CN100531180C - Request scheduling system and method for providing difference service quality on network relay point - Google Patents

Request scheduling system and method for providing difference service quality on network relay point Download PDF

Info

Publication number
CN100531180C
CN100531180C CN 200510059615 CN200510059615A CN100531180C CN 100531180 C CN100531180 C CN 100531180C CN 200510059615 CN200510059615 CN 200510059615 CN 200510059615 A CN200510059615 A CN 200510059615A CN 100531180 C CN100531180 C CN 100531180C
Authority
CN
Grant status
Grant
Patent type
Prior art keywords
queue
request
external
response
quality
Prior art date
Application number
CN 200510059615
Other languages
Chinese (zh)
Other versions
CN1842072A (en )
Inventor
李卓俊
林盈达
温硕彦
田庆明
赖源正
郑博文
Original Assignee
财团法人工业技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Grant date

Links

Abstract

本发明涉及一种在网络中继点上提供差别服务品质的外部请求排程系统,执行于介在一个以上客户终端与应用服务器之间的网络中继点上的计算机可执行程序,包含有:探查器,用以探查该应用服务器中一个以上请求项目及其对应的响应大小,并记录于请求/响应表中;分类器,用以建立服务品质策略表,并负责接收来自各该客户终端一个以上外部请求进行外部请求分类;队列器,根据该服务品质策略表配置一个以上队列,用以根据分类结果暂存该外部请求;及排程器,用以设定各该队列对应的权重值,以及队列指针及排程窗户值,并根据各该外部请求的该响应大小以循环方式开始进行各该队列中各该外部请求的传送处理。 The present invention relates to a method of providing an external differential service quality request scheduling system relay point in the network, via computer executable program to execute on a network relay point between one or more client terminals and application servers, comprising: probe , a probe to the application server in response to a request size than its corresponding item, and recorded in the request / response table; classifier for establishing a quality of service policy table, and for receiving from each of the more than one client terminal external requests external request classification; queue is configured based on the quality of service policy table than a queue for temporarily storing the classification result according to an external request; and a scheduler for setting the weight value corresponding to each of the queue, and and schedule window queue pointer value, and in a cyclic manner each of the transfer processing is started in each of the external request queue based on the size of each of the response to an external request.

Description

在网络中继点上提供差别服务品质的请求排程系统及方法 Provide differentiated quality of service request scheduling system and method for a relay point in the network

技术领域 FIELD

本发明涉及一种应用于互联网服务的外部请求排程系统及其方法,特别是涉及一种位于网络中继点利用请求分类、回应大小及窗式控制来进行外部请求排程的系统及其方法。 The present invention relates to an external request used in the Internet service scheduling system and method, particularly to a network relay point located using the classification request, response size, and window control system and method for scheduling external request .

背景技术 Background technique

随着互联网环境中各种应用服务的蓬勃发展,使得应用服务器的使用量已经越来越大,再加上互联网使用人口数量的与日俱增,目前几乎所有的应用服务器的负载量也越来越大。 With the rapid development of the Internet environment in a variety of applications and services, making use of the application server it has been growing, coupled with the increasing use of the Internet population, almost all of the currently loading application server is also growing.

事实上,当应用服务器的负载量过大时,就容易造成位于客户终端的使用者等待服务的时间变长,'严重的话还可能导致应用服务器的服务完全瘫痪。 In fact, when the application server load is too large, it is easy to cause time at the customer's end users waiting for service longer, 'the worst case, may also lead to an application server service completely paralyzed. 因此,许多应用服务器便希望能够通过差别服务的方式,让具有较高等级的使用 Therefore, many application servers they want to be able through differentiated services, so using a higher level

者可以获得到较多的应用服务器吞吐量(Throughput),从而縮短这些使用者通过客户终端发出外部请求后所需等待服务的时间,让使用者可以获得较好的服务品质,提高这些使用者的满意度。 Can get to more application server throughput (Throughput), thus shortening the time required to issue these users waiting for service after a request by the external client terminal, so that users can get a better quality of service, enhance these users satisfaction.

而目前与差别服务相关的现有技术内容,多是直接在应用服务器上进行更改,通过修改其操作系统或者是另外提供排程程序的方式来实现对使用者的差别服务,然而此种做法的缺点在于必须耗费应用服务器的资源来处理排程问题,并且与应用服务器的相依性过大,反而容易造成应用服务器稳定性不足以及移植性不佳的问题。 At present, the content of the prior art related to the difference in services, mostly to make changes directly on the application server, to achieve user of differentiated services by modifying its operating system or program is another way to provide scheduling, but such practices the disadvantage is that you must consume resources of the application server to handle scheduling problems, and the dependencies of the application server is too large, but likely to cause insufficient application server stability and poor portability issues.

其它现有技术,如美国专利US006742016B1,其所提供的差别服务技术内容为在应用服务器上设置一个请求接收器(Acc印tor),然后再利用其中的优先权控制模块(Priority Control Module Process)来控制外部请求传送至应用服务器上的顺序,其中优先权控制模块会将所有来自客户终端的外部请求分为两个等级,其决定传送顺序的依据有两个:(l)较高等级的外部请求数量不小于事先定义好的传送数量时,则传送较高等级中与定义传送数量相同的外部请 Other prior art, as described in U.S. Patent No. US006742016B1, differentiated services they provide to the technical content provided on the server application a request receiver (Acc printing Tor), then use the control module where the priority (Priority Control Module Process) to the external request order control is transferred to the application server, wherein the control module sets the priority of all external requests from the client terminal is divided into two levels, which is determined according to the transmission order are two: external request (l) a higher level of the number is not less than pre-defined number of good transmission, is transmitted with the higher level transmit the same number of defined outer please

5求至应用服务器;(2)较高等级的外部请求数量小于事先定义好的传送数量时, 则将较高等级与较低等级的外部请求混合传送至应用服务器,传送的数量与定义传送数量相同。 5 request to the application server; external requests (2) a higher level when the number is less than the predefined number of good transmission, then the higher level and the lower level of the external hybrid delivery request to the application server, the number of transfer defined number of transmission the same. 此种做法,主要是将所有的外部请求依照优先权的不同而分为两大类,特别强调具有较高等级的外部请求所获得应用服务器处理的数量一定比具有较低等级的外部请求多,借此达到差别服务的目的。 Such an approach, mainly to all external requests according to different priorities and divided into two categories, with particular emphasis on the number of application server processes the request with an external higher level than obtained in some external request with a lower rank, thereby achieve the purpose of different services. 虽然此种做法可以保证不同等级的外部请求被处理的个数不同,但并不表示具有较高等级的使用者可以获得较好的服务品质,其主要的原因在于每个外部请求对于应用服务器所耗费的资源并不相同,即每个外部请求的回应大小均不同,因此处理较多较高等级的外部请求数量并不表示应用服务器可以提供较多资源给这些较高等级的外部请求,所以拥有较高等级的使用者并不能够享受到较佳的服务品质,也即使用者所感受到的服务等待时间并不一定能够被縮短。 While such an approach ensures that the number of different levels of different external request being processed, but does not mean that the user has a high level may obtain better quality of service, the main reason is that each request for an external application server cost of resources is not the same, that is, each response size external requests are different, so the number of requests processed more external higher level does not mean that the application server can provide more resources to these higher levels of external requests, therefore, have a higher level of user is not able to enjoy a better quality of service, that is experienced by the user of the service does not necessarily waiting time can be shortened.

至于美国专利US006728748B1,则利用一台路由伺服主机(Routing Host) 来接收所有客户终端所传送来的外部请求,然后将这些外部请求作外部请求分类,再通过连接于路由主机与多台应用服务器之间的适应性策略引擎(Adaptive Policy Engine)来与路由主机和各个应用服务器中所设置的代理器进行沟通,利用代理器所传回的各种应用服务器属性信息,来决定哪种类别的外部请求该分配给哪台应用服务器处理。 As U.S. Patent No. US006728748B1, the use of a routing server host (Routing Host) to receive external requests for all client terminal transmitted, and then these external requests for external request classification, and then connected to the routing host and multiple application servers by external requests adaptive policy engine (Adaptive policy engine) to communicate between the host and routing each application server in the proxy settings, using the agent returned by a variety of application servers attribute information to decide which category which server processes the application is assigned to. 此种做法,则是采取利用应用服务器丛集(Server Cluster)的方式来解决差别服务的问题,除了对应用服务器作更改之外还必须花费大量的服务器硬件成本才能够实现差别服务。 Such an approach, it is a way to take advantage of application server cluster (Server Cluster) to solve the problem of different services, in addition to making changes to the application server must also spend a lot of server hardware costs can be achieved differentiated services.

为了提供更好的差别服务品质,以及减少因为差别服务所产生对应用服务器在运作上的影响,并且能够增加差别服务机制在应用上的兼容性,确实有必要对目前既有差别服务技术作改善,以早日实现互联网环境中真正的差别服务机制。 In order to provide better service quality differences, as well as reducing the impact on the application server in the operation of the resulting difference in service, and to increase the compatibility differences in the application of the service mechanism, it is necessary to do both currently serve different techniques for improvement to the early realization of the Internet environment real difference service mechanism.

发明内容 SUMMARY

本发明所要解决的技术问题在于提供一种应用于互联网环境的外部请求排程系统及其方法,将排程机制完整实现于介于客户终端以及应用服务器之间的网络中继点,因此可以在不更改客户终端以及应用服务器的前提之下,遵循既有的互联网服务模式来实现真正的差别服务。 The present invention is to provide a technical problem to be applied to the external Internet environment system and method for scheduling request, the scheduling mechanism implemented in a complete network between the relay point between the client terminal and an application server, and therefore may be under no changes to the client terminal and the application server under the premise of follow existing Internet service model to achieve true differentiated services.

对于差别服务中最重要的外部一请求分类机制,本发明除了采取依据外部请求所包含的网络层内容(如:来源端位址、目的端位址、端口号及协议类型) 来进行外部请求分类之外,更具有依据外部请求所包含的应用层内容(标头信息及内容)来进行外部请求分类,可以增加差别服务的应用层面,使差别服务应用更多元化。 The most important difference with respect to a request for an external service classification mechanism, in addition to the present invention is based upon external request to take the content layer included in the network (example: the source end address, destination address, protocol type and port numbers) to external requests classified addition, more according to application layer content (content and header information) contained in the external request external request classification, the difference can be increased service application layer, the service application so that the difference is more diversified.

另外本发明更进一步以权重值设定来提供不同应用服务器服务资源的技术手段来作为不同等级的外部请求的差别服务方式,除了通过权重值与外部请求所对应的回应大小的控制来决定外部请求的处理顺序之外,还可以通过权重值让具有较高等级的使用者平均享有比较低等级的使用者高的应用服务器资源,同时也縮短平均等待服务的时间,也即让使用者实际获得到较高的服务品质。 Further the present invention still further to a weight value is set to provide the technical means different application server service resources as different approaches in different levels of external requests, in addition to the control by a weight value to the external request corresponding to the response size to determine the external request outside the processing sequence, it can also allow users to weight values ​​with a higher level of average users enjoy a high level of application server resources are relatively low, but also shorten the average waiting time of service, that allows users to get real high service quality.

为了避免大量的外部请求传送对应用服务器造成过大的处理负担,本发明还提出利用设定排程窗户值的方式以窗式控制机制来控制已排程好的外部请求的传送时间点,所有已分类未传送处理的外部请求,都将被暂存在队列中等待允许传送,如此将可以确保应用服务器的运作维持在最佳的状态,提高整体排程运作的稳定性。 In order to avoid a large number of external transfer requests cause excessive processing burden on the application server, using the embodiment of the present invention also provides schedule setting window values ​​to control a window mechanism to control transmission of the scheduled time points good external requests, all classified not transmitted external request process, will be temporarily stored in the queue for allowed transmission, thus ensuring the operation of the application server is maintained at an optimum state, improve the stability of the overall operation of the schedule.

本发明所公开的在网络中继点上提供差别服务品质的外部请求排程系统, 该系统设置于介于一个以上客户终端与一应用服务器之间的网络中继点上,包含有: 一探査器,用以探查该应用服务器中一个以上请求项目及其对应的回应大小,并记录于一请求/回应表中,该回应大小是该应用服务器在执行外部请求所指定的请求项目时所需传送的实际字节数量; 一分类器,用以建立一服务品质策略表,并负责接收来自各该客户终端一个以上外部请求进行一外部请求分类; 一队列器,根据该服务品质策略表配置一个以上队列,用以根据分类结果暂存该外部请求;及一排程器,用以设定各该队列对应的一权重值,以及一队列指针及一排程窗户值,并根据各该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理,其中,该排程窗户值 Provide differentiated service quality on a network relay point outside the claimed invention as disclosed scheduling system, the system is provided on a network relay point between a terminal and a server interposed between more than one client application, comprising: a probe , a probe to the desired application server when transmitting a response to the above request and the corresponding size of the project, and recorded in a request / response table, the size of the response is performed outside the application server request specified requested items the actual number of bytes; a classifier for establishing a quality of service policy table, and the client is responsible for receiving from each terminal a request for a more external classification external request; a queue, a configuration based on the more than one quality of service policy table queue for temporarily storing the classification result according to the external request; and a scheduler for setting each of the queues corresponding to a weight value, and a scheduling queue pointer and a window value, and in accordance with the request of each of the outside the response size in a cyclic manner each of the transfer processing is started in each of the external request queue, wherein the schedule window value 示同一时间内可传送给该应用服务器处理的外部请求的数量。 The number may be transmitted to an external application server process requests in the same time shown.

本发明所公开的在网络中继点上提供差别服务品质的外部请求排程方法, 执行于介于一个以上客户终端与一应用服务器之间的网络中继点上,包含有: 探查该应用服务器中一个以上请求项目及其对应的回应大小,并记录于一请求/回应表中,该回应大小是该应用服务器在执行外部请求所指定的请求项目时所需传送的实际字节数量;建立一服务品质策略表,开始接收来自各该客户终端一个以上外部请求进行一外部请求分类,并根据分类结果暂存各该外部请求至对应一个以上队列;及设定各该队列的一权重值,以及一队列指针及一排程 The present invention disclosed herein provide different quality of service on a network relay point external request scheduling method executed in more than one client terminal between the relay point between a network application server, comprising: the probe application server in response to a request size than its corresponding item, and recorded in a request / response table, the size of the response is the actual number of bytes transferred in the desired application execution server specified external request requested items; establishing a quality of service policy table, the client begins to receive from each terminal a request for an external more external classification request, and temporarily stores each of the external request based on the classification result corresponds to more than one queue; and setting a weight value of each of the queues, and a scheduling queue pointer and a

窗户值,并根据各该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理,其中,该排程窗户值表示同一时间内可传送给该应用服务器处理的外部请求的数量。 Window value, and in a circular manner starts transfer processing each of the queue of each of the external request based on the response to the size of each of the external request, wherein the scheduling window value represents an external may be transmitted to the same time of the application server process the number of requests.

本发明的功效,在于可以在不更改客户终端以及应用服务器的前提之下, 遵循既有的互联网服务模式来实现真正的差别服务。 Efficacy of the present invention is that it can not change under the client terminal and the application server under the premise of follow existing Internet service model to achieve true differentiated services.

以下结合附图和具体实施例对本发明进行详细描述,但不作为对本发明的限定。 The present invention will be described in detail in conjunction with the accompanying drawings and the specific embodiments, but not limit the present invention. 附图说明 BRIEF DESCRIPTION

图11A为本发明在网络中继点上提供差别服务品质的外部请求排程系统的系统架构图; FIG. 11A outside the present invention provides a differential service quality of the system architecture of FIG request scheduling system on a network relay point;

图1B为本发明在网络中继点上提供差别服务品质的外部请求排程系统的服务品质策略表原始码示意图; Figure 1B a schematic view of the present disclosure provide differentiated service quality source of a relay point in the network quality of service policy table external request scheduling system;

图1C为本发明在网络中继点上提供差别服务品质的外部请求排程系统的队列器与排程器初始状态示意图; Figure 1C a schematic view of the present invention to provide a scheduler queue with different quality of service of the initial state of the external request scheduling system on a network relay point;

图1D为本发明在网络中继点上提供差别服务品质的外部请求排程系统的外部请求处理与排程窗户值运作示意图; FIG 1D of the present invention to provide different quality of service on a network relay point external request the value of the external operating system scheduling request processing and scheduling a schematic diagram window;

图2A为本发明在网络中继点上提供差别服务品质的外部请求排程方法的流程图;及 Figure 2A flowchart scheduling method of the present invention to provide differentiated service quality on a network relay point external request; and

图2B及图2C为本发明在网络中继点上提供差别服务品质的外部请求排程方法的细部流程图。 2B and 2C detailed flow chart of the scheduling method of providing quality of service in different network relay point external requests present invention.

其中,附图标记: Wherein, reference numerals:

10 20 31 33 35 1020313335

客户终端外部请求 External client terminal requests

分类器队列器 Category queue is

351A、 351B、 351C 352 第二队列 351A, 351B, 351C 352 second queue

15 30 32 34 351 15303234351

第一队列的外部请求352A、 352B、 352C Outside of the first request queue 352A, 352B, 352C

互联网 the Internet

外部请求排程系统请求/回应表服务品质策略表第一队列 External request scheduling system request / response to the first queue table service quality policy table

第二队列的外部请求353 第三队列 Outside of the second request queue third queue 353

353A、 353B、 353C 第 353A, 353B, 353C first

36 排程器 Scheduler 36

361 第一队列计数器 First queue counter 361

362 第二队列计数器 A second queue counter 362

363 第三队列计数器 363 Third queue counter

364 队列指针 Queue pointer 364

365 排程窗户值 365 Scheduling window value

40 应用服务器 40 Application Server

50 原始码 50 source

51 分类项目 51 Classification Project

52 分类规则 52 classification rules

步骤ioo-探査一应用服务器中一个以上请求项目及其对应的回应大小, 并记录于一请求/回应表中 Step ioo- a probe request to an application server to respond to more than one item and the corresponding size, and recorded in a request / response table

步骤200-建立一服务品质策略表,开始接收一个以上外部请求进行一外部请求分类,并根据分类结果暂存各该外部请求至对应一个以上队列 Step 200- establishing a quality of service policy table receiving one or more external starts an external request classification request, and temporarily stores each of the external request queue according to a corresponding one or more classification results

步骤300-设定各该队列的一权重值,以及一队列指针及一排程窗户值, 并根据各,外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理 Step 300 to set a weight value of each of the queues, and a queue pointer to the window and a scheduled value, and accordingly each of the outer size of the response requests in a cyclic manner each of the transfer processing is started in each of the external request queue is

步骤301-设定该排程窗户值,初始化该队列指针使其指向最先该队列, 逐一扫瞄所有队列,将有至少一个外部请求的队列加入有效队列列表步骤302-有效队列列表中是否有至少一个队列? Step 301 sets the schedule window value, initializing the queue pointer to point to the first queue, the queue one by one all the scanning, there will be a list of valid queue if there is at least one external request queue is added in step 302 the list of valid queue at least one queue?

步骤303-初始化该队列指针使其指向有效队列列表中最先该队列,增加最先该队列对应的该权重值于最先该队列的一计数器中步骤304-该队列中仍有未处理的该外部请求? Step 303 initializes the queue pointer to point to the first list of valid queue the queue, the queue corresponding to the first increase in the weight value of the first counter step in a queue in the queue 304- the still untreated external requests? 步骤305-将该队列加入有效队列列表,若已经存在则不需重复加入步骤306-读取该队列中该外部请求,并判断该外部请求所对应的该回应 Add step 305 the queue active queue list, need not be repeated if it already exists is added to the outer step 306- read request in the queue, and determines whether the response request corresponding to the outer

大小 size

步骤307-该回应大小^该计数器? The response to the size of the step 307- ^ the counter? 步骤308-该排程窗户值=0?步骤309-该计数器_该回应大小,该排程窗户值-1,并传送该外部请求进行处理 Step 308- the schedule window value = 0? _ Step 309 the response size of the counter, the value -1 scheduling window, and transmits the external request is processed

步骤310-移动该队列指针至下一队列,并增加该队列所对应的该权重值至该队列的计数器中 Step 310- moving the queue pointer to the next queue, and increase the weight of the queue corresponding to the queue to the counter weight value of

步骤311-将该队列从有效队列列表移除步骤312-完成此回合所有队列的排程? Step 311- the queue step removed from the list of valid queue 312- complete this round of all the queues schedule?

具体实施方式 detailed description

本发明是一种应用于互联网15环境中,利用执行于网络中继点上的计算机可执行程序来提供差别服务品质的外部请求排程系统30及其方法。 The present invention is a method used in the Internet environment 15 to provide differential service quality of the external system 30 and method of scheduling request using a computer-executable program executed on a network relay point.

所谓的网络中继点,是指介于一个以上客户终端10与应用服务器40之间的网络节点,如:网关器、路由器、代理服务器及服务器负载平衡器等。 A so-called network relay point refers to more than one client network node interposed between the terminal 10 and the application server 40, such as: a gateway, a router, a proxy server, and server load balancer. 任何客户终端10所发出的外部请求20都必须要通过互联网15,通过此网络中继点才能够被顺利传送至应用服务器40上,然后由应用服务器40提供对应的服务与回应。 Any external client terminal 10 requests sent via the Internet 20 must be 15, through this network can be a relay point was successfully transferred to the application server 40, and then provides a corresponding service by the application server 40 and responses.

图1A为本发明的外部请求排程系统30的系统架构图,由图中可知,外部请求排程系统30包含有:用以探查应用服务器40中所有请求项目及其对应回应大小的探査器31、负责记录前述请求项目以及对应回应大小的请求/回应表32、用以对客户终端10所发出的外部请求20进行外部请求分类的分类器33、 负责储存分类项目51的服务品质策略表34、根据服务品质策略表中分类项目51来配置队列并对应储存分类后的外部请求20的队列器35,以及负责根据外部请求20的回应大小以循环方式来传送各队列中外部请求20至应用服务器40的排程器36等几个部分。 FIG 1A of the present invention an external request scheduling system architecture diagram of the system 30, can be seen from the figure, external request scheduling system 30 includes: an application server 40 to probe all the items and their corresponding request response size profiler 31 , responsible for recording the requests and corresponding request response item size / response table 32, external to the client terminal 10 requests 20 issued a request for an external classification classifier 33, is responsible for service quality classification items stored policy table 51, 34, after configuring a queue according to the quality of service policy table classification item 51 and the corresponding storage classified external request queue 20 35, and is responsible for the request response size 20 according to the outside in a circular manner to transmit in each queue external request 20 to the application server 40 the scheduler 36 sections the like.

图1B为分类器33作为外部请求分类依据的服务品质策略表34的原始码50示意图,原始码50中包含有一个以上的分类项目51,分类项目51是外部请求分类的主要依据,基本上可以依照外部请求20中所包含的网络层内容来订定外部请求的分类规则52,如:来源端位址、目的端位址、端口号及协议类型等,也可以利用外部请求20中所包含的应用层内容来进行分类规则52 的订定,如:标头信息及内容等。 1B is a classifier 33 as an external source of Sort requested quality of service policy table 34 is a schematic diagram 50, the source 50 includes one or more classification item 51, item 51 is the main basis for classification external request classification basically SUMMARY in accordance with an external request the network layer 20 included to provide for classification rules 52 external requests, such as: the source end address, destination address, protocol type, port numbers, and the like, can also be used external requests contained 20 application layer content to provide for the classification rules 52, such as: the content of header information, and the like. 虽然图1B中原始码是采取可延伸标记程序语言(Extensible Markup Language; XML)来撰写,但是本发明并未对此作任 Although FIG. 1B is a source program to take Extensible Markup language (Extensible Markup Language; XML) to write, but the present invention is not for any of this

10何限定,其它程序语言都可以作为可行的实施方案。 10 defines any other programming language can be used as viable embodiments.

图1C为队列器35以及排程器36的初始状态示意图,将再进一步进行说 FIG 1C is a schematic view of the initial state of the queue 35, and scheduler 36, are then further be explained

明。 Bright. ' '

其中,队列器35中配置有一个以上的队列,队列的数量是依照服务品质策略表34中分类项目51而决定的,以图1C为例,表示在服务品质策略表34 中将所有的外部请求20区分为三个不同的分类项目51,.因此在队列器35中会配置三个不同的队列,即第一队列351、第二队列352及第三队列353,用来储存不同的外部请求20。 Wherein, the queue 35 is configured with more than one queue, the queue number in accordance with the quality of service policy table 34 determined classification items 51 to FIG 1C, for example, represents all external requests in the quality of service policy table 34 20 is divided into three different regions 51 ,. Accordingly classification items in the queue 35 will be arranged three different queues, i.e., the first queue 351, queue 352, and a third second queue 353, for storing different external requests 20 .

此外,在初始状态下排程器36中包含有对应于每个队列的计数器(配置数量也与队列相同),即第一队列计数器361、第二队列计数器362、第三队列计数器363,在排程器36中会对每个计数器都设定有权重值,如:第一队列计数器361的权重值为600、第二队列计数器362的权重值为300、第三队列计数器363的权重值为100,此权重值主要是用来控制外部请求.20所能够取得的应用服务器40服务资源比例,可以依照不同需求作调整设定。 Furthermore, scheduler 36 is included in the initial state of each counter corresponding to the queue (queue number and configuration is also the same), i.e. a first queue counter 361, the counter 362 a second queue, a third queue counter 363, the row 36 will process in each counter is set with the weight value, such as: a first queue counter weight 361 is weight 600, weight of the second queue counter weight 362 is 300, the third queue counter weight 363 is weight 100 this weight value is mainly used to control the ratio of the external service resource request 40 as the application server can obtain .20 can be adjusted according to different requirements set. 排程器36 中还包含有用来控翁j各个队列中外部请求20传送处理的队列指针364和排程窗户值365,队列指针364主要是依照Deficit Round Robin的演算方式来进行循环移动,当移动到队列时,将进行队列中等待传送的外部请求20的回应大小(回应大小可从请求/回应表32中取得)与对应队列的计数器中权重值的比较,来决定是否可以传送该外部请求20,而排程窗户值365则是设定用来控制在同一时间内可传送给应用服务器40的外部请求20数量,此设定值也可根据应用服务器40的处理能力来进行设定。 36 scheduler further comprises for controlling queue pointer 364 and a scheduling window value Weng j each queue external request transmitted 20 processing 365, the queue pointer 364 is carried out primarily cyclic shift in accordance with the calculation mode Deficit Round Robin, and when the mobile when the queue, the external queue awaiting transmission request response size (response size from the request / response table 32 acquired) 20 and the comparison weight values ​​of weights counter corresponding to the queue to determine whether to transmit the external request 20 , while the scheduling window value is set 365 to control the application may be transmitted to the server at the same time the number of external requests 20 40, this set value may be set according to the application server 40 in processing capability.

以下将通过图2A来说明本发明外部请求排程方法的流程。 Outside will be described below the present invention is a method of scheduling request procedure through FIG. 2A. 首先,步骤100, 利用探查器31探査应用服务器40中一个以上请求项目及其对应的回应大小(是指应用服务器40在执行外部请求20所指定的请求项目时所需要传送的实际字节数量大小),并记录于请求/回应表32中,探查方式是采取阶层式方式进行,由应用服务器40中第一阶层的请求项目开始探查其回应大小,当第一阶层的探查结束后再根据第一阶层中的其它链接关系继续第二阶层的请求项目探查,持续递归探查直到完成所有请求项目的回应大小记录后结束;接着, 步骤200,通过事先建立的一份服务品质策略表34,由分类器33开始接收来自各客户终端10 —个以上的外部请求20然后进行外部请求分类,并根据分类结果暂存各外部请求20至对应一个以上队列,在服务品质策略表34中包含有 The actual bytes to be transmitted First, in step 100, probe 31 using a profiler application server 40 a request response size than its corresponding item (means 20 requests the application server 40 designated by the execution request item number of external size ), and recorded in the request / response table 32, the probe is to take a hierarchical manner way, the requested items in the first class of the application server 40 starts its probe response size, when the first level after the end of the probe according to the first other links relationship hierarchy continues to request a second-class exploration projects, exploration continued recursively until the end of recording in response to the size of the project after all the requests; Next, step 200, by a quality of service policy table 34 previously established by the classifier 33 starts receiving from each client terminal 10 - 20 or more of external requests and external requests classified, and temporarily stores the respective external request queue 20 according to a corresponding one of the above classification results, the quality of service contained in the policy table 34 has

一个以上不同的分类项目、51,用来对所有的外部请求20作分类,队列配置的数量则依照服务品质策略表34来决定;最后,步骤300,在排程器36的计数器中加上各队列所对应的权重值,并同时产生一个队列指针364及设定一个排程窗户值365,然后以循环方式根据外部请求20的回应大小开始进行各队列中各外部请求20的传送处理。 More than one different classification items, 51, 20 used for classification request to all the external, the number of queues is determined in accordance with the configuration of quality of service policy table 34; Finally, step 300, together with the respective counter in scheduler 36 the weights of the queues corresponding to a weight value, and simultaneously generates a queue pointer 364 and a schedule setting window value 365, and then respond to the request in a cyclic manner based on the external size of 20 starts each respective queue 20 external request transfer processing.

有关步骤300的细部流程,请参考图2B,说明如下。 For details of the process step 300, refer to Figure 2B, described below. 首先,步骤301设定好排程窗户值365,并同时进行初始化将队列指针364指向最前面的一个队列,逐一扫瞄所有队列,将有至少一个外部请求20的队列加入有效队列列表, 排程器36只会排程在有效队列列表中的所有队列。 First, in step 301 the window value is set up schedule 365, and simultaneously initialize the queue pointer 364 points to the front of a queue, the queue one by one all the scanning, will have at least one external request queue 20 is added valid queue list, schedule 36 will schedule all active queues in the queue list. 接着,步骤302,判断在有效队列列表中是否有至少一个队列存在,若是没有,则代表在队列器35中没有任何外部请求存在,则排程器36不需要进行排程,结束整个运作方法(步骤300)的流程运作。 Next, step 302 determines whether the list of valid queue at least one queue exists, if not, it means there is no external request queue 35, the scheduler 36 does not need to be scheduled, the operating end of the entire process ( step 300) the process works.

若是有效队列列表中存在有至少一个队列,则步骤303,初始化该队列指针364使其指向有效队列列表中最先队列,增加最先队列对应的权重值于最先队列的一计数器中。 If there is valid queue list at least one queue, step 303, initializes the queue pointer 364 to point to the first queue valid queue list, increase the weight of the first weight values ​​corresponding to the queue counter in a first queue. 接着步骤304,判断该队列中仍有未处理的该外部请求20, 若该队列中仍有未处理的该外部请求20,则步骤305,将该队列加入有效队列列表,若是该队列已经存在于有效队列列表则不需重复加入;然后步骤306, 开始读取队列指针364所指向的队列中的外部请求20,并通过对请求/回应表32的读取来判断此外部请求20所对应的回应大小;步骤307,当回应大小小于或等于计数器中的数值,步骤308,且排程窗户值365不为0时,步骤309, 则将计数器减去回应大小,将排程窗户值365减去1,然后传送外部请求20 给应用服务器40去进行处理。 Next step 304 determines that the queue is still unprocessed external request 20, still in the queue if the external request untreated 20, step 305, the queue is a list of valid queue is added, if the queue is already present in the the list of valid queue need not be repeated added; then, step 306, starts reading the external queue pointer 364 points to the request queue 20, and through the request / response table reading unit 32 in addition to determining a request response corresponding to 20 size; step 307, when the response size smaller than or equal to the value of the counter in step 308, and the schedule window value 365 is not zero, step 309, then the counter is decremented by the size of the response, the schedule window 365 subtracts the value 1 , 20 then transmits the request to the external server 40 to the application for processing.

步骤304,当所指向的队列中仍有未处理的外部请求20时,将重复上述该队列指针364所指向的队列中读取外部请求20进行排程的动作(即重复歩骤305、步骤306、步骤307、步骤308及步骤309的流程)。 Step 304, when the external request is directed queue 20 is still untreated, will repeat the operation of the queue pointer 364 points to an external read request queue 20 for the schedule (i.e., step 305 is repeated ho, step 306, step 307, step 308 and step 309 of the flow). 步骤304,当所指向的队列中的外部请求全部处理完时,则将该队列从有效队列列表中移除,步骤312,然后判断排程器36是否已经完成此回合所有队列的排程,若是,则重新回到步骤302继续执行。 Step 304, when the external request is directed completely processes the queue, then the queue is removed from the active list of queues, step 312, then scheduler 36 determines whether all the complete round scheduling queue, and if so, then back to step 302 to continue. 若是尚末完成此回答所有队列的排程,步骤310, 则移动该队列指针364至下一队列,并增加该队列所对应的权重值至该队列的该计数器中,完成后再回到步骤304继续执行。 If Shangmo complete answer all scheduling queue, step 310, the queue pointer 364 is moved to the next queue, and increase the weight of the queue corresponding to the weight value of the counter in the queue, then returns to step 304 to complete the continue. 另外,步骤307,当该回应的大小大于该计数器的数值时,也会造成步骤310,该队列指针移动至次一队列并增加该队列所对应的权重值至该队列的该计数器中。 Further, in step 307, when the response is greater than the size of the value of the counter, will cause step 310, the queue pointer is moved to the right to increase the time a queue and the queue corresponding to the queue weight value of the counter.

前面提到排程窗户值365的作用在于控制同一时间内可以传送给应用服务器40处理的外部请求20数量。 The aforementioned schedule window 365 is to control the value of the action of external processing application 40 may be communicated to the server 20 requests the same amount of time. 因此,每当步骤309传送一个外部请求20 时,排程窗户值365会自动减去1,当排程窗户值365等于0的时候,所有的外部请求20都会被暂时保留在队列中等待应用服务器40的处理,此时便会进入到图2C的步骤A,步骤341,等待并判断是否收到来自应用服务器40的回应。 Thus, step 309 whenever a request external transmission 20, the schedule window 365 automatically subtracts the value 1, when the schedule window value 365 is equal to 0, all external requests 20 will be temporarily retained in the queue waiting for an application server 40 process, the FIG. 2C at this time will enter the step a, step 341, and determines whether or not to wait for response from the application server 40. 当应用服务器40处理完一个外部请求20时,会主动传回一个对应的回应给外部请求排程系统30,步骤342,此时排程窗户值365便会自动加上1,表示应用服务器40可以继续处理其它外部请求20,此时可以恢复对外部请求20 的传送处理,因此将再通过步骤B回到图2B中的步骤308,.继续后续的运作流程。 When finished processing the application server 40 a request external 20, will automatically transmit a response back to the corresponding external request scheduling system 30, step 342, this time schedule window 365 will automatically add a value indicating the application server 40 may continues to process other external requests 20, 20 can now resume the transfer processing of an external request, so Jiangzai step returns to step B in FIG. 2B 308 ,. continue with the subsequent operation of the process.

最后通过图1C及图1D来作实施例的说明。 Finally, to make the embodiment described by FIG. 1C and FIG. 1D. 由图1C中可知经过探査、分类及初始化步骤后的队列器35中配置有三个队列(表示有效队列列表中有三个队列):即第一队列351、第二队列352及第三队列353,每个队列中均已经储存有各自的外部请求20,如:第一队列351的外部请求有351A、 351B及351C(根据请求/回应表32可得知,其回应大小判别结果分别为300、 200及150)、第二队列352的外部请求有352A、 352B及352C(其回应大小判别结果分别为250、 300及150)、第三队列353的外部请求有353A、 353B及353C(其回应大小判别结果分别为200、 150及250);而在排程器36中对应于每个队列所配置的计数器分别为:第一队列计数器361(其权重值大小设定为600)、 第二队列计数器362(其权重值大小设定为300.)及第三队列计数器363(其权重值大小设定为100),排程器36中还包含一个队列指针364(在初始状态时指向第一队列351所属的计数器),以及 After probing understood, the classification and queue initialization step 35 in FIG. 1C is configured by three queues (queue list indicates valid queue has three): namely a first queue 351, queue 352, and a third second queue 353, each queues are already stored in the respective external request 20, such as: a first external request queue 351 are 351A, 351B and 351C (table 32 can be learned according to the request / response, which response size determination result of 300, 200 and 150), a second external request queue 352 are 352A, 352B and 352C (which are response size determination result 250, 300 and 150), a third external request queue 353 are 353A, 353B and 353C (which response size determination result respectively 200, 150 and 250); and 36 in the scheduler corresponding to each counter of queues are: a first queue counter 361 (which is set to a weight value of the size 600), the second queue counter 362 ( its weight is set to 300. the value of the size) and the third queue counter 363 (which size is set to a weight value 100), the scheduler 36 further comprises a queue pointer 364 (in the initial state of the first point 351 belongs queue counter), and 个排程窗户值365 (设定为1)。 A scheduling window value 365 (set to 1).

当开始进行外部请求排程处理时,由于队列指针364指向第一队列351 所属的第一队列计数器361上,因此先在第一队列计数器361中加入对应的权重值600并开始进行第一队列351的外部请求20处理,此时将读取第一队列中的第一个外部请求351A,由于其回应大小为300小于第一队列计数器361 中的数值600,且此时排程窗户值365不为0,因此排程器36便将此外部请求后将第一队列计数器361数值600减去0f传送的第一个外部请求351A的回应大小300后成为300并同时将排程窗户f直减去1后变成0,接着继续读取第一队列351中的第二个外部请求351B,此时虽然其回应大小200小于第一队列计数器361数值300,但是因为排程窗户值365 When starting the right external request scheduling process, since the queue pointer 364 points to the first queue 351 belongs to a first queue counter 361, and thus to join the first queue corresponding to the weight value of the counter 361 and begin a first queue 600 351 the external request process 20, a first outer case first read request queue 351A, since the size of its response queue 300 is smaller than the value of the first counter 361 is 600, and the value at this time schedule window 365 is not after 0, so the scheduler 36 put addition request after the first outer portion 361 of the first queue counter 600 minus the value of the transfer request response size 0f 300 351A becomes 300 and simultaneously is subtracted a linear schedule window f 1 after becomes 0, and then continues to read the second external requests in the first queue 351 351B, although the case 200 is smaller than the size of the first response queue 361 counter value 300, but since the scheduling window value 365

;为0,因此将暂停传送任何的外部请求20,并等待排程窗户值365不为0。 ; It is 0, suspending any external transfer request 20, and waits for the schedule window value 365 is not zero.

当应用服务器40在接收到第一队列351的第一个外部请求351A后便开始进行处理,处理完毕后便传回一个回应给排程器36,再通过互联网15传回给客户终端10,此时排程窗户值365便可以增加1由0变成为1,当排程器36 发现排程窗户值不为0时,则恢复传送前述的第二个外部请求351B,然后将 When the application server 40 first receives a first external request queue 351 351A after processing is started, the process is completed after a response returned to the scheduler 36, and then returned to the client terminal 10 through the Internet 15, this when scheduling windows 365 can increase the value changes from 0 to 1 1, when the scheduler 36 schedules found window value is not 0, then resume the transfer request to the second external 351B, then

)第一队列计数器361数值300减去所传送的第二个外部请求351B的回应大小200后成为100并同时将排程窗户值减去1后变成0。 After) the second counter 361 value outside the first queue 300 subtracts the transmitted request response size becomes 200 351B 100 and simultaneously schedule window value obtained by subtracting 1 becomes 0. 接着继续读取第三个外部请求351C,而此时因为第三个外部请求351C的回应大小150超过第一队列计数器361数值100,因此便将队列指针364移至第二队列352所属的第二队列计数器362,并增加对应的权重值300至第二队列计数器362中。 Then continues to read the third external request 351C, at which point 351C as the third external request response 150 exceeds the first queue size counter 361 value of 100, and therefore put to a second queue pointer 364 belongs to a second queue 352 queue counter 362, and adds the corresponding weight value 300 to a second queue counter 362.

; 开始读取第二队列352中的第一个外部请求352A,此时虽然其回应大小 ; Start reading a first external requests in the second queue 352 352A, its response time although the size

250小于第二队列计数器362数值300,但是由于排程窗户值365仍为0,因此暂停传送第二队列352中的第一个外部请求352A,直到应用服务器40处理完毕前一个外部请求20后(即第一队列351的第二个外部请求351B)传出回应给排程器36,再透过互联网15传回给客户终端10,之后将排程窗户值365 The second queue 250 is less than the counter value of 362 300, but since the value of the schedule window 365 remains 0, and therefore the first suspension a second external transmit queue 352 of the request 352A, the processing until the completion of the application server 40 a request to the external 20 ( i.e., the first queue 351 requesting the second external 351B) outgoing response to scheduler 36, via the Internet 15 and then returned to the client terminal 10, then the value of the schedule window 365

'增加1后才进行后续的传送处理。 '1 increases after subsequent transfer process.

在同样的运作模式下继续进行处理,当队列指针364停留在第三队列353 所属的第三队列计数器363上等待移动时,表示已经完成一个循环的队列处理,此时队列指针364将依据余额累计式循环排程算法(Def icit Round Robin) 模式重新被初始化设定到第一队列计数器361上继纟i进行处理,此时同样会在第一队列计数器361中增加对应的权重值600,因此新的第一队列计数器361 数值由原先的100变成700,继续进行对第一队列351中的第三个外部请求351C的处理。 Processing continues under the same operation mode, when the queue pointer 364 rests on the third queue 353 queues the third counter 363 waits for the mobile belongs, represents a complete cycle of the processing queue, the queue pointer 364 at this time based on the accumulated balance circulation scheduling algorithm (Def icit Round Robin) is re-initialized mode set to the following processing on a first queue i Si counter 361 at this time will also increase the weight value corresponding to the queue 600 in a first counter 361, so the new the first queue 100 from the original value of the counter 361 becomes 700, the process 351C proceeds to the third external requests in the first queue 351.

当处理中的队列没有其它等待处理的外部请求20时,队列指针364将会自动移至下一个队列进行处理。 When the queue is not processing other pending external request 20, the queue pointer 364 automatically moves to the next queue for processing. 如此的运作流程将持续到所有队列中的所有外'部请求20处理完毕为止。 Such operation process will continue until all queues all external 'portion 20 until the request is processed. 要注意的是,当某一个队列没有任何外部请求20 It is noted that when a queue does not have any external requests 20

14时,该队列将会从有效队列列表中移除,在有新的外部请求20进入队列后, 该队列才会被重新加入到有效队列列表。 14, the queue will be removed from active list of queues, the new 20 into the external request queue, which will be re-added to the active queue list. 队列指针只会指向有效队列列表中的队列,而且加上权重值到队列的计数器,没有在有效队列列表中的队列将不会被进行排程,当然其计数器也不会被加上权重值。 Queue pointer will point to a valid queue list queue, and the queue plus the weight value of the counter, there is no active queue list in the queue will not be scheduled, of course, its counter will not be combined with the weight value.

图1D为上述实施例的另一示意图,由图中可以确认整个外部请求排程系统30对于所有外部请求20的处理顺序逻辑,并且显示当排程器36每传送出一个外部请求20给应用服务器40时,便会使得排程窗户值365减去1,而每当应用服务器40处理完一个外部请求时,便会使得排程窗户值365增加1 。 36 per transmit a request external 1D is a schematic view of another embodiment of the above-described embodiment, it was confirmed from the figure the entire external request scheduling system 30 for all external request processing sequence logic 20, and when the display scheduler 20 to the server application 40, so that it will schedule window 365 subtracts the value 1, the application server 40 whenever an external request is finished processing, will be such that the schedule window 365 increases the value 1.

当然,本发明还可有其他多种实施例,在不背离本发明精神及其实质的情况下,熟悉本领域的技术人员可根据本发明作出各种相应的改变和变形,但这些相应 Of course, the present invention may have various other embodiments without departing from the spirit and the essence of the present invention, those skilled in the art can make various corresponding modifications and variations according to the present invention, these corresponding

Claims (11)

  1. 1、一种在网络中继点上提供差别服务品质的外部请求排程系统,该系统设置于介于一个以上客户终端与一应用服务器之间的网络中继点上,其特征在于,包含有: 一探查器,用以探查该应用服务器中一个以上请求项目及其对应的回应大小,并记录于一请求/回应表中,该回应大小是指执行该外部请求所指定的该请求项目时应用服务器所需要传回的实际字节数量大小; 一分类器,用以建立一服务品质策略表,并负责接收来自各该客户终端一个以上外部请求进行一外部请求分类; 一队列器,根据该服务品质策略表配置一个以上队列,用以根据分类结果暂存该外部请求;及一排程器,用以设定各该队列对应的一权重值,以及一队列指针及一排程窗户值,并根据各该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理,其中, A relay point in the network to provide different quality of service external requests scheduling system, the system is provided on a network relay point between a terminal and a server interposed between more than one client application, wherein, comprising : a profiler application server to the probe request more than one size of the project and its corresponding response, and recorded in a request / response table, the size refers to the response to the external request is performed to the specified application program requests the server returns the required size of the actual number of bytes; a classifier for establishing a quality of service policy table, and the client is responsible for receiving from each terminal a request for a more external classification external request; a queue, a service based on the quality strategy table configure more than one queue for temporarily storing the classification result according to an external request; and a scheduler for setting each of the queues corresponding to a weight value, and a queue pointer to the window and a scheduled value, and in a cyclic manner each of the transfer processing is started in each of the external request queue based on the size of each of the response to an external request, wherein, 排程窗户值是用以控制在同一时间内可传送至应用服务器的该外部请求数量。 Schedule windows for controlling the value of the external server may transmit to the application at the same time the number of requests.
  2. 2、 根据权利要求1所述的在网络中继点上提供差别服务品质的外部请求排程系统,其特征在于,该外部请求分类是指依照该外部请求的一网络层内容进行分类。 2, the external differential quality of service provided on a network relay point according to claim 1 request scheduling system, characterized in that the external request classification means to classify the content of the network layer in accordance with an external request.
  3. 3、 根据权利要求1所述的在网络中继点上提供差别服务品质的外部请求排程系统,其特征在于,该外部请求分类是指依照该外部请求的一应用层内容进行分类。 3, the external differential quality of service provided on a network relay point according to claim 1 request scheduling system, characterized in that the external request classification means to classify the content of the application layer in accordance with an external request.
  4. 4、 根据权利要求1所述的在网络中继点上提供差别服务品质的外部请求排程系统,其特征在于,该队列指针是以余额累计式循环排程算法的方式进行循环方式的移动。 4, there is provided according to claim 1, the relay point in the network requests the external service quality difference scheduling system, characterized in that the moving mode of the circular mode of queue pointer is accumulated balance circulation scheduling algorithm.
  5. 5、 根据权利要求1所述的在网络中继点上提供差别服务品质的外部请求排程系统,其特征在于,该排程器中还包含对应于各该队列的一个以上计数器, 用以累加该队列对应的该权重值。 5. The difference in the quality of service provided externally with the network relay point according to claim 1 request scheduling system, characterized in that the scheduler also includes each of the queues corresponding to more than one counter for accumulating the queue corresponding to the weight value.
  6. 6、 一种在网络中继点上提供差别服务品质的外部请求排程方法,该方法执行于介于一个以上客户终端与一应用服务器之间的网络中继点上,其特征在于,包含有下列步骤.-探查该应用服务器中一个以上请求项目及其对应的回应大小,并记录于一请求/回应表中,该回应大小是指执行该外部请求所指定的该请求项目时应用服务器所需要传回的实际字节数量大小;建立一服务品质策略表,开始接收来自各该客户终端一个以上外部请求进行一外部请求分类,并根据分类结果暂存各该外部请求至对应一个以上队列;及设定各该队列的一权重值,以及一队列指针及一排程窗户值,并根据各该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理,其中,该排程窗户值用以控制在同一时间内可传送至应用服务器的该外部请求数量。 6. A method of providing quality of service in different network relay point external request scheduling method executed on more than one client on the network between the relay point between a terminal and an application server, wherein, comprising .- probe the following steps in more than one application server and request responses corresponding to the size of the item, and recorded in a request / response table, the size refers to the response to the external request is performed when the requested item specified by the application server require returns the number of bytes of actual size; establishing a quality of service policy table starts receiving external client terminal from each of the more than one external request a classification request, and temporarily stores each of the external request queue according to a corresponding one or more classification results; and setting each of the one weighting value of the queue and a queue pointer, and a scheduling window value, and in a circular manner starts transfer processing each of the external requests of each of the queues based on the response to the size of each of the external request, wherein, the schedule window value for controlling the application may be transmitted to the external server at the same time the number of requests.
  7. 7、 根据权利要求6所述的在网络中继点上提供差别服务品质的外部请求排程方法,其特征在于,该外部请求分类是指依照该外部请求的一网络层内容进行分类。 7, according to claim quality provide differentiated services over a network relay point outside the scheduling request 6, characterized in that the external request classification means to classify the content of the network layer in accordance with an external request.
  8. 8、 根据权利要求6所述的在网络中继点上提供差别服务品质的外部请求排程方法,其特征在于,该外部请求分类是指依照该外部请求的一应用层内容进行分类。 8, according to claim quality provide differentiated services over a network relay point outside the scheduling request 6, characterized in that the external request classification means to classify the content of the application layer in accordance with an external request.
  9. 9、 根据权利要求6所述的在网络中继点上提供差别服务品质的外部请求排程方法,其特征在于,该队列指针是以余额累积式循环排程算法的方式进行循环方式的移动。 9, there is provided according to claim 6 in the external network relay point difference method of quality of service scheduling request, wherein the mobile mode of the circular mode of queue pointer is crediting circulation scheduling algorithm.
  10. 10、 根据权利要求6所述的在网络中继点上提供差别服务品质的外部请求排程方法,其特征在于,该设定各该队列的该权重值,以及该队列指针及该排程窗户值,并根据该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理的步骤,还包含下列步骤:(a) 设定该排程窗户值,初始化该队列指针使其指向最先该队列,并增加该队列对应的该权重值于该队列的一计数器中;(b) 读取该队列中该外部请求,并判断该外部请求所对应的该回应大小;(c) 当该回应大小小于该计数器且该排程窗户值不为0时,将该计数器减去该回应大小,将该排程窗户值减去l,然后传送该外部请求至应用服务器进行处理;及(d)移动该队列指针至下一队列,并增加该队列所对应的该权重值至该队列的该计数器中;其中在步骤(c)中,当该队列仍有未处理的该外 10, the external differential quality of service provided on a network relay point according to claim 6 scheduling request method, wherein the weighting of each of the set value of the queue, and the queue pointer and the schedule window value and proceeds to a step in a cyclic manner to each of the transfer processing each of the external request queue based on the magnitude of the response to external requests, further comprising the steps of: (a) setting the schedule window value, initializing the queue pointer to point to the first queue, and increase the weight value corresponding to the queue to the queue of a counter; (b) the external read request of the queue, and determines whether the external request corresponding to the response size; ( c) when the size is less than the response counter and the schedule window value is not 0, the counter is decremented by the size of the response, the schedule window minus L, and then transmits the external processing request to the application server; and (d) moving the queue pointer to the next queue, and the queue to increase the weight corresponding to a weight value of the counter in the queue; wherein in step (c), when the queue still has the outer untreated 请求或该回应大小不大于该计数器时,将持续重复执行步骤(b)及步骤(c);在一个循环回合中尚未完成所有队列的排程时,将持续重复执行步骤(b) 至步骤(d);当完成该循环回合中所有队列的排程后,将重新初始化该队列指针使其指向最先的该队列,并增加该队列对应的该权重值于该队列的该计数器中,然后重新执行步骤(b)至步骤(d)。 When the request or the response size not greater than the counter will continue repeating steps (b) and step (C); when a loop round has not been completed for all of the queue scheduling, will continue repeating steps (b) to step ( d); upon completion of the cycle round all the scheduling queue, the queue pointer is re-initialized to point to the first queue, and to increase the weighting of the value of the counter to a queue corresponding to the queue, and then re- step (b) to step (d).
  11. 11、根据权利要求10所述的在网络中继点上提供差别服务品质的外部请求排程方法,其特征在于,该设定该队列的该权重值,以及该队列指针及该排程窗户值,升根据各该外部请求的该回应大小以循环方式开始进行各该队列中各该外部请求的传送处理的步骤,还包含当接收到来自该应用服务器处理完毕该外部请求的回应时,将该排程窗户值增加1的步骤。 11, claim different quality of service provided on a network relay point outside the scheduling request 10, characterized in that the weighting value is the setting of the queue, and the queue pointer value and the schedule window , l a cyclic manner the start of each step of each of the external request queue of the transfer processing based on the size of each of the response to external requests, further comprising when the application server receives the response from the processed external requests, the schedule window value is increased in step 1.
CN 200510059615 2005-03-29 2005-03-29 Request scheduling system and method for providing difference service quality on network relay point CN100531180C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 200510059615 CN100531180C (en) 2005-03-29 2005-03-29 Request scheduling system and method for providing difference service quality on network relay point

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 200510059615 CN100531180C (en) 2005-03-29 2005-03-29 Request scheduling system and method for providing difference service quality on network relay point

Publications (2)

Publication Number Publication Date
CN1842072A true CN1842072A (en) 2006-10-04
CN100531180C true CN100531180C (en) 2009-08-19

Family

ID=37030916

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200510059615 CN100531180C (en) 2005-03-29 2005-03-29 Request scheduling system and method for providing difference service quality on network relay point

Country Status (1)

Country Link
CN (1) CN100531180C (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6728748B1 (en) 1998-12-01 2004-04-27 Network Appliance, Inc. Method and apparatus for policy based class of service and adaptive service level management within the context of an internet and intranet

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6728748B1 (en) 1998-12-01 2004-04-27 Network Appliance, Inc. Method and apparatus for policy based class of service and adaptive service level management within the context of an internet and intranet

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
1531282A 2004.09.22

Also Published As

Publication number Publication date Type
CN1842072A (en) 2006-10-04 application

Similar Documents

Publication Publication Date Title
Schmidt et al. An overview of the real-time CORBA specification
US7680897B1 (en) Methods and systems for managing network traffic
Dogar et al. Decentralized task-aware scheduling for data center networks
US20050034130A1 (en) Balancing workload of a grid computing environment
US8504689B2 (en) Methods and systems for cloud deployment analysis featuring relative cloud resource importance
US20030182464A1 (en) Management of message queues
US20080183800A1 (en) Mobile device management proxy system
EP1526682A2 (en) Network and interface selection on a computing device
US20100234009A1 (en) Centrally managing user-specified configuration data for a configurable device
US5802306A (en) Supporting multiple client-server sessions from a protocol stack associated with a single physical adapter through use of a plurality of logical adapters
US5682478A (en) Method and apparatus for supporting multiple, simultaneous services over multiple, simultaneous connections between a client and network server
US20030018784A1 (en) System and method for processing requests from newly registered remote application consumers
US6976258B1 (en) Providing quality of service guarantees to virtual hosts
US7321940B1 (en) Iterative architecture for hierarchical scheduling
US20030084144A1 (en) Network bandwidth optimization method and system
US7437447B2 (en) Method and system for authenticating a requestor without providing a key
Carlstrom et al. Application-aware admission control and scheduling in web servers
Munir et al. Friends, not foes: synthesizing existing transport strategies for data center networks
Nolte et al. Real-time server-based communication with CAN
US20060047794A1 (en) Application of genetic algorithms to computer system tuning
CN1670698A (en) Transparent computing based computing apparatus and method
US20090300161A1 (en) Method and system for using feedback in accessing network services
CN102611735A (en) Load balancing method and system of application services
CN102082692A (en) Method and equipment for migrating virtual machines based on network data flow direction, and cluster system
US20070180453A1 (en) On demand application scheduling in a heterogeneous workload environment

Legal Events

Date Code Title Description
C06 Publication
C10 Request of examination as to substance
C14 Granted