JP3735022B2 - Request aggregator - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a request aggregation device that aggregates requests from a plurality of clients connected to a server via a network.
[0002]
[Prior art]
The client / server system is a computer system that performs distributed processing using a network. The server performs services such as file management, database management, data provision and processing, and communication functions in response to client requests. A plurality of clients connected to the server via a network requests a service from the server, receives a result of data processing performed by the server, and displays a result. Since each of the client and the server performs processing in a distributed manner, load distribution of computer processing can be achieved.
[0003]
[Problems to be solved by the invention]
Conventionally, when the number of requests (requests) from a client increases, the number of processes at the server increases by the amount of increase in requests, so the processing load on the server increases and the processing speed at the server decreases. In addition, if the number of requests increases too much, there is a problem that the server does not function in some cases. Further, when the number of requests from clients increases, there is a problem that the amount of network traffic to the server increases.
[0004]
The present invention has been made to solve the above problems, and has as its object to suppress the traffic volume in the network, reduce the load on the server, and make the response to the client faster.
[0005]
[Means for Solving the Problems]
The request aggregation device of the present invention includes a queue in which requests sent from a client connected to a server via a network to the server are registered in the order in which the requests arrive at the request aggregation device, and requests registered in the queue. The client information memory in which the information of the sent client is registered, the sending information memory in which the sending status of the request registered in the queue is registered to the server, and when a request is sent from the client, a request response number is given to this request And store it in the queue, register the information of the client that sent the request in the client information memory, and order the order of the requests stored in the queue so that the server that takes a short time to respond to the request starts Sort, in the queue registration A request control circuit that sends the received requests to the server in the order of registration, and the request control circuit includes a new request that is the same as the new request sent from the specific client to the server. The request is not registered in the queue, but the information of the specific client that sent the new request is registered in the client information memory as the client that sent the advance request, and the request response is the same as the server response number that identifies the response sent from the server. Information on the client that sent the number request is acquired from the client information memory, and responses are sent to all clients in the acquired client information.
According to the present invention, the same request sent from a plurality of clients is sent as one request to the server.
[0006]
In the above invention, the request control circuit, the order of the requests stored in the queue, so rearranged in short order of the time required for the server's response to the request, the time required for a server response to the request, is set in advance Just do it .
Further, in the above invention, a cache device that temporarily stores a response from the server together with request information requesting the response is newly provided, and the request control circuit stores the request sent from the client to the server in the cache device. If the request is the same as the request, the response stored together with the request in the cache device may be sent to the client. In this case, the cache device may erase the stored content after a predetermined time.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
<Embodiment 1>
FIG. 1 is a configuration diagram showing the configuration of the request aggregation device 100 according to the first embodiment of the present invention. The request aggregating apparatus 100 includes a queue 101 in which requests from clients such as the client A 201 are registered in the order in which the requests arrive at the request aggregating apparatus 100. The request aggregating apparatus 100 also includes a transmission information memory m1 in which information indicating whether a request has been transmitted from the request aggregating apparatus 100 is registered, and a client information memory m2 in which information of all clients that have transmitted the same request is registered. And.
[0008]
The request aggregation device 100 also includes a request control circuit 102 that performs information input / output control and various data management of the queue 101, the transmission information memory m1, and the client information memory m2.
The reception of a request from each client and the output of a response to each client are performed via the client side input / output unit 103. Similarly, sending a request to a server (not shown) and inputting a response from the server are performed via the server-side input / output unit 104.
[0009]
Registered in the queue 101 are a request from each client, a response number for uniquely determining a response from the server to the request, and a storage address in the transmission information memory m1 in which detailed information regarding the request is registered. . When the TCP / IP (Transmission Control Protocol / Internet Protocol) is adopted as a communication protocol, the response number is, for example, a combination of a transmission source IP address and a TCP (UDP) port number. A unique combination for each request registered in the queue 101 is determined by the request control circuit 102.
[0010]
In the transmission information memory m1, a registration time (time) of a request registered in the queue 101, a transmission identifier for identifying whether the request has been transmitted to the server, and a client in which information of the client A 201 that transmitted the request is registered. The storage address in the information memory m2 is registered.
In the client information memory m2, information on the client A 201 that sent the request and a storage address in the client information memory m2 in which the client information following the client information is registered are registered.
[0011]
When the TCP / IP is adopted as the communication protocol, the registered client information is, for example, IP header information, TCP header information, and arbitrary information of a TCP upper layer protocol. Arbitrary information of the upper layer protocol of TCP is a command such as “GET” and a file name when HTTP (Hypertext Transfer Protocol) is used for the upper layer protocol of TCP. When the request is a representative request in which requests from a plurality of clients are aggregated, information on the second and subsequent clients is registered in the client information following the client information.
[0012]
Further, the request control circuit 102 sends the request with the earliest registration time among the requests registered in the queue 101, that is, the request arriving at the request aggregation device 100 earliest to the server, and sets the sending identifier to “Yes”. To. Setting the transmission identifier to “Yes” indicates that the request has been transmitted to the server.
[0013]
Hereinafter, the operation of the request aggregation device 100 will be described.
First, when the request A from the client A 201 arrives at the request aggregating apparatus 100, the request control circuit 102 stores information including the request in the queue 101, the transmission information memory m1, and the client information memory m2 in the following procedure. sign up.
[0014]
The request control circuit 102 first determines whether or not the request A is already registered in the queue 101. Here, since the request is not yet received in the request aggregating apparatus 100, the request A is not in the queue 101, and the request control circuit 102 determines that there is no request A. As a result of this determination, the request control circuit 102 first registers the request A, the response number 3033, and the address m1-0 of the transmission information memory m1 in the queue 101.
[0015]
Next, the request control circuit 102 registers the registration time 8:30, the transmission identifier No, and the address m2-100 of the client information memory m2 in the storage area indicated by the address m1-0 of the transmission information memory m1.
Next, the request control circuit 102 registers the information of the client A201 in the storage area indicated by the address m2-100 of the client information memory m2.
As described above, each information is registered as shown in FIG.
[0016]
Next, as shown in FIG. 3, when the request A arrives at the request aggregating apparatus 100 from the other clients B202 and C203, the request control circuit 102 determines whether or not the information of the request A already exists in the queue 101. judge. Here, since the request A is already registered in the queue 101, the request control circuit 102 determines that there is information on the request A. As described above, when the same request already exists, the request control circuit 102 does not register the request A from the client B 202 and the client C 203 in the queue 101 and uses the client B 202 as information following the information of the client A 201. , The information of the client C203 is registered in the client information memory m2. As a result, as shown in FIG. 3, each piece of information regarding the new request is added.
[0017]
When a response to the request A (response number 3033) from a server not shown in FIGS. 1 to 3 arrives at the request aggregating apparatus 100, the request control circuit 102 first sends the same response number (3033) from the queue 101. Extract information. Next, the request control circuit 102 extracts the address m1-0 of the transmission information m1 from the response number 3033, and the address m2-100 of the client information memory m2 registered in the storage area in the transmission information memory m1 indicated by this address. To get.
[0018]
As a result, the request control circuit 102 receives the client A201 information that is client information indicated by m2-100 in the client information memory m2, the information of the client B202 following the client A201, and the information of the client C203 following the client B202. The response from the server is copied to each client based on the extracted information, and sent to each client. After sending the response from the server, the request control circuit 102 deletes the information in the queue 101 relating to the request A, the sending information memory m1, and the client information memory m2.
[0019]
FIG. 4 is a configuration diagram illustrating a configuration example of a client-server network using the request aggregation device 100 according to the first embodiment.
Requests sent from the client A 201 and the clients B 202 to n 204 connected to the network to the server 301 such as a Web server or an FTP server are received by the request aggregating apparatus 100, and the same request for the server 301 is received as one representative. Send out as a request. As a result, the number of requests to the server 301 can be reduced, and the amount of traffic to the server 301 can be reduced.
[0020]
FIG. 5 is a configuration diagram illustrating a configuration example when the request aggregation device 100 is mounted on the gateways of the network 1 and the network 2. Even in such a configuration, the request aggregating apparatus 100 collectively sends the same request to the server 301 as a representative request, so that the amount of traffic from the network 1 to the network 2 can be reduced.
When a response from the server is sent to a predetermined client, the information in the queue 101, the send information memory m1, and the client information memory m2 are deleted. Therefore, when a request A is sent from another client thereafter, necessary information is again registered in the queue 101, the sending information memory m1, and the client information memory m2, and the request A is sent again to the server. Become.
[0021]
Note that a content addressable memory (CAM) may be used for the queue 101. The associative memory is a memory that does not indicate the location of the memory as an address (address) but can give the memory content as search information and convert the memory location where the content matches to the address, and searches for a part of the data When given as information, information related to this is taken out associatively. Therefore, for example, if the request A has already been stored, the request A that has already been registered is taken out when the request A is given. Thus, the request control circuit 102 determines whether or not the request A has already been registered. it can.
[0022]
<Embodiment 2>
Next, a second embodiment of the present invention will be described.
In the present embodiment, the request control circuit 102 changes the status of the request registered in the queue 101 in accordance with the response time from the server to the request. FIG. 6 is an explanatory diagram showing the status of requests registered in the queue 101 of the request aggregation device 100. As shown in FIG. 6, in the queue 101, the requests are arranged in the order of registration time. In the following, it is assumed that the requests A, B, and C from the client respectively request the files A ′, B ′, and C ′ from the server. In addition, it is assumed that the response time of the server to the request can be calculated in advance or can be measured.
[0023]
In the present embodiment, it is assumed that it takes 2 seconds to download the files A ′ and B ′ from the server, and it takes 10 minutes to download the file C ′. In this case, the response time from the server to the C request is extremely long for the A and B requests.
In such a case, the request control circuit 102 sets the registration time of the first C (request) registered in the transmission information memory m1 to the third request rank, and advances the registration time of the requests A and B.
[0024]
By changing the request order, the average response time to the clients that have requested the requests A and B is shortened. In the case of the present embodiment, the response to the client that sent the request C is 4 seconds longer and 10 minutes and 4 seconds, but the actual delay with respect to the expected value of 10 minutes from the server is large. It doesn't matter because it doesn't change. As described above, according to the present embodiment, the response time from the server to requests can be leveled.
[0025]
<Embodiment 3>
Next, a third embodiment of the present invention will be described. FIG. 7 is a configuration diagram illustrating a configuration of the request aggregation device 100 according to the third embodiment. As shown in FIG. 7, the request aggregation device 100 according to the present embodiment includes a cache device 105. The cache device 105 may be built in the request aggregation device 100 or may be connected and used as an external peripheral device. The cache device 105 registers a request from the client and a response from the server. In the case of FIG. 7, the response to the request A is A ′, and the response to the request B is B ′.
[0026]
Next, the request control circuit 102 determines whether the request from the client that subsequently arrives is the same as the request already in the cache device 105. If it is determined that the requests are the same, the request control circuit 102 registers in the cache device 105 with the client that sent the subsequent arrival request without sending the subsequent arrival request to the server. Send a response.
[0027]
In the case of FIG. 7, if the request that arrives subsequently is A, a response A ′ in the cache device 105 is sent to the client that sent the request. If the request that arrives subsequently is B, a response B ′ in the cache device 105 is sent out.
Note that the cache device 105 may erase the stored content after a predetermined time. Further, the erasure of the cache device 105 may be controlled from the outside.
[0028]
As described above, according to the request aggregating apparatus 100 of the present embodiment, a combination of a request that has already occurred and this response is cached, and when the same request occurs, the cached response is sent out. As a result, the number of requests from the client to the server can be reduced, the traffic can be reduced, and the response to the client can be made faster.
[0029]
<Embodiment 4>
Next, a fourth embodiment of the present invention will be described.
FIG. 8 is a configuration diagram illustrating a configuration of the request aggregation device 100 according to the fourth embodiment. Thereafter, as shown in FIG. 8, when a request A having a response number of 3033 is registered in the queue 101 for the common request A from the clients A 201, B 202, and C 203, and the request A has already been sent to the server. Is described.
[0030]
When an arbitrary request A from another client D (not shown) arrives at the request aggregating apparatus 100 after the request A is sent to the server, in the case of FIG. It is determined that the same request A (representative request A) exists. When the transmission identifier in the request information memory m1 related to the request A that has been determined to be already registered in the queue 101 is Yes, the request control circuit 102 uses the information related to the client D as the client information related to the representative request A in the client information memory. Register at m2.
[0031]
As a result, as shown in FIG. 9, the information of the client D is registered in the information address of the client following the information of the client C in the client information memory m2, and the address (m2-103) is registered. In this case, since the information in the queue 101, the transmission information memory m1, and the client information memory m2 has not yet been deleted, this is a stage before sending a response from the server to the client.
Therefore, if a response to the representative request A is sent from the server, this response is also duplicated and sent to the client D, so that the number of requests sent to the server can be reduced and the traffic volume can be reduced. Can do.
[0032]
【The invention's effect】
As described above, according to the present invention, since the same request sent from a plurality of clients is sent to the server as one request, the amount of traffic in the network is kept low, and the load on the server is reduced. It is possible to obtain an excellent effect that the response to the client is reduced and the response is made faster.
[Brief description of the drawings]
FIG. 1 is a configuration diagram illustrating a configuration of a request aggregation device according to an embodiment of the present invention.
FIG. 2 is a configuration diagram showing a state of a request aggregating apparatus according to an embodiment of the present invention.
FIG. 3 is a configuration diagram illustrating another state of the request aggregation device according to the embodiment of the present invention.
FIG. 4 is a configuration showing a network configuration using the request aggregation device of the embodiment.
FIG. 5 is a configuration showing a network configuration using the request aggregation device of the embodiment.
FIG. 6 is a configuration diagram illustrating a queue state in a request aggregation device according to another embodiment of the present invention.
FIG. 7 is a configuration diagram showing a configuration of a request aggregation device according to another embodiment of the present invention.
FIG. 8 is a configuration diagram showing a configuration of a request aggregation device according to another embodiment of the present invention.
FIG. 9 is a configuration diagram showing a configuration of a request aggregation device according to another embodiment of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 100 ... Request aggregation apparatus, 101 ... Queue, 102 ... Request control circuit, 103 ... Client side input / output part, 104 ... Server side input / output part, 201 ... Client A, m1 ... Transmission information memory, m2 ... Client information memory.

Claims (4)

A request aggregating apparatus that is arranged between a plurality of clients and servers connected via a network and aggregates requests sent from the clients to the server,
A queue in which requests sent from the client to the server are registered in the order in which they arrived at the request aggregation device;
A client information memory in which information of a client that has sent a request registered in this queue is registered;
A transmission information memory in which a transmission state of the request registered in the queue to the server is registered;
When a request is sent from the client, a request response number is attached to the request and the request is stored in the queue, information on the client that sent the request is registered in the client information memory, and the request stored in the queue A request control circuit that rearranges the order of the requests so that the server takes a short time to respond to the request to the top, and sends the requests registered in the queue to the server in the order of registration,
The request control circuit includes:
If the same advance request as the new request sent from the specific client to the server is already registered in the queue, the new request is not registered in the queue, and the information of the specific client that sent the new request is Register in the client information memory as the client that sent the advance request,
Information of the client that sent the request with the same request response number as the server response number that identifies the response sent from the server is acquired from the client information memory, and all the clients in the acquired client information are obtained. A request aggregating apparatus that sends the response. The request aggregation device according to claim 1,
The time required for the server to respond to the request is preset, and the request aggregating apparatus is characterized in that: The request aggregation device according to claim 1 or 2 ,
A cache device for temporarily storing a response from the server together with request information for requesting the response;
When the request sent from the client to the server is the same as the request stored in the cache device, the request control circuit sends a response stored together with the request in the cache device to the client. A request aggregating apparatus. 4. The request aggregating apparatus according to claim 3 , wherein the cache device erases stored contents after a predetermined time .

Images