JP2006324793A - Communication band allocation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system capable of allocating a communication band matching not only a user's requested band and network performance, but also the processing capability of a user terminal and a communication device to the user terminal. <P>SOLUTION: The communication band allocation system measures at least one out of the effective values of the processing performance of the hardware of the user terminal, the effective value of the software, and the effective value of the processing performance of the communication device, and determines a communication band to be allocated to the user terminal on the basis of the measured value. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technology for allocating a communication band used for data communication to an information terminal (hereinafter referred to as a “user terminal”) such as a user's desktop PC, notebook PC, or PDA connected to a network. The present invention relates to a technology for allocating communication bands according to the effective processing capacity of each user terminal.

  Conventionally, voice and moving images have been received by a television / radio using a dedicated voice / moving image line. However, in recent years, construction of a large-capacity network has progressed, and services for distributing voice and moving images via the network have become popular. However, in network distribution, the bandwidth that can be used by one user changes every moment according to the route status and the number of connected users. For example, when a large amount of data suddenly flows into the network, a bandwidth that can be used by the user is significantly narrowed, and a problem occurs in that a voice or a moving image being received is interrupted. This has been an unstable factor when transmitting voice, moving images, and the like over a network.

  In order to solve the above problems, a system has been proposed in which a management server allocates a necessary bandwidth to a user terminal in advance when the user terminal is connected to a network, thereby securing a necessary bandwidth for data transmission. In such a "system that requests a necessary transmission band in advance and secures a necessary band for data transmission", the required band is large depending on whether the transmission data is an audio / still image or a moving image. Although consideration has been given to changing the size, the size of the requested bandwidth is generally set relatively roughly by the user's sense.

However, in the above method, a user tends to request allocation of a larger band than actually used in order to request a sufficient band. As a result, the bandwidth allocated to the user is not fully utilized, and the network utilization efficiency is reduced. Therefore, in Patent Document 1 and the like, the performance of the communication interface (for example, 100 Mbps for a wired network, 11 Mbps for a wireless network, etc.), the required bandwidth calculated from the quality of the content desired by the user, and the bandwidth of the network path are shown. In comparison, a method has been proposed in which the management server allocates the lowest bandwidth value to the user terminals.
JP 2002-271383 A.

  However, in the method described in Patent Document 1, although the bandwidth value to be allocated is determined in consideration of the limitation conditions in the system standard, user requirements, network performance, etc., the processing of the user terminal or the communication device Since the effective value of the capacity is not taken into consideration, the bandwidth utilization may be allocated in excess of the value that the user terminal or communication device can actually process, and the line utilization efficiency is not sufficiently improved. . For example, in the bandwidth allocation method described in Patent Document 1, the user terminal compares the requested bandwidth value of 30 Mbps of the user with 100 Mbps, which is the standard value of the communication device, when content is distributed, and sets the requested bandwidth to 30 Mbps. On the side, considering that the bandwidth of the network path is 50 Mbps, the user's requested bandwidth value of 30 Mbps is determined as it is as the allocated bandwidth. However, when the user terminal actually has a processing capacity of only 10 Mbps, an excessive bandwidth of 20 Mbps is allocated to the user terminal. The same problem also occurs when the effective value of the processing capability of the communication device is significantly lower than the standard value.

  That is, since the distribution server cannot detect that the effective value of the processing capability of the user terminal or the communication device is lower than the processing capability in the network path, it may not be possible to perform sufficiently efficient bandwidth allocation to the user terminal. It happens. When a large number of user terminals are connected to the distribution server, by inefficient bandwidth allocation as described above, even if a user requests a bandwidth, it cannot accept the connection. It is assumed that something will be rejected.

  The present invention has been made in view of such circumstances, and not only a user request bandwidth and network performance but also a communication bandwidth suitable for the effective value of the processing capability of the user terminal and the communication device is provided to the user terminal. It is intended to provide a system that can be assigned.

  As a result of diligent research in view of the above problems, the inventor considered the effective processing capability of the user terminal and the communication device itself in addition to the bandwidth in the user system standard, the requested bandwidth of the user, and the network performance. Thus, the inventors have come up with the idea of determining the bandwidth to be allocated to the user terminal within a range that does not exceed the effective processing capacity of the user terminal or the communication device itself.

  That is, the present invention is a system for allocating a communication band to a user terminal connected to a network, the effective value of hardware processing performance, the effective value of software processing performance in the user terminal, and a communication device The communication bandwidth allocation system is characterized in that at least one of the effective values of the processing performance is measured and a communication bandwidth to be allocated to the user terminal is determined based on the measured value.

  In the communication bandwidth allocation system of the present invention, the effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the processing performance of the communication device are actually unit time for each of the hardware, software, and communication device. It is expressed as the maximum amount of data that can be processed per hit.

  In the communication bandwidth allocation system of the present invention, the effective value of the hardware processing performance is calculated using at least one of the data processing performance of the CPU, ROM, RAM, and hard disk of the user terminal. To do.

  In the communication band allocation system according to the present invention, a method for calculating an effective value of processing performance of the hardware is changed in accordance with a change in the hardware configuration of the user terminal.

  In the communication band allocation system according to the present invention, the effective value of the processing performance of the software is calculated using hardware parameters relating to the software presented by the OS of the user terminal.

  In the communication bandwidth allocation system of the present invention, the hardware parameters related to the software presented by the OS of the user terminal are the amount of data written to and read from the IO device by the software, the CPU usage rate of the software, and other than the software It includes at least one of a CPU usage rate of the process, a communication bandwidth required by the software, and a buffer capacity capable of temporarily storing data.

  In the communication band allocation system according to the present invention, a method for calculating an effective value of the processing performance of the software is changed according to a type of information to be processed by the software.

  In the communication band allocation system of the present invention, the effective value of the processing performance of the communication device is calculated using hardware parameters related to the communication device presented by the OS of the user terminal.

  In the communication band allocation system of the present invention, the hardware parameters related to the communication device presented by the OS of the user terminal are the amount of data transmitted and received by the communication device and the maximum value of the processing performance according to the standard of the communication device. It is characterized by including at least one.

  In the communication band allocation system of the present invention, the method for calculating the effective value of the processing performance of the communication device is changed according to the communication system standard of the communication device.

  In the communication bandwidth allocation system of the present invention, the user terminal continues at least one of an effective value of hardware processing performance, an effective value of software processing performance, and an effective value of processing performance of the communication device in the user terminal. And the communication bandwidth allocated to the user terminal is dynamically changed based on the measured value.

  In the communication band allocation system according to the present invention, at least one measurement value among the effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the processing performance of the communication device in the user terminal is set at a constant time interval. And the communication bandwidth allocated to the user terminal is dynamically changed.

  In the communication band allocation system according to the present invention, at least one measurement value of the effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the processing performance of the communication device in the user terminal is the same as the previous value. The communication band allocated to the user terminal is dynamically changed based on the comparison result with the measured value.

  In the communication band allocation system of the present invention, when a connection request to a network is received from a new user terminal, the sampling is performed on all user terminals currently connected to the network, and sufficient to allocate to the new user terminal It is characterized by checking whether there is a proper communication band.

  In the communication band allocation system according to the present invention, the bandwidth is determined based on at least one measurement value of an effective value of hardware processing performance, an effective value of software processing performance, and an effective value of processing performance of the communication device in the user terminal. The communication band is allocated to the user terminal only when the communication band to be allocated to the user terminal is a value that does not exceed the communication band of the route on the network with which the user terminal communicates. It is characterized by that.

  As described above, according to the communication bandwidth allocation system of the present invention, not only the user request bandwidth and network performance but also the communication bandwidth corresponding to the effective value of the processing capability of the user terminal and the communication device is provided to the user terminal. Therefore, the utilization efficiency of the communication band in the network is improved.

  The best mode for carrying out the communication band allocation system of the present invention will be described below in detail with reference to the accompanying drawings. FIGS. 1-11 is a figure which illustrates embodiment of this invention, In these figures, the part which attached | subjected the same code | symbol represents the same thing, and a basic structure and operation | movement are the same. To do.

The communication band allocation system of the present invention includes the following components.
・ Management server that manages the allocation of bandwidth to each user terminal connected to the network ・ Relay devices such as routers and switching hubs ・ User terminals such as workstations, desktop PCs, notebook PCs and PDAs ・ Effective hardware of user terminals Monitors that monitor data processing capabilities (hereinafter referred to as “hardware performance”) and effective data processing capabilities of software (hereinafter referred to as “software performance”), communication devices such as wired LAN adapters and wireless LAN PC cards, and processing details Display device

  In the following, for convenience of explanation, a case where the user terminal and the communication device are integrated will be described as an example. However, the user terminal and the communication device may be configured as separate hardware. In addition, a case where the user terminal and the display device have separate hardware configurations will be described as an example, but these may be integrated. Moreover, although the example which implement | achieves the monitor which performs communication band monitoring with the software memorize | stored in ROM in a user terminal is shown, this monitor may operate | move on a management server, or a user terminal, You may comprise by the hardware different from a management server. Hereinafter, the user terminal and the communication device are collectively referred to as a user system. A data transmission path from the CPU of the user terminal to the interface of the communication device of the user terminal is called a bus.

  The communication band allocation system of the present invention includes a wireless LAN system capable of band control such as HiSWAN (High Speed wireless Access Network) of ARIB STD-T70, a wired network having a band guarantee function such as ATM, etc. Thus, the present invention is applied to a system that controls the bandwidth of a user terminal by a management server. Moreover, in the communication band allocation system of the present invention, an example is shown in which the user terminal communicates with the management server via a relay device such as a router, but since it is not necessary to add any modification to the relay device, The present invention is also applicable to a system that performs bandwidth control using a relay device such as a router such as Diffserv.

  In the communication band allocation system of the present invention, first, before a management server that controls allocation of communication bands allocates a band to a user terminal, the user terminal requests a necessary band from the management server. Immediately before this bandwidth request, the monitor inspects the user system, extracts the application to be used and various hardware parameters (details will be described later), and obtains the effective value of the processing performance of the hardware of the application and the user system as a whole. Calculate and create a software performance list and a hardware performance list (described later). The user terminal makes a bandwidth request to the management server via the communication device, and the management server allocates the requested bandwidth to the user terminal. The management server requests recalculation of the values of the above-mentioned software performance list and hardware performance list of all user terminals at an arbitrary timing, extracts the values, and performs management such as dynamic bandwidth reallocation based on that value. Do. When a LAN network is constructed, for example, a base station having a router function corresponds to the management server. Further, the user terminal and the management server may be directly connected, or one or more relay devices such as routers and switching hubs may exist on the network.

Configuration of Communication Band Allocation System FIG. 1 shows an internal configuration of an information processing apparatus having a communication function constituting a user terminal in the communication band allocation system of the present embodiment, and a configuration of a network to which the information processing apparatus is connected. FIG. FIG. 1A is a diagram showing an outline of a system including a user terminal. As shown in FIG. 1A, the information processing apparatus (hereinafter referred to as “user terminal”) in the present embodiment operates on the CPU 1-1 that controls the operation of the operating system OS1-2 and the like on the OS1-2. Application 1-3, an interface to the communication device 2, a ROM 1-5 for storing a program such as an OS, a RAM 1-4 for temporarily storing instruction data of the CPU, and stored user data and programs. A storage device 1-6 typified by a hard disk, a graphics card 1-7 for sending an image to a display device, a monitor 3 for monitoring, and a bus line BL (hereinafter referred to as “bus”) connecting them. Have.

  The user terminal can be connected to the network via the communication device 2. A plurality of relay devices 5 and a management server 6-1 exist in the network. The management server 6-1 is equipped with a RAM 6-2, which stores sampling periods [s] (details will be described later) of all user terminals. For convenience, in this example, the number of relay devices 5 and the management server 6-1 are both one. For convenience, in this example, the user terminal is connected to the management server 6-1 via the relay device 5, but may be directly connected to the management server 6-1. Furthermore, the RAM 6-2 in the management server 6-1 is not limited to the RAM as long as it has a method capable of storing data. In addition to the OS, the ROM 1-5 and the storage device 1-6 store application software used for communication. The graphics card 1-7 outputs data received from an application or OS to the display device 4. The monitoring monitor 3 provided in the user terminal extracts performance information from each of the user terminal and the communication device 2, and after determining the bandwidth required by each, returns the performance information value to the communication device 2.

  FIG. 1B is a diagram illustrating a schematic configuration example of the monitor 3 for monitoring. As shown in FIG. 1B, the monitor 3 has a software performance list 7, and stores applications, requested bandwidth, hardware parameters presented from the OS, and the like. The monitor 3 further has a hardware performance list 8, and stores hardware parameters, bus widths, standard values of communication devices, and the like presented from the OS. The comparator (required bandwidth calculator) 9 receives the software performance and the hardware performance from the software performance list 7 and the hardware performance list 8, respectively, and determines the bandwidth value required for the network based on the comparison. The requested bandwidth value is passed to the communication device 2.

  FIG. 2 is a diagram showing a data configuration example of the software performance list 7 and the hardware performance 8 stored in the monitor 3 shown in FIG. FIG. 2A shows a data table of the software performance list 7. Specifically, the software performance list 7 includes an application 10, a requested bandwidth [bps] 11, an IO size [byte] 12, a cumulative IO size [byte] 13, a total buffer size [byte] 14, and a cumulative total. A buffer size [byte] 15, a CPU usage rate [%] 16 of the application before communication, a CPU usage rate [%] 17 of the application during communication, a usage rate [%] 18 of the entire CPU before communication, The CPU usage rate [%] 19, processing time [s] 20, processing bandwidth [bps] 21 (the previous value in parentheses) and sampling period [s] 22 are stored. ing.

  FIG. 2B shows a data table of the hardware performance list. Specifically, the hardware performance list includes a communication device 30, a bus width [bps] 31, a standard bandwidth [bps] 32, a transmission / reception (data) size [byte] 33, and a cumulative transmission / reception size [byte] 34. , Communication time [s] 35, NIC bandwidth [bps] 36 (the previous value in parentheses), and sampling time [s] 37 are stored. The description of each parameter shown in FIGS. 2A and 2B will be given later.

  Next, the operation of the communication band allocation system of this embodiment will be described. The flow of processing for requesting bandwidth by the user terminal in this system is shown in the flowcharts of FIGS. Further, in the present system, the flow of processing for allocating the bandwidth to the user terminal by the management server is shown in the flowchart of FIG.

(A) Method for calculating software performance The processing performance of an application cannot be determined simply by a description language, a description method, or a software element such as a compiler. When hardware elements such as the performance of the CPU that executes the process, the speed at which the program is read from the ROM or RAM, the data saving speed to the storage device, and the display data transfer speed to the graphics card are closely involved There are many. In addition, depending on the type of application, there is a type in which a special buffer is constructed on a RAM or a storage device, and data from a network can be temporarily stored even when the application main body is not running.

However, in order to directly measure the effective performance of hardware such as CPU, ROM, RAM, storage device, and graphics card, the user terminal hardware is directly modified every time the technology according to this embodiment is applied. This must be done, but this is cumbersome and impractical. Therefore, the following items are extracted from the hardware parameters presented by the OS installed on the user terminal, and are used as materials for obtaining the processing performance of the application and hardware.
1) Write / read size to IO device for each application [byte] (hereinafter referred to as total IO size)
2) Application CPU usage [%]
3) CPU usage rate [%] used by processes other than the above applications (hereinafter referred to as “CPU usage rate”)
4) Bandwidth required by application [bps] (initial value: hereinafter referred to as “requested bandwidth”)
5) Total buffer size [byte]

  Of the parameters 1) to 5), each value assigned to the application used by the user is extracted and added to the software performance list. By extracting these, even if the effective speed of the CPU, ROM, RAM, storage device, graphics card, etc. is not known, the data can be exchanged between the application itself and the IO device (user terminal in this system). You can know if you can. In addition, by considering how much data the buffer can handle, the processing performance of the entire application can be understood. The total IO size and the total buffer amount are values obtained by adding up the data amount from the time of starting the application to the present time.

  First, a bandwidth corresponding to the processing performance of the entire application is adopted as an initial value of the bandwidth requested to the management server before the start of communication. The value depends on the data processed by the entire application. For example, if the user wants to view music of 128 Kbps, the application uses 128 Kbps as the required bandwidth, and if he wants to play a 1 Mbps video, the application uses 1 Mbps as the required bandwidth. In some applications, the user can manually set the bandwidth value in advance. In this case, the value is set as the required bandwidth.

  However, since the initial value is a theoretical value or a desired value, it may be an excessive value as described above. Therefore, the actual performance is measured using the parameters 1) to 5). In addition, since the actual performance changes from time to time, the management server allocates an appropriate band according to the situation to the user terminal by sampling the above parameters at the timing notified from the management server. The management server writes the sampling period [s] to the software performance list via the monitor of the user terminal at the first connection. The management server stores the sampling period [s] of all connected user terminals in the RAM.

  First, the CPU usage rate (1) of the application before communication is extracted. Next, the total IO size at the start of communication is added to the software performance list as the accumulated IO size, and the CPU usage rate (2) of the current application is monitored as needed. Here, a period in which the CPU usage rate of the application before communication (1) <the CPU usage rate of the current application (2) is a time during which data processing is performed by the application, and the CPU of the application before the communication The time when the usage rate (1) = the CPU usage rate (2) of the current application is reached is when the data processing of the application is finished. The difference between the total IO size at this time and the IO size at the start of communication, that is, the accumulated IO size is the data size processed by the application in this communication, and the difference between the total buffer amount and the accumulated buffer amount is This is the data size that could be stored as a buffer in communication.

Here, if the processing time [s] is a period in which the CPU usage rate of the application before communication (1) <the CPU usage rate of the current application (2), the processing performance [bps] of the application main body during communication Is given by:
((Total IO size)-(Total IO size)) / (Processing time)
Hereinafter, a value obtained by this equation is referred to as an “application processing band”.

The processing performance [bps] of the buffer during communication is given by the following equation.
(Total buffer size)-(Total buffer size) / (Processing time)
Hereinafter, a value obtained by this equation is referred to as a “buffer processing band”.

The processing performance [bps] of the entire application is given by the following equation.
(Application processing bandwidth) + (Buffer processing bandwidth)
Hereinafter, a value obtained by this equation is referred to as a “processing band”.

  The time when the communication device performs communication and the time when the application communicates do not necessarily match. For example, it is conceivable that the application keeps processing even after the communication is completed, so it is not simply necessary to observe whether or not the communication device is operating. By monitoring the CPU usage rate of the user terminal from the start of communication, the time for the application to process transmission / reception can be determined.

  After the initial value is passed to the monitor and communication is started, the value of the parameter is measured during the sampling time. When the sampling time elapses and the recalculation time comes, the calculation is performed based on the formula relating to the processing band, and the posted processing band is written in the software performance list 7 shown in FIG. Then, the processing performance of the application finally determined by the management server is “the smaller of (requested bandwidth) or (processing bandwidth)”. Thereafter, during the communication period, the processing band is recalculated every time the sampling time comes.

  When the bandwidth is allocated for the second time, the immediately previous processing bandwidth is compared with the current processing bandwidth. The immediately preceding processing band is the requested band before communication. The processing bandwidth this time is the processing bandwidth calculated at the sampling time. The smaller of these two values is used as the processing performance of the application, and is written in the software performance list. At the time of bandwidth allocation after the third time, the processing performance of the previous application is compared with the processing performance of the current application, and the smaller one is set as the processing performance of the application for the next bandwidth request.

  Each time a bandwidth is allocated from the management server, the user terminal writes the total IO size and the total buffer amount as a cumulative IO size and a cumulative buffer amount in the software performance list, and resumes data processing until the next sampling time.

  If the amount of data that can be processed by the application decreases due to circumstances such as the burden on the CPU becoming heavy, and the processing cannot be performed without reducing the bandwidth allocated to the user terminal, the next processing bandwidth will be automatically reduced. . In this case, the current processing band may be used as the processing performance of the application as it is. On the other hand, if there is a possibility that the processing capacity of the CPU can be used and the application can process more data, it is necessary to consider the increment in which data can be processed and set the next processing bandwidth. In preparation for this case, it is preferable to monitor the current usage rate (3) of the entire CPU in advance. If this value is not 100 [%], it means that the CPU processing performance can still be assigned to the application. Therefore, when the recalculation is performed after the sampling time has elapsed, the processing capacity of the application is set so that the current application CPU usage rate (2) + current CPU usage rate (3) = 100 [%]. Can be increased. Eventually, the lower one of the processing bandwidth and the required bandwidth at this time is adopted as the processing performance of the application.

  FIG. 3 is a flowchart showing the flow of the above-described software performance measurement process. In FIG. 3, the system waits until communication preparations are made (step S101). When entering the communication preparation stage, first the application, its required bandwidth, and sampling time are extracted from the user terminal, and the IO size and CPU usage rate before communication are extracted from the hardware parameters presented by the OS of the user terminal. Then, a software performance list is created (step S102). It is determined whether communication has been performed using the above application and the processing bandwidth has been calculated (step S103). If the processing bandwidth is not listed in the software performance list, it is determined whether communication is in progress. (Step S104). If the application has been used before and the processing band has been described, the process proceeds to step S111. If communication is not yet performed, the requested bandwidth is passed to the monitor (step S105), and the start of communication is awaited (step S106). If the communication is in progress, the CPU usage rate and the overall CPU usage rate of the application being communicated are always measured, added to the software performance list, and this value is updated as needed (step S107).

  When the sampling time elapses from the start of communication and the software performance recalculation time is reached (step S108), the processing time is calculated from the CPU usage rate of the application during communication and the CPU usage rate of the application before communication. Are added to the list (step S109), the processing bandwidth is calculated taking into consideration the overall CPU usage rate and the CPU usage rate of the application (step S110), and then the previous value of the processing bandwidth and the processing bandwidth are compared. The larger value is compared with the requested bandwidth, and the smaller value is passed to the monitor as the processing performance of the software (step S111). Then, the current value of the IO size is added to the software performance list as the accumulated IO size (step S112). If the communication is completed before the sampling time has elapsed (step S113), the software performance measurement process is terminated. The process in step S110 will be described in detail later.

(B) Method of calculating user terminal hardware performance Next, the hardware processing performance of the user terminal is obtained. With regard to this hardware performance, performance such as the access speed to the CPU, ROM, RAM, and hard disk mounted on the user terminal and the processing performance of the graphics card is taken into account in the calculation of the software performance. The hardware performance is a bus bandwidth [bps] (hereinafter referred to as “bus width”). The value of the bus width of the user terminal is added to the hardware performance list shown in FIG. The bus width is also remeasured at each sampling time determined by the management server in the same manner as described above.

  FIG. 4 is a flowchart showing the flow of the hardware performance measurement process of the user terminal at the time of hardware performance measurement. As shown in FIG. 4, the present system waits until communication preparations are made (step S201). When entering the communication preparation stage, the bus width is extracted from the user terminal, added to the hardware performance list (step S202), and the bus width is passed to the monitor (step S203). Next, it is determined whether or not communication is currently being performed (step S204), and if communication is not currently being performed, communication start is waited (step S205). It is determined whether or not the sampling time has elapsed since the start of communication, and it has become the recalculation time of the hardware performance of the user terminal (step S206). If the recalculation time is reached, the bus width is measured again. (Step S207), the bus width is passed to the monitor (Step S203). If the communication is terminated before the sampling time has passed in step S208, the user terminal hardware performance measurement process is terminated.

(C) Communication device hardware performance calculation method As with the software performance calculation method described above, it is possible to directly measure the effective processing performance by modifying the communication device hardware for the purpose of applying this system. It is not realistic to do so. Therefore, regarding the hardware parameters of the communication apparatus, the following parameters presented from the OS are extracted and written to the hardware performance list.
1) NIC (Network Interface Card) send / receive size [byte] (hereinafter referred to as “total send / receive size”)
2) Theoretical value [bps] in the standard of the communication system (initial value: hereinafter referred to as “standard band”)
It should be noted that the total transmission / reception size is a cumulative value from the time of activation.

  First, a standard band is adopted as an initial value before starting communication as processing performance of the communication device. The standard band is determined by the communication method of the communication device and the standard of the communication device. For example, if the wired network is 100 Mbps, the standard band is 100 Mbps, and if the wireless network is 11 Mbps, the standard band is 11 Mbps. Depending on the communication device, the user may be able to manually set the bandwidth value in advance. In this case, the set value is set as the standard band.

Next, as with the calculation of software performance, the actual processing performance is obtained at the sampling time. For this purpose, it is necessary to observe the transmission / reception size for a certain period of time. First, the transmission / reception size before the start of communication is set as the cumulative transmission / reception size, and is added to the hardware performance list. The difference between the total transmission / reception size and the cumulative transmission / reception size after observing for a certain period is the data size processed by the communication device in the current communication. The observation time may be the time when the communication device transmits / receives data to / from the user terminal, that is, the time when data flows through the bus, so if this is the communication time [s], the observation time is
(Total transmission / reception size)-(Total transmission / reception size)) / (Communication time)
It is represented by Hereinafter, the value obtained by this equation is referred to as “NIC bandwidth”. The NIC bandwidth value obtained from the above equation is added to the hardware performance list shown in FIG. The communication device hardware performance to be finally requested is “the smaller of (standard bandwidth) or (NIC bandwidth)”.

  Thereafter, the NIC bandwidth is recalculated every time the sampling time is reached during communication, but from the second time onward, the higher one of the previous NIC bandwidth and the current NIC bandwidth is compared with the standard bandwidth. For example, if the NIC bandwidth this time is lower than the previous value, the amount of data that can be processed by the communication device will drop due to circumstances such as a load on the communication device or a network failure. This means that it has become necessary to request less bandwidth. The fact that this NIC bandwidth is higher than the previous value means that the amount of data that can be processed by the communication device will increase due to factors such as a light load on the communication device and a lighter network processing load. This means that processing can be performed even if more bandwidth is allocated to the user terminal.

  FIG. 5 is a flowchart illustrating the flow of the hardware performance measurement process of the communication device in the hardware performance measurement. As shown in FIG. 5, the system waits until communication preparations are made (step S301). In the communication preparation stage, first, the transmission / reception size and sampling time are extracted from the user terminal, the standard band is extracted from the communication device, and added to the hardware performance list (step S302). Next, it is determined whether or not the NIC has been previously communicated and the NIC band has been calculated (step S303). If the communication apparatus has been used before, the hardware performance If the NIC band is included in the list, the process proceeds to step S311. If the NIC band is not included in the hardware performance list, it is determined whether communication is in progress (step S304). If not yet communicated, the standard bandwidth is passed to the monitor (step S305), and the start of communication is awaited (step S306). During communication, the communication time is always measured (step S307).

  It is determined whether or not the sampling time has elapsed from the start of communication (step S308), and if it is the recalculation time of the hardware performance of the communication device, the communication time is added to the list (step S309), After calculating the NIC band (step S310), the NIC band is compared with the previous value of the NIC band, the larger one is compared with the standard band, and the smaller one is passed to the monitor as the communication device hardware performance ( Step S311). Then, the current value of the transmission / reception size is added to the hardware performance list as the cumulative transmission / reception size (step S312). If it is not the recalculation time in step S308, the process proceeds to step S313. If the communication is completed before the sampling time has elapsed (step S313), the communication device hardware performance measurement process is terminated. If communication is not completed in step S313, the process returns to step S307.

  Note that once the processing bandwidth and NIC bandwidth are obtained, if the software and hardware conditions do not change even at the next connection opportunity, it is not the bandwidth that requires only the initial value at the first communication. By comparing the values described in the hardware performance list with the initial values, an appropriate bandwidth can be requested from the time of the first communication. Also, for example, up to the previous time, if a video playback application and a 100 Mbps wired network communication device were used, but this time the application does not change, but when trying to use a wireless LAN communication device, The software performance can use the processing bandwidth of the previously used video playback application. That is, a value can be used independently for each of software and hardware. In this way, by measuring the performance of software and hardware during communication, even if the user terminal performs some processing other than communication and each parameter fluctuates, an appropriate value can be corrected each time. In addition, when the hardware configuration of the user terminal, the application to be used, or the communication device is changed, the list can be newly added to the list by extracting the parameters again.

(D) Processing by User Terminal Monitor The user terminal monitor determines the bottleneck among the measured software performance, user terminal hardware performance, and communication device hardware performance values. The lowest value among the above three values is the bottleneck in the user system. Even if a bandwidth exceeding that value is allocated, the given bandwidth is used for the entire user system due to the performance limitations of the software and hardware. It will not be possible.

  FIG. 6 is a flowchart showing a flow of processing in which the monitor shown in FIG. 1 outputs the requested bandwidth to the communication device. As shown in FIG. 6, the system waits for communication preparation (step S401). When the communication preparation is completed, the bandwidths passed from the software, the user terminal, and the communication device are compared, and the smallest value among them is notified to the communication device as the bandwidth required for the network (step S402). The communication apparatus notifies the management server of the requested bandwidth value (step S403). During communication, after waiting for the sampling time and when it is time to recalculate the requested bandwidth (step S404), the bandwidths passed from the software, user terminal, and communication device are compared, and the smallest value is determined by the user system. Is notified to the communication device as a bandwidth that can be processed (step S405). The communication device notifies the management server of the requested bandwidth value (step S406). When the communication is finished, the software performance measurement process is finished (step S407). In this way, the monitor of the user terminal can investigate a value that becomes a bottleneck in communication.

(E) Processing Bandwidth Calculation Processing in Software Performance Measurement FIG. 7 is a flowchart showing the flow of processing (step S110) for calculating the processing bandwidth of the application in FIG. In FIG. 7, after the process of step S109 of FIG. 3, first, an application processing band is calculated (step S501). Next, the CPU usage rate [%] of the current application and the usage rate [%] of the entire CPU are measured, and it is determined whether or not the total value has reached 100 [%] (step S502). If the total value does not reach 100 [%], the next application usage rate is calculated (step S503), and the application processing bandwidth is recalculated so that the application processing bandwidth increases corresponding to the increase in the CPU usage rate. (Step S504). After the recalculation or when the total value has not reached 100 [%] in step S502, it is investigated whether there is a buffer (step S505). If there is a buffer, the buffer processing bandwidth is calculated (step S506). If there is no buffer, the buffer processing bandwidth is set to 0 (step S507). Finally, the processing bandwidth is calculated (step S508), and the value is passed to step S111 in FIG.

(F) Bandwidth allocation processing by management server Next, the operation of the management server will be described. The management server sets the sampling period when the first request bandwidth is received from the monitor of the user terminal via the communication device. The actual performance of each user terminal is measured at each sampling period, and the bandwidth is reallocated. Further, when it is found that the bandwidth becomes insufficient when the user terminal newly requests connection during communication, it is preferable to remeasure the performance without waiting for the sampling time. Thereby, more flexible bandwidth control is possible. As for the sampling period, the sampling period once set is not changed for convenience in this example. However, if the sampling period is dynamically changed depending on the situation, more effective band allocation is possible.

  If there is only one user terminal connected to the network, the sampling period may be long. This is because if only one user terminal uses the bandwidth of the network, it is not necessary to quickly reallocate the bandwidth according to the actually measured processing performance even if it is somewhat inefficient. However, when a large number of user terminals are connected to increase the required network bandwidth, the sampling period of each user terminal is shortened, and the effective processing performance of each user terminal is immediately measured to improve efficiency. It is necessary to perform good bandwidth allocation. As described above, the management server can dynamically allocate a bandwidth to each terminal so that more user terminals can be connected.

  FIG. 8 is a flowchart illustrating a flow of processing in which the management server measures the performance of the user terminal and performs bandwidth allocation and reallocation. In FIG. 8, the management server first confirms a connection request from a new user terminal (step S601). If there is no request, the process proceeds to the process of determining whether there is a connected user terminal (step S611). If there is a request, the requested bandwidth of the new user terminal is measured (step S602), and it is determined whether bandwidth can be allocated to the new user terminal (step S603). If the assignment is possible, the process proceeds to the assignment process (step S609). If assignment is not possible, the presence of another connected user terminal is confirmed (step S604). If no other user terminal is connected, the connection of the new user terminal is rejected because there is no available bandwidth. (Step S610). If another user is connected, the actual performance of all connected user terminals is measured even if it is not the sampling time (step S605). If a user terminal that can reduce the bandwidth is not found at this time, the connection of the new user terminal is rejected (step S610). If a user terminal that can reduce the bandwidth is found, the bandwidth of the user terminal is reduced. (Step S607) It is determined again whether or not it is possible to allocate a bandwidth to the new user terminal (Step S608). If the allocation is possible, the bandwidth is allocated to the new user terminal, and the sampling time is set (step S609). If the assignment is not possible, the connection of the new user terminal is rejected (step S610).

  After that, the management server observes the connected user terminal (step S611), and when there is a connected user terminal, determines whether or not the band reallocation time has come (step S612), and the reallocation time Is reached, the actual performance of the user terminal at the sampling time is measured (step S613), and the bandwidth of the user terminal is reassigned (step S614). Subsequently, it is determined whether or not communication has been completed for all user terminals (step S615). If completed, the operation of the management server is terminated. If not completed, the process returns to step S601 to wait for a new connection request again.

  As described above, the management server performs bandwidth allocation suitable for the user system. As a result, when the user terminal and the communication device are lower than the bandwidth processing performance of the entire network, it becomes possible to efficiently allocate the network bandwidth to the user terminal, and as a result, a larger number of user terminals can be connected. Become. Also, the above does not describe the bandwidth control on the network, but the bandwidth allocation value determined as described above is lower than the lowest bandwidth value in the entire network on the route from the user terminal to the management server. It is clear that the bandwidth value only needs to be requested to the network. This is because it is useless to allocate a bandwidth larger than the bandwidth of the entire network to the user terminal.

  FIG. 9 is a diagram illustrating a procedure in which the management server assigns a bandwidth to one user terminal. 9A to 9C show a user terminal having a configuration similar to that shown in FIG. 1 (some components are omitted). As shown in FIG. 9A, first, before starting communication, the monitor scans the user terminal to determine the required bandwidth. Now, a user terminal application is requesting 4 Mbps video playback. At this time, the vacant bandwidth of the bus is 16 Mbps, the standard value of the communication device is 11 Mbps, and both are larger than 4 Mbps. Therefore, the monitor determines the bandwidth of 4 Mbps as the requested bandwidth, and transmits this value to the management server. The management server allocates 4 Mbps to the user terminal.

  Thereafter, as shown in FIG. 9B, when the sampling period is actually observed to some extent, it is assumed that the application is actually only capable of processing 1.5 Mbps. This is because the CPU is processing other than the above application and the usage rate reaches 100%, and the processing capacity of the application is suppressed to 1.5 Mbps. This situation is monitored by the monitor, and the management server scans the monitor to detect that the 4 Mbps bandwidth allocated to this user terminal is not being used effectively, and to allocate the bandwidth allocated to the user terminal. To reduce. At this time, since the effective value of the communication device is 4 Mbps and the vacant bandwidth of the bus band is 12 Mbps, the bandwidth value assigned to the user terminal is 1.5 Mbps.

  After that, as shown in FIG. 9C, when the CPU processing load is increased and the processing capacity of the application is increased again, the monitor displays the CPU usage rate of the application and the overall CPU usage rate. In consideration of the above, the management server is requested to increase the allocated bandwidth. In the case of FIG. 9C, the CPU usage rate of the application can be increased from the current 20% to 40%. Then, it becomes possible to double the processing capacity that is currently 1.5 Mbps. Therefore, the monitor requests a bandwidth of 3 Mpbs from the management server.

  FIG. 10 is a diagram illustrating a procedure in which the management server performs bandwidth allocation control for a plurality of user terminals. As shown in FIG. 10A, it is assumed that user terminal A and user terminal B are connected to a network, and the management server allocates 4 Mbps and 5 Mbps bandwidths to the respective user terminals. At this time, the free bandwidth of the network is 1 Mbps. The management server measures the effective processing performance of both user terminals and reallocates the bandwidth each time a specified sampling time comes.

  Here, as shown in FIG. 10B, when the user terminal C newly makes a connection request, first, the management server monitors the request from the user terminal C and tries to secure a bandwidth of 3 Mbps. However, since the free bandwidth of the network is only 1 Mbps, the requested bandwidth cannot be allocated to the user terminal C as it is. Therefore, the management server monitors each of the user terminals A and B again. At this time, the management server preferably performs re-monitoring even if the sampling time has not come. As a result, it is assumed that the user terminal A uses the allocated 4 Mbps bandwidth as shown in the figure, but the user terminal B is found to use only 3 Mbps of the allocated 5 Mbps.

  In this case, as shown in FIG. 10C, the management server reallocates the 3 Mbps bandwidth to reduce the bandwidth allocated to the user terminal B, and the bandwidth released from the user terminal B. Therefore, the total available bandwidth of the network is 3 Mbps, so that the requested bandwidth of 3 Mbps can be allocated to the user terminal C.

  FIG. 11 is a diagram showing a comparison between a conventional technique for allocating a bandwidth from a management server to a user terminal and the system of the present invention. In the general method in the prior art shown in FIG. 11A, the allocated bandwidth [bps] to the user terminal based on the application 60, the user requested bandwidth [bps] 61, and the user terminal requested bandwidth [bps] 62. ] 63 is determined. However, the bandwidth [bps] 64 of the network path is not considered. In the method of Patent Document 1 shown in FIG. 11B, an application 70, a user request bandwidth [bps] 71, a communication device processing performance [bps] 72, a user terminal request bandwidth [bps] 73, Based on the bandwidth [bps] 74 of the network path, the allocated bandwidth [bps] 75 for the user terminal is determined. However, the effective performance [bps] 76 of the user terminal is not taken into consideration, and there is a case where excessive bandwidth allocation is performed to the user terminal. On the other hand, in the method of the present invention shown in FIG. 11C, the application 80, the requested bandwidth [bps] 81 of the user, the communication device [bps] 82, the effective performance [bps] 83 of the user terminal, The allocated bandwidth [bps] 86 is determined based on the requested bandwidth [bps] 84 of the user terminal and the network route [bps] 85.

Application Example of the System Next, an application example of the above-described communication band allocation system of the present embodiment will be described.
[First application example]
Depending on the configuration of the user terminal, there may be an area for temporarily storing data to be transmitted and received, such as a FIFO and a cache, in addition to the buffer managed by the application. Hereinafter, such a temporary storage area for data is referred to as a cache [byte]. By having this cache, even when the processing capacity of the application used by the user or the user system is low, processing can be performed by allocating a bandwidth that is larger by the cache. When the system having this cache is applied to the communication band allocation system of the present embodiment described above, the calculation formula of the hardware processing capability in the user terminal can be expressed as follows.
(Bus width) + (cache) / (communication time)

  At this time, the application may not operate at all while data is stored in the buffer. In this case, the processing bandwidth of the application becomes 0, but it is actually necessary to continue communication. In this case, an additional field is provided in the hardware performance list. For example, when “1” is described here, the processing bandwidth of the application may be ignored to calculate the bandwidth.

  The cache may be realized by a dedicated hardware mechanism in the user system, or may be managed by the application or software such as an OS like a virtual memory. Of course, it may be installed for a user terminal connected to a management server or a communication device.

[Second application example]
In the communication band allocation system of the present embodiment described above, it is assumed that a bandwidth higher than the user's requested bandwidth is not allocated. However, if a powerful CPU is installed and the performance of the application does not become a bottleneck, an allocation is performed. There may be cases where more data can be processed as more bandwidth is allocated. For example, FTP (File Transfer Protocol) software corresponds to this. Since it is possible that the user has set the requested bandwidth without knowing such a situation, if the above situation is applicable, the user's requested bandwidth is ignored and the processing bandwidth is set as the processing performance of the application. Also good.

[Third application example]
In the communication band allocation system of this embodiment described above, the sampling time is set to the same value for both software and hardware, but may be set to different values. For example, the usage status of the CPU by the application does not change much over a long period of time. However, when data is exchanged intermittently, the software sampling time is increased and the hardware sampling time is decreased. Thus, efficient sampling adapted to the operation mode of the system can be performed.

[Fourth application example]
In the communication bandwidth allocation system of the present embodiment described above, when the sampling time elapses, various processing capacities are recalculated, and the management server reallocates the bandwidth to the user terminal each time. The bandwidth may be allocated only when the above changes. In this case, calculation is performed based on the formula of (current request bandwidth) / (previous request bandwidth), which is a coefficient a, and the value of a is within a predetermined allowable range (for example, 0.7 <a < It is only necessary to reallocate the bandwidth only when it exceeds (range 1.3). In this application example, band reallocation is performed only when the effective performance of the user system exceeds or exceeds a predetermined amount, and the minute performance change of the user system is excluded to some extent or completely. is there. As a result, for example, in a system in which communication is temporarily interrupted when switching the bandwidth allocation, the effective performance of the user system is allocated once even if the effective performance of the user system is somewhat lower than the frequency bandwidth reassignment processing. As a result, it is possible to improve the processing efficiency.

  As described above, according to the communication band allocation system of the present embodiment, the management system can actually process the user system by allocating the band in consideration of the application used by the user and the actual processing performance of the user terminal and the communication device. Therefore, it is possible to prevent the allocation of an excessive communication band, and thus more efficient band allocation is realized. In addition, by appropriately monitoring the status of software and hardware in the user system and dynamically reflecting it in the allocation of bandwidth, appropriate bandwidth allocation according to the change in the status in the user system is realized.

  While the communication band allocation system of the present invention has been described with reference to specific embodiments, the present invention is not limited to these. A person skilled in the art can make various changes and improvements to the configurations and functions of the invention according to the above-described embodiments or other embodiments without departing from the gist of the present invention.

  The communication band allocation system of the present invention can be configured from an information processing device having a communication function, a mobile terminal, a mobile phone, and the like. In addition, as a program for causing a computer to execute the communication band allocating method of the present invention, it can be applied to an information processing apparatus having a communication function, a mobile terminal, a mobile phone, and the like.

It is a figure which shows roughly the internal structure of the information processing apparatus provided with the communication function which comprises the user terminal in the communication band allocation system of this invention, and the structure of the network to which this information processing apparatus is connected. (A) is a figure which shows the outline | summary of the system containing a user terminal, (b) is a figure which shows the example of schematic structure of the monitor 3 for monitoring. It is a figure which shows the data structural example of the software performance list | wrist stored in the monitor shown in FIG. 1, and hardware performance. It is a flowchart figure which shows the flow of a measurement process of software performance. It is a flowchart which shows the flow of a measurement process of the hardware performance of a user terminal in the case of a hardware performance measurement. It is a flowchart which shows the flow of the measurement process of the hardware performance of the communication apparatus in the case of a hardware performance measurement. It is a flowchart which shows the flow of the process which the monitor shown in FIG. 1 outputs a request zone | band to a communication apparatus. 4 is a flowchart showing a flow of processing (step S110) for calculating a processing bandwidth of an application in FIG. It is a flowchart which shows the flow of the process in which a management server measures the performance of a user terminal, and allocates and reallocates a band. It is a figure which shows the procedure in which a management server allocates a band to one user terminal. It is a figure which shows the procedure in which a management server controls the bandwidth allocation with respect to several user terminals. It is the figure which compared and showed the system by the prior art which allocates a band with respect to a user terminal from a management server, and the system of this invention.

Explanation of symbols

2 Communication Device 3 Monitor 4 Display Device 5 Relay Device 6-1 Management Server 7 Software Performance List 8 Hardware Performance List 9 Comparator (Request Bandwidth Calculator)
10 Application 11 Required bandwidth [bps]
12 Total IO size [byte]
13 Cumulative IO size [byte]
14 Total buffer size [byte]
15 Cumulative buffer size [byte]
16 CPU usage rate of application before communication [%]
17 CPU usage [%] of application during communication
18 Total CPU usage before communication [%]
19 CPU usage rate during communication [%]
20 Processing time [s]
21 Processing bandwidth [bps] The value in parentheses is the previous value. 22 Sampling cycle [s]
30 Communication device 31 Bus width [bps]
32 Standard bandwidth [bps]
33 Send / receive size [byte]
34 Total send / receive size [byte]
35 Communication time [s]
36 NIC bandwidth [bps] The value in parentheses is the previous value 37 Sampling time [s]

Claims (15)

A system for allocating communication bandwidth to user terminals connected to a network,
Communication that measures at least one of an effective value of hardware processing performance, an effective value of processing performance of software, and an effective value of processing performance of a communication device in the user terminal, and is assigned to the user terminal based on the measured value A communication band allocation system characterized by determining a band.   The effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the communication device processing performance are expressed as the maximum amount of data that can be actually processed per unit time by the hardware, software, and communication device, respectively. The communication band allocation system according to claim 1, wherein:   The effective value of the processing performance of the hardware is calculated using at least one of the data processing performance of the CPU, ROM, RAM, and hard disk of the user terminal. Communication bandwidth allocation system.   4. The communication band allocation system according to claim 3, wherein a calculation method of an effective value of processing performance of the hardware is changed according to a change in a hardware configuration in the user terminal.   5. The communication according to claim 1, wherein the effective value of the processing performance of the software is calculated using a hardware parameter related to the software presented by an OS of the user terminal. Bandwidth allocation system.   The hardware parameters related to the software presented by the OS of the user terminal include the amount of data written to and read from the IO device by the software, the CPU usage rate of the software, the CPU usage rate of processes other than the software, the software 6. The communication band allocation system according to claim 1, further comprising at least one of a communication band required by the host and a buffer capacity capable of temporarily storing data.   The communication band allocation system according to claim 5 or 6, wherein a method for calculating an effective value of the processing performance of the software is changed according to a type of information to be processed by the software.   The effective value of the processing performance of the communication device is calculated by using a hardware parameter related to the communication device presented by an OS of the user terminal. Communication bandwidth allocation system.   The hardware parameter related to the communication device presented by the OS of the user terminal includes at least one of the amount of data transmitted and received by the communication device and the maximum value of the processing performance according to the standard of the communication device. The communication band allocation system according to any one of claims 1 to 8, wherein the communication band allocation system is characterized in that:   The communication band allocation system according to claim 8 or 9, wherein a method for calculating an effective value of processing performance of the communication device is changed according to a communication method standard of the communication device.   The user terminal continuously measures at least one of the effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the processing performance of the communication device in the user terminal. The communication band allocation system according to any one of claims 1 to 10, wherein a communication band allocated to the user terminal is dynamically changed based on a value.   At least one measurement value among the effective value of the hardware processing performance in the user terminal, the effective value of the software processing performance, and the effective value of the processing performance of the communication apparatus is sampled at a predetermined time interval and assigned to the user terminal. The communication band allocation system according to claim 11, wherein the communication band is dynamically changed.   Based on a comparison result between at least one measurement value of the effective value of the hardware processing performance, the effective value of the processing performance of the software, and the effective value of the processing performance of the communication device in the user terminal and the previous measurement value. The communication band allocation system according to claim 11 or 12, wherein a communication band allocated to the user terminal is dynamically changed.   When a connection request to the network is received from a new user terminal, the sampling is performed on all user terminals currently connected to the network, and it is confirmed whether there is sufficient communication bandwidth to be allocated to the new user terminal. The communication band allocation system according to claim 12, wherein:   It should be assigned to the user terminal determined based on at least one measurement value among the effective value of the hardware processing performance, the effective value of the software processing performance, and the effective value of the processing performance of the communication device in the user terminal 2. The communication band is allocated to the user terminal only when the communication band is a value that does not exceed the communication band of the route on the network through which the user terminal communicates. 14. The information device according to any one of 14.

Images