on the use of The corresponding application depends on the number of envelopes opened for other applications used to receive a portion of the agent allocation. 7 As stated in item 4 of the patent scope for a system for dynamically allocating bandwidth on a network link, basin Φ, — y Xi Yu, Xiu Zhong, 'an application allocation is based on the corresponding application relative to receiving This review is based on the assigned priority of the envelopes of other applications that are part of the proxy distribution. 8 · Please refer to item 4 of the patent scope for each application allocation on a network link based on the dynamically allocated bandwidth system, of which 76 200304298 corresponds to the application receiving a part of the agent allocation Depending on the previously allocated consumption of other applications. '9 · The system used to divide the bandwidth on a network link as claimed in item 4 of the scope of the printed patent' where 'I _ application allocation is based on The number of envelopes opened for other applications that are part of the agent allocation, the assigned priority of the envelopes, and the previously allocated consumption of those envelopes depend. 10 · As claimed in item 4 of the patent scope, a system for allocating bandwidth on a network link, wherein each agent module is suitable for a plurality of applications in a computer system. In the case of envelopes, the application allocation for which the allocation is so applied becomes an envelope allocation for each of the envelopes. 1 1 · The system for dynamically allocating bandwidth on a network link, as in item i 0 of the scope of patent application, wherein the agent module is suitable for the relative situation between the consumption activities of the envelopes And dynamically and repeatedly update the envelope assignments. 1 2 · The system for dynamically allocating bandwidth on a network link _ road as described in item i 0 of the scope of patent application, wherein each envelope assignment is based on the envelope used for the corresponding one relative to the one used to receive the application Assignment of part of the other envelopes depends on the assigned priority. 1 3 · A system for dynamically allocating bandwidth on a network link as claimed in item i 0 of the scope of patent application, wherein each envelope allocation is based on the corresponding envelope for the corresponding application allocation for receiving the application. Part of the other envelopes are consumed by previous allocations. 77 200304298 1 4 · If the system for dynamically allocating bandwidth on the 丄 th road of the scope of the patent application, two of them: the corresponding envelope in a network link is divided into ^ sets of allocation system based on the corresponding application Assigned priority and other seals from such envelope = part :. The previous allocation of the weight loss is 1 5 · If the system of patent application scope item i is used to dynamically allocate bandwidth on a network, the zhongzhong _ 路 链 鲜 -Ύ an agent module system p contains a bandwidth manager, and is suitable for '' '' σ One of the multiple applications on the agent computer-one application bandwidth management cry 2: Originally used for each of the multiple envelopes related to those applications Bandwidth Manager. Τ # .If the system of patent application scope No. 15 is used to allocate bandwidth on a network ... The agent bandwidth management] is the agent suitable for receiving its related agent computer Server allocation, and: assign this agent allocation 'to be below the agent's bandwidth manager' Each application bandwidth manager receives a low application allocation for its related application, and ^ per-application bandwidth manager is suitable for sub-allocation of the other, using 4 application points g &amp; , So that the female bandwidth manager under the bandwidth manager of the application receives a sleeve allocation for its corresponding sleeve. 1 7 · If the system of Patent Application No. 6 is used to allocate bandwidth on a network link, each envelope bandwidth management system is constructed to report the consumption data to the upstream. Envelope Bandwidth Manager 78 200304298 The application bandwidth manager above the application. 1 8. As a system for dynamically allocating bandwidth on a network chain according to item 7 of the scope of patent application, where __ application bandwidth managers are built to manage bandwidth based on the application The effective envelope priority of the envelopes below the device is assigned next, and the application allocation is assigned, wherein the effective envelope priority for a particular envelope is based on the two assigned priorities of the envelope and according to the The envelope uses the previous envelope assignment. 19. A system for dynamically allocating bandwidth on a network link according to item 16 of the scope of the patent application, where 应用 application bandwidth management M is constructed to report consumption data upstream to the application frequency. The agent bandwidth manager above the bandwidth manager. 2 0. Please refer to item 9 of the patent scope for a system for dynamically allocating bandwidth across a network key, where #a proxy bandwidth manager is constructed based on the location of the proxy. Bandwidth Management Envelope Valid Envelope Priority _ Secondly Assigned by the Agent ^ Too Effective Envelope Priority for a Specific Envelope is based on a specified priority of the envelope and according Depending on the previous envelope assignment. 2 1 · As claimed in item 16 of the patent scope, a system for dynamically allocating bandwidth across a network link, where #a proxy bandwidth management benefit is constructed based on the proxy frequency Wide bandwidth manager and the application bandwidth manager located below the agent =! Manager and the application bandwidth &quot; Liyi and the envelope bandwidth manager located below the application bandwidth manager 79 200304298 Interaction between them, dynamically and periodically updating application allocation and envelope allocation. 2 2 · For the system for dynamically allocating bandwidth on one network link according to item 21 of the scope of patent application, the envelope allocation is updated more frequently than the application allocation. 2 3 · If the system for dynamically allocating bandwidth on a network link is described in item 1 of the scope of the patent application, each agent module is suitable for: assigning an envelope priority to an An envelope opened on the agent computer related to the agent module; a bandwidth-provided envelope is allocated to the envelope according to the priority of the envelope; and whether the envelope is being used to implement a predefined network After such detection, the parent changes the envelope priority of such an envelope, and updates the envelope assignment to calculate such an amendment. 2 4 types of agent modules built on a computer,-for network bandwidth used by a computer for 7 reasons, the agent module includes: an agent bandwidth manager; and An application bandwidth manager, the agent module is constructed to start an application bandwidth manager for each of a plurality of applications running on the computer, and the agent bandwidth management is established. The controller and the application bandwidth managers are interacting with each other to allocate the available network frequency of the computer 80 200304298 to become the user's individual application allocation for each of these applications. 5. If the construction of item 24 in the scope of patent application constitutes an agent module that is carried on a meat, each application allocation is based on the pair- relative to the other receiving the network bandwidth available to the computer. The condition of the envelope applied. Knife ^ • If the construction of item 25 in the scope of patent application constitutes an agent module loaded on an electronic device, in #, each application allocation is based on. The corresponding application depends on the number of open envelopes used by other applications that receive a portion of the network bandwidth available to the computer. ^ 2 7 · If the item No. 25 in the scope of the patent application constitutes an agent module loaded on an electronic device, each application allocation is based on the corresponding application relative to the network frequency available to the computer A wide portion of the envelope of other applications is assigned priority. 2 8 · If the construction of item 25 in the scope of patent application constitutes an agent module loaded on an electric shame, wherein each application allocation is based on the corresponding application with respect to the network bandwidth available to the computer Part of the other app depends on the previously allocated consumption. zy · If the construction of item 25 of the patent application scope constitutes an agent module loaded on a computer, each application allocation is based on the corresponding application relative to the portion of the network bandwidth available to the computer The number of envelopes opened for other applications, the priority assigned to those envelopes, and the previously allocated consumption of those envelopes. 3 〇 · If the application of the scope of the patent application No. 24 constitutes an agent module loaded on a computer, it includes an envelope bandwidth manager, 81 200304298 5 Hai agent module system constitutes a plurality of Open the mother of the sleeve on this computer to start a sleeve bandwidth manager like this. Like 3 1 · If the construction of the scope of patent application No. 30 constitutes a proxy module loaded on a computer, wherein the proxy module is constructed so that, when executed on the computer, the applications When one has a plurality of related envelopes, the application bandwidth manager associated with such an application dynamically interacts with the envelope bandwidth managers of the envelopes to allocate each of the envelopes. The app assignment becomes an envelope assignment. 3 2 · If the construction of item 3 i in the scope of patent application constitutes an agent module loaded on a computer, wherein the application bandwidth manager and the envelope frequency finders are constructed dynamically and The envelope assignments' are updated repeatedly in response to changing the state at the envelopes. It looks like 3 3 · If the construction of the scope of the patent application No. 24 constitutes an agent module loaded on a computer, wherein the agent bandwidth manager and the application bandwidth controller are constructed dynamically and The individualized application allocations are updated repeatedly in response to changing the status of the applications. 3 4 · If the construction of item 24 of the patent application scope constitutes an agent module on the computer, where the agent module is suitable for: assigning an envelope priority to one of the agent modules Is an envelope opened on the computer loaded on it; an envelope that provides a bandwidth based on the priority of the envelope is assigned to the envelope; and whether the envelope is being used to implement a predefined network parent Easy, and after such detection, modify the envelope 82 200304298 priority of such envelope 'and update the envelope assignment to calculate such a modification. 3 5 types of agent modules built on a computer, which are used to manage the network frequency modules used by such computers include: ° an agent bandwidth manager; and an envelope Bandwidth manager, the agent module is constructed to start each of a plurality of envelopes that are opened on the computer, such an envelope "manager", the agent bandwidth adapter and The envelopes ν page width S benefit system is constructed to interact to allocate the network bandwidth available to the computer to the individualized envelope allocation for each of the envelopes. 4 Agent Bandwidth Manager and The envelope bandwidth managers are constructed to update the envelope allocations based on the consumption activities of the envelopes with each other, and to renew the envelope allocations. A structure is constructed to load an agent module on a computer. Group, which is used to manage the network bandwidth made by such a computer, the agent module includes: an agent bandwidth manager; an application bandwidth manager, the agent module is constituted For multiple executions On each of the applications on the computer, an application bandwidth manager is started, wherein the agent bandwidth manager and the application bandwidth managers are constructed to interact with each other to allocate the available computer bandwidth. Network bandwidth becomes an individualized application allocation for each of these applications; and an envelope bandwidth manager, the agent module is constructed to constitute each of a plurality of envelopes opened on the computer Start an envelope bandwidth manager such as 83 200304298, where, for an application with multiple envelopes, the corresponding application bandwidth manager and the envelope bandwidth management and ancestry constitute an interaction, The application allocation is divided into envelope allocations for each of the _ envelopes. 3 7 · —A system for dynamically managing bandwidth consumption on a network link by a computer, which includes · · A software-based agent module, the software-based agent module is constructed and executed on the computer, and the agent module is further built Success: _ When multiple envelopes on the computer are open, assign an envelope priority to each of the multiple envelopes; sub-allocations can be used by the overall bandwidth of the computer to become each of the multiple envelopes Envelope allocation, wherein the allocation of the envelopes of each of the plurality of envelopes is determined according to the priority of the envelopes over the other envelopes; and when one of the envelopes is being used to implement a Pre-defined online transactions, and after such detection, modify such an envelope "&quot; Xuan Envelope Priority, and update the envelope assignment to calculate such an amendment. 3 8 · If the scope of patent application is 37 A system for dynamically managing bandwidth consumption on a network link by a computer, wherein the pre-defined network transaction is only related to a network source requested by the computer Part of the data, so that when the envelope is used to obtain the part of the information, the priority of the envelope is modified, and when the envelope is used to obtain For other parts of the information from the road source, the priority of the envelope is not modified. 3 9 · — A system for dynamically managing bandwidth consumption on a network link by a computer, which includes: a software-based agent module, which is built using a software agent module It is constructed and executed on the computer, and the agent module Z is further constructed: It refers to a envelope priority to an envelope opened on the computer; an envelope that provides a network bandwidth according to the priority of the envelope is assigned to The envelope; and / or test is that the envelope system is being used to implement a pre-defined online transaction, and after such a primary test, modify the envelope priority of such an envelope, and update the envelope assignment to calculate this Its modification. J 〇. The system for dynamically managing bandwidth consumption on a network link by a computer as described in item 39 of the scope of patent application, where it is easy to complete the network after you have completed the pre-defined network 'The envelope priority is restored to an unchanged value. 〆41. A system for dynamically allocating bandwidth on a network link ', the first 4 network links are interconnected with a plurality of computers, and the system includes a plurality of agent computers' wherein' Each agent computer is capable of executing one or more applications using one or more envelopes, the system includes a plurality of agent errors, and each of the agent modules is adapted to execute on the agent. 4b Fari ++, and monitoring and repeated reporting-the agent is a computer and its related agent 11 computers-a jacket status; 85 200304298: a control module 'which is suitable for execution on the control computer,苴 ^ '^ Molding mode is suitable for reducing the relative envelope status of these substitute modules: == Computer relative envelope state, and dynamically enter the bandwidth of the network configuration on the network link of these proxy groups Each agent module is further suitable for: for assigning one when the envelopes are opened with the brain system; the agent related to the agent module has an envelope priority to the Each of a plurality of envelopes Detecting when these envelopes are in use— # The various systems are being used to implement a pre-defined online transaction, and during the pre-defined online transaction, the modification of the envelope for such an envelope takes precedence right. 4 2: A method of dynamically managing bandwidth in a decentralized network system. A decentralized network system has a network link interconnected with multiple computers, which includes: Designation: Envelope priority Weight to an envelope, wherein the envelope is combined with an application running on one of the plurality of computers; Provide an envelope allocation to the envelope; where the envelope allocation defines the network bandwidth that can be used by the delta envelope; use the envelope to monitor the implemented network transactions to identify whether the envelope is being used for implementation A pre-defined network transaction and after detecting that the system is being used ☆ After the pre-defined network transaction is implemented, the envelope priority is modified, and the envelope allocation is updated to calculate this. modify. For example, if the 42nd item of the scope of patent application is dynamically managed in one minute 86 200304298, the method of ": bandwidth in the road system" further includes: providing an application to be allocated to the application, where the application is allocated Set. The network bandwidth used by the money is related to the envelope, and _. Dynamically modify the application allocation to calculate the modification of the priority of the envelope ^ 44. As in the 42nd scope of the patent application A method for dynamically managing bandwidth in a distributed network system, which further includes: providing a computer to allocate to the computer on which the application is running, and the computer's distribution system can be defined. The network frequency t ^ used by such a computer is related to the envelope allocation; and the computer allocation is dynamically modified to calculate the envelope priority. Modifications and diagrams: as shown on page 87