JP5291366B2 - Flow control device - Google Patents

Description

  The present invention relates to a flow rate control technique for limiting the number of requests accepted via a network.

  In general, when operating a server that accepts requests via a network, a flow control device that controls the request flow rate of web access by limiting the upper limit of the number of requests so that requests are not received beyond the processing capacity of the server Is placed in front of the web server. As a result, it is possible to prevent a situation in which the processing capability of the server is punctured and the response is not returned, and to maintain a response from the web server in an appropriate response time.

  Conventionally, in the system industry, it is common to obtain an agreement on system performance requirements in units of TPS (Transaction Per Sec), which is the number of transactions that can be processed per second, and depending on the number of simultaneous connections with the processing system per unit time There is little consensus on performance requirements. On the other hand, in the conventional flow control device, the flow control is usually executed by setting the upper limit of the number of simultaneous connections.

For example, when a new session establishment request is received from an end user terminal, Patent Literature 1 compares the upper limit value of the number of sessions that can be connected per unit time with the number of connection sessions established in the most recent unit time. A server load balancer that discards a session establishment request when the number of established sessions is equal to or greater than an upper limit value is disclosed.
JP 2004-246833 A

  However, as described in Patent Document 1, if the TPS is secured by guaranteeing the number of simultaneous connections in the flow rate control device, there is a risk that the subsequent system will have excessive performance. For example, as shown in FIG. 1, it is assumed that the number of simultaneous connections between the client and the processing system is set to “10” in order to secure 10 TPS. In this case, as shown in FIG. 1A, if the requests reach the system on average, the number of connections is two. However, as shown in FIG. 1B, consider a case where 10 requests that require an average processing time of 200 milliseconds are concentrated at a time. In this case, when the number of simultaneous connections is 10, as shown in FIG. 1C, a system having a processing performance of 50 TPS at the maximum must be prepared. In the flow control by determining the upper limit of the number of simultaneous connections in this way, it is necessary to make the system so as to have a performance exceeding the requirements, assuming that requests are concentrated at one time. Thus, there is a problem that the performance requirements agreed with the client do not match the actual specifications of the system.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a flow rate control technique that enables the design of a processing system that does not become excessive performance while satisfying the performance requirements of the subsequent processing system. There is.

  One embodiment of the present invention is a flow control device. This device receives an information processing request for a system that executes a predetermined process from a client terminal, and based on a reception request count unit that counts the number of requests received within a unit time, and a notification of a processing result from the system to the client terminal, A processed request count unit that counts the number of requests that have been processed within a unit time, and when the number of requests counted by the processed request count unit reaches a predetermined upper limit within the unit time, a new An error processing unit that returns an error response to the client terminal that issued the request, a request holding unit that stores the requests received by the received request counting unit in a queue, and the number of simultaneous connections that can be processed simultaneously on the system So as not to exceed Comprising the simultaneous connection restriction unit that passes requests in list holding unit in the system, the.

  According to this aspect, the acceptance of requests is limited based on the number of requests that the processing system has actually finished processing. On the other hand, the delivery of the request to the processing system is processed based on the number of simultaneous connections with the processing system. As a result, requests are received only for the number of times the processing system has actually finished processing, so the requirements for the TPS determined with the orderer of the processing system and the number of requests received in the flow control device are limited. And become the same.

  The request holding unit includes a type determination unit that determines the type of the received request, and a queue control unit that stores a request for each request type in a queue and determines the timing for passing the request to the system for each request type. Also good. According to this, it is possible to execute priority control for controlling the order in which the received request is passed to the processing system based on the priority according to the type.

  The simultaneous connection restriction unit is a type determination unit that determines the type of the received request, a queue that stores the requests received from the request holding unit for each type in a queue, and a queue that determines the timing for passing the request to the system for each request type And a control unit. According to this, it is possible to execute priority control for controlling the order in which the received request is passed to the processing system based on the priority according to the type.

  The type determination unit includes a priority setting unit in which the priority of processing in the subsequent system is set for each request type, and a timing determination unit that extracts a request from a queue holding a request type with a high priority. Also good. According to this, it becomes possible to preferentially process a specific type of request.

  The concurrent connection restriction unit is a unit in the system based on the processing time holding unit that calculates and holds the average processing time for each request type in the system, the request stored in the queue for each request type, and the average processing time. A request selection unit that selects requests so that the number of requests that can be processed in time is maximized. According to this, it is possible to increase the number of request processes per unit time in the processing system while securing the number of simultaneous connections.

  It should be noted that any combination of the above-described components and a representation of the present invention by a method, apparatus, system, recording medium, and computer program are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, the flow control which enables the design of the processing system which does not become excess performance can be performed, satisfying the performance requirement of the processing system arrange | positioned after a flow control apparatus.

  FIG. 2 is an overall view of the client server system 100 including the flow control device 20 according to the embodiment of the present invention. The client terminal 12 is a computer terminal that includes at least an input device and a display device and has a function of connecting to a network. A user can access various websites via a network 14 such as the Internet or a LAN (Local Area Network) using a browser installed in the client terminal 12. The client terminal 12 may be a personal computer (PC), a mobile phone, a PDA (Personal Digital Assistant), a POS register, a vending machine, or the like depending on the type of processing requested to the processing system 16.

  The processing system 16 is a web server or an application server, receives a processing request from the client terminal 12 via the network 14, and returns a response to the request to each client terminal 12. The processing system 16 is, for example, a billing system, a reservation system, a payment system, etc., but there is no limitation on the target processing.

  The flow control device 20 is disposed between the processing system 16 and the network 14 and receives a processing request for the processing system 16. Then, the concentration of requests with respect to the processing system 16 is detected, and reception limit control for requests exceeding the processing capacity of the system, priority control for specific requests, and the like are executed. In the present embodiment, the flow control device 20 focuses on the number of transactions that can be processed by the processing system 16 per unit time instead of performing flow control based on the number of simultaneous connections and the number of simultaneous executions in the conventional flow control device. Then, flow control is executed. This will be described later.

  FIG. 3 is a functional block diagram showing a detailed configuration of the flow control device 20 of FIG. Each block shown here can be realized in hardware by an element such as a CPU and a memory of a computer, and is realized in software by a computer program having the functions described below. It is drawn as a functional block realized by the cooperation. Therefore, those skilled in the art will understand that these functional blocks can be realized in various forms by a combination of hardware and software.

  The reception request count unit 30 receives a processing request from each client terminal to the processing system 16. Then, the number of requests accepted within a predetermined unit time (for example, 1 second) is counted.

  The request holding unit 32 stores (queues) the request received by the received request counting unit 30 in a queue. With this queuing, it is not necessary to synchronize the timing of requests from client terminals and the timing of request processing in the processing system 16, and requests can be reliably processed as long as the number of requests is within the request upper limit value described later. it can.

  The simultaneous connection restriction unit 34 extracts the requests queued in the request holding unit 32 so as not to exceed the number of simultaneous connections set as the number of requests that can be processed simultaneously on the processing system, and passes the requests to the processing system 16. .

  Thus, the simultaneous connection restriction unit 34 controls the flow rate of requests flowing to the processing system 16 by defining the number of connections that can be connected to the subsequent processing system 16.

  The processed request count unit 24 relays the processing result notification from the processing system 16 to the client terminal 12. When this notification is received, the processed request count unit 24 determines that the operation corresponding to the request is completed in the processing system 16. The processed request count unit 24 counts the number of requests that have been processed within a predetermined unit time (for example, 1 second). However, the length of this unit time is preferably the same as the unit time for counting the number of accepted requests in the accepted request counting unit 30.

  The error processing unit 26 monitors the number of processed requests within the unit time counted by the processed request counting unit 24. When the processed request reaches a predetermined upper limit value within the unit time (hereinafter referred to as “processed request upper limit value”), the client that has issued a request that has reached the reception request count unit 30 thereafter An error response indicating that the processing system 16 is busy and cannot accept the request is returned to the terminal 12. In this way, by detecting the release of a connection due to the end of processing for a request, it is possible to restrict reception of a request that exceeds the processing capability of the processing system 16.

  The setting unit 22 changes parameters such as the number of simultaneous connections, the processed request upper limit value, and unit time based on a command from the system administrator.

  Hereinafter, the operation of the flow control device 20 according to the present embodiment will be described.

  The processed request count unit 24 counts the number of requests processed by the processing system 16 within a predetermined unit time, and for a request that exceeds a prescribed processed request upper limit value, the error processing unit 26 causes the client terminal to An error response is returned to 12. At the same time, the simultaneous connection restriction unit 34 sequentially passes the requests queued in the request holding unit 32 to the processing system 16 so that the connection between the processing system 16 and the flow rate control device 20 is equal to or less than the maximum number of connections. With such a configuration, the flow control device 20 can be controlled based on the number of requests that can be processed by the processing system 16 per unit time instead of the number of simultaneous connections.

  The unit time and the processed request upper limit value (that is, the number of requests that can be accepted within the unit time) can be arbitrarily set according to the characteristics of the processing system. For example, a combination of up to 10 requests within a unit time of 1 second or up to 200 requests within a unit time of 60 seconds is possible.

  As described above, when the number of received requests and the number of processed requests are counted within a unit time, naturally, they may not match each other. It can be determined that this difference occurs because the processing is performed over a plurality of unit times. Therefore, since the connection with the processing system corresponding to the request related to this difference is not released, even if the upper limit value for accepting requests is limited considering only the number of processed requests, the processing system may not be busy. Absent.

  4 and 5 show a configuration in which the request type is determined, and the flow rate control according to the request type can be performed. To achieve this, 1. 1. A method of storing a request in a different queue immediately after receiving a request; There are two types of methods: queuing requests regardless of their type and dividing them by type before passing them to the processing system.

  FIG. 4 corresponds to the first method of flow rate control according to the request type, and shows a configuration when the request type is determined by the request holding unit 32.

  The request holding unit 32 includes a type determination unit 40, a priority setting unit 44, a queue control unit 41, and a timing determination unit 46. The type determination unit 40 determines the type of request received by the reception request count unit 30. As an example, the request URL is compared with a character string pattern prepared in advance to determine the type of request. As used herein, the type refers to processes such as page browsing, purchase procedure, authentication, and settlement performed in the processing system, but is not particularly limited as long as it can be distinguished based on the request URL.

  The priority setting unit 44 holds the priority for each type determined by the type determination unit 40. This priority is set in advance by a system administrator or the like. For example, when the request is classified into three types A, B, and C in the type determination unit 40, the priority of A is set to 1, the priority of B is set to 2, and the priority of C is set to 3, and so on. Yes. The smaller the numerical value of the priority, the higher priority is processed by the processing system.

  The queue control unit 41 stores the request received from the received request count unit 30 in the dedicated queue 42 for each request type specified by the type determination unit 40. For example, all requests of type A are stored in queue A.

  The timing determination unit 46 determines, based on the priority, which type of queue among the requests held in the queue 42 for each type is to be passed to the simultaneous connection restriction unit 34. As described above, when the priority levels of the requests of types A, B, and C are 1, 2, and 3, respectively, the timing determination unit 46 starts with the simultaneous connection restriction unit 34 first from the request stored in the queue A. To pass to. When there are no more requests in the queue A, the requests in the queue B are subsequently passed to the simultaneous connection restriction unit 34. In this case, the simultaneous connection restriction unit 34 flows the request passed from the request holding unit 32 to the processing system 16 within a range not exceeding the number of simultaneous connections.

  Even when the above configuration is adopted, the number of processed requests counted by the processed request counting unit 24 need not be counted for each type.

  With the configuration as described above, it is possible to execute priority control for controlling the order in which the received request is transferred to the processing system based on the priority according to the type. That is, priority control according to a request type having a preset priority can be performed.

  FIG. 5 corresponds to the second method of flow control according to the request type, and shows a configuration when the request type is determined by the simultaneous connection restriction unit 34.

  The simultaneous connection restriction unit 34 includes a type determination unit 50, a priority setting unit 54, a queue control unit 51, and a timing determination unit 56. The type determination unit 50 determines the type of request queued in the request holding unit 32. The priority setting unit 54 holds the priority for each type determined by the type determination unit 50. These functions are the same as those already described in FIG.

  The queue control unit 51 stores the request received from the request holding unit 32 in the dedicated queue 52 for each request type specified by the type determination unit 50. Based on the priority, the timing determination unit 56 determines which of the types of requests held in the queue 52 for each type is to be passed to the processing system 16. As described above, when the priority levels of the requests of types A, B, and C are 1, 2, and 3, respectively, the timing determination unit 56 first passes the requests stored in the queue A to the processing system 16. Like that. When there are no more requests in queue A, the requests in queue B are subsequently passed to processing system 16.

  With the configuration as described above, it is possible to execute priority control for controlling the order in which the received request is transferred to the processing system based on the priority according to the type.

The number of simultaneous connections may be set for each request type. As an example, when the total number of simultaneous connections is “10”, the type A simultaneous connections may be set to a maximum of “5”. By doing so, it is possible to prevent an event in which only the request stored in the queue A continues to be passed to the processing system.
In addition, the unit time in the received request count unit 30 and the processed request count unit 24 and the upper limit value of requests that can be processed may be set for each type of request. For example, when the processing time in the processing system 16 differs greatly depending on the type of request, the processing efficiency in the processing system may be improved by appropriately adjusting the unit time and the number of received requests.

FIG. 6 is a flowchart illustrating the flow rate control process according to the present embodiment.
First, the reception request count unit 30 receives a request for the processing system 16 from each client terminal 12 via the network 14 (S10). The acceptance request counting unit 30 counts the number of accepted requests (S12). The request holding unit 32 holds the request received by the reception request counting unit 30 in the queue (S14). In this case, it may be held in a different queue for each request type. At this time, information for specifying the client terminal 12 that has transmitted each request, such as an IP address, may be held in a predetermined memory (not shown).

  The processed request count unit 24 counts the number of requests that have already been processed per unit time (S16). Then, it is determined whether or not the number of processed requests has reached a predetermined processed request upper limit value (S18). If the upper limit is reached (Y in S18), the error processing unit 26 The reception request count unit 30 is inquired to identify the client terminal 12 that has issued the request thereafter, and an error response indicating that the request cannot be received is transmitted to the client terminal 12 (S20). If the number of processed quests has not reached the upper limit (N in S18), S20 is skipped.

  The simultaneous connection restriction unit 34 counts the number of requests passed to the processing system 16 every unit time (S22), and is held in the request holding unit 32 so that the number of requests does not exceed the number of simultaneous connections. The flow control for passing the requests to the sequential processing system 16 is executed (S24). In the request holding unit 32, when requests are held in different queues for each type, priority control for determining which type of request is passed to the system may be executed.

  As described above, in the flow control device according to the present embodiment, the number of requests that can be accepted by the flow control device is different from the number of requests passed to the processing system. That is, the acceptance of requests is limited based on the number of requests that the processing system has actually finished processing. On the other hand, the delivery of the request to the processing system is processed based on the number of simultaneous connections with the processing system. As a result, requests are received only for the number of times the processing system has actually finished processing, so the requirements for the TPS determined with the orderer of the processing system and the number of requests received in the flow control device are limited. And become the same. Therefore, the system design becomes easy. Moreover, it is not necessary to make the processing system have excessive performance as compared with the case of controlling by the number of simultaneous connections as in the conventional flow control device.

  Even if requests arrive in a short period of time, the number of processed requests per unit time will not reach the upper limit if the number of requests being processed by the processing system is within the TPS range. . Accordingly, an error response is not returned from the error processing unit to the client terminal, but the request is queued and the request is sent to the processing system within the range of the number of simultaneous connections. Therefore, the capacity of the processing system can be effectively utilized even when requests are concentrated.

  The present invention has been described based on some embodiments. It is understood by those skilled in the art that these embodiments are exemplifications, and that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are also within the scope of the present invention. By the way.

  It should also be understood by those skilled in the art that the functions to be fulfilled by the constituent elements described in the claims are realized by the individual functional blocks shown in the present embodiment or their linkage.

  When requests are held in different queues for each request type, the request to be passed to the processing system may be determined according to the processing time for each request type in the processing system. As an example, the simultaneous connection restriction unit 34 includes a processing time holding unit and a request selection unit (not shown). The processing time holding unit calculates and holds the average processing time for each request type in the processing system. The request selection unit selects a request based on the request stored in the queue for each request type and the average processing time so that the number of requests that can be processed in the unit time by the processing system is maximized. With such a configuration, it is possible to increase the number of request processes per unit time in the processing system while securing the number of simultaneous connections.

It is a figure explaining the conventional flow control apparatus. 1 is an overall view of a system including a flow control device according to an embodiment of the present invention. It is a functional block diagram which shows the detailed structure of a flow control apparatus. It is a figure which shows the case where a request classification is determined in a request holding part. It is a figure which shows the structure in the case of determining a request classification in a simultaneous connection restriction | limiting part. It is a flowchart explaining the flow control processing concerning one embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Client terminal, 14 Network, 16 Processing system, 20 Flow control apparatus, 22 Setting part, 24 Processed request count part, 26 Error processing part, 30 Accepted request count part, 32 Request holding part, 34 Simultaneous connection restriction part, 40 Type determination unit, 42 queues, 50 type determination unit, 52 queues.

Claims (6)

A request holding unit that stores an information processing request received from a client terminal for a system that executes a predetermined process in a queue;
Based on the notification of the processing result from the system to the client terminal, a processed request count unit that counts the number of requests that have been processed in the system within a predetermined unit time;
An error processing unit that returns an error response to a client terminal that has issued a new request when the number of requests counted by the processed request count unit reaches a predetermined upper limit within the unit time;
When the error processing unit returns an error response , requests in the request holding unit are sent to the system so that the number of requests that can be processed simultaneously on the system and the number of simultaneous connections different from the upper limit value is not exceeded. A concurrent connection restriction section to pass,
A flow rate control device comprising: The request holding unit
A type determination unit that determines the type of the received request;
A queue control unit that stores a request in a queue for each request type;
A timing determination unit that determines the timing of passing a request to the system for each request type;
The flow control device according to claim 1, comprising: The simultaneous connection restriction unit includes:
A type determination unit that determines the type of the received request;
A queue control unit that stores requests received from the request holding unit in a queue for each type;
A timing determination unit that determines the timing of passing a request to the system for each request type;
The flow control device according to claim 1, comprising: A priority setting unit in which the priority of processing in the subsequent system is set for each request type;
The flow rate control apparatus according to claim 2, wherein the timing determination unit extracts a request from a queue that holds a request type having a high priority. The simultaneous connection restriction unit includes:
A processing time holding unit that calculates and holds an average processing time for each request type in the system;
A request selection unit that selects requests so that the number of requests that can be processed within a unit time in the system is maximized based on the requests stored in the queue for each request type and the average processing time;
The flow rate control device according to claim 2, further comprising: A request holding function for storing in a queue an information processing request received from a client terminal for a system that executes a predetermined process;
A processed request count function that counts the number of requests that have been processed in the system within a predetermined unit time based on a notification of a processing result from the system to the client terminal;
An error processing function for returning an error response to the client terminal that has issued a new request when the number of processed requests counted reaches a predetermined upper limit value within the unit time;
When the error processing function returns an error response, the number of requests that can be processed at the same time on the system and the requests stored in the queue are not exceeded so that the number of simultaneous connections that is different from the upper limit value is not exceeded. Connection control function to pass to
A flow control program to make a computer realize.

Images