JP2012155564A - Server cloud network for communication nodes - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently use server resources in a network comprising servers controlling communication in the network and routers receiving links as communication nodes.SOLUTION: A server cloud network 1 for communication nodes is configured to satisfy following three points: (1) a server cloud 100 virtualizing all servers 110 is to be established, and connections of the servers 110 in the server cloud 100 is to be performed by using a router network 150 comprising existing routers 140 that receive lines or links to which user terminals 160 are connected, and the links. Namely, a dedicated network is not to be established anew for creating a server cloud; (2) a principal address S is to be set for the server cloud 100; (3) all of the routers 140 are to be connected to the server cloud 100 by using the principal address S.

Description

  The present invention relates to a technique for efficiently using server resources in a network in which a server that controls communication in a network and a router that accommodates a link are used as communication nodes.

Some networks handled by NGN (Next Generation Network) or the like include a server that controls communication in a network and a router that accommodates a link as communication nodes (see Non-Patent Document 1).
The router accommodates the terminal used by the user and executes routing control, converts the destination address received from the terminal into an address addressed to the server, eliminates access from non-registered users, Or implement congestion control. Moreover, since the router accommodates the link, the installation location is limited.

The server is communicably connected to a plurality of routers installed in a predetermined narrow area (region), and controls communication in the network. For example, when the network is an IP (Internet Protocol) network, the server is a SIP (Session Initiation Protocol) server that performs call control, a RADIUS (Remote Authentication Dial In User Service) server that performs authentication, or a DNS that performs address resolution. (Domain Name System) server, and application servers that control various services.
Further, the server is installed with necessary resources in advance in accordance with the amount of processing load predicted (including not only short-term time fluctuation prediction but also medium- to long-term demand prediction). At that time, the server sets resources so as to have a sufficient margin so that it can cope with a predicted fluctuation in processing load. Therefore, a state where the server resources are not necessarily used efficiently has occurred.

Hideki Kasahara, et al., “Network infrastructure technology to support next generation networks”, NTT Technical Journal, Vol.19, No.4, p.38-43, (2007)

  Therefore, an object of the present invention is to efficiently use server resources in a network in which a server that controls communication in a network and a router that accommodates a link are used as communication nodes.

  According to the present invention, a server cloud network of a communication node configured with a plurality of servers that control communication in a network and a plurality of routers that accommodate links as communication nodes, virtualizes the plurality of servers to form a single server A server cloud in which one representative address is set as a destination to be used when communicating with the router, a line to which a terminal used by a user is connected, or the router accommodating the link and the link And a router network configured to connect the terminal and the server in a communicable manner.

  According to such a configuration, the server cloud network of the communication node includes a server cloud that virtualizes servers and functions as one server. Therefore, in the server cloud, it is possible to share the processing of a server with a large processing load by other servers, and the server resources can be used efficiently. Further, since a router network composed of a router accommodating a terminal and a link is used for connection for a server in the server cloud, it is not necessary to newly construct a server network in the server cloud. Therefore, it is possible to reduce the cost for constructing a server cloud network of communication nodes. Further, since one representative address is set in the server cloud, any router can always connect to the server cloud.

  In the server cloud network of the communication node according to the present invention, the server cloud includes a load balancer that allocates processing to the server, and the load balancer sets the representative address. .

  According to such a configuration, the load balancer in the server cloud allocates processing to each server. At this time, if the load balancer has a function of monitoring the processing load of each server and allocating the processing to a server with a low load according to the processing load in the server cloud, server resources can be used efficiently.

  Further, in the server cloud network of the communication node, the present invention is configured such that the router belongs to any of a plurality of groups, the server cloud is provided for each group, and each of the server clouds sets a different representative address. It is characterized by that.

  According to such a configuration, even if each of the server cloud network of communication nodes divided into a plurality is operated by different operators, in each server cloud, a server with a large processing load Processing can be shared by other servers, and server resources can be used efficiently.

  ADVANTAGE OF THE INVENTION According to this invention, a server resource can be used efficiently in the network comprised as a communication node the server which controls the communication in a network, and the router which accommodates a link.

It is a figure which shows the present network. It is a figure showing the conceptual diagram of the server cloud-ized network of the communication node of this invention. It is a figure which shows the structural example of the server cloud in this embodiment, (a) represents the case where one load balancer is provided, (b) represents the case where a load balancer is provided in a hierarchical structure. It is a figure which shows the structural example provided with the server cloud of this embodiment for every area. It is a figure which shows the structural example of the router of this embodiment. It is a figure which shows the example of a process sequence at the time of the terminal registration in a SIP server in the server cloud-ized network of the communication node of this embodiment. It is a figure which shows the process sequence example for the session setting after terminal registration.

  A mode for carrying out the present invention (hereinafter referred to as “embodiment”) will be described in detail with reference to the drawings as appropriate. In the following description, the case where the network is an IP network and the server is a SIP server will be described as an example. However, the present invention is not limited to this, and the network is a protocol different from IP and the server corresponds to the protocol in the network. Various servers that control the communication may be used.

  The outline of the present embodiment will be described with reference to FIGS. FIG. 1 shows an example of a current network configured with a server and a router as communication nodes. Moreover, FIG. 2 represents the conceptual diagram of the server cloud-ized network of the communication node in this embodiment. 1 and 2 each show a map of Japan. However, the map is not limited to Japan, and may be a narrow area in one region or a wide area across the world.

  The server 110 shown in FIG. 1 executes communication control in the network, and the router 140 accommodates a terminal 160 and a link used by the user. In the current network shown in FIG. 1, the server 110 and the router 140 are installed with predetermined resources at predetermined locations, and are operated in a fixed manner. The router 140 and the server 110 to which the router 140 is connected are determined in advance as a pair. The server 110 stores registration information of the terminal 160 accommodated by the router 140 determined as a pair of the server 110. The terminal 160 is, for example, a PC (Personal Computer), and transmits and receives packets. In FIG. 1, the relay router and the relay server that transfer the packet via the link are not shown, but this relay router is also included in the router 140, and the relay server is also included in the server 110. To do.

  When a terminal 160 communicates with another terminal 160, the terminal 160 registers (REGISTER) the terminal 160 in advance with the server 110 to which the terminal 160 should connect before performing communication. At this time, the terminal 160 acquires the address of the router 140 instead of the address of the server 110 by DHCP (Dynamic Host Configuration Protocol). When the router 140 receives a packet from the terminal 160, the router 140 converts the destination of the received packet from the address of the router 140 to the address of the server 110 and transmits the packet to the server 110.

  In the server 110, resources are set so as to have a sufficient margin so as to cope with a predicted change in processing load. Therefore, in the current network, the server 110 does not necessarily use the resource efficiently.

  FIG. 2 is a conceptual diagram of the server cloud network 1 of communication nodes according to the present embodiment, and the same components as those in FIG. In FIG. 2, the relay router and the relay server that transfer the packet via the link are not shown, but this relay router is also included in the router 140 and the relay server is also included in the server 110. To do.

  In the server cloud network 1 of communication nodes, the installation location is limited geographically because the router 140 accommodates a link among the communication nodes. However, the server 110 may be installed anywhere as long as communication can be performed between the router 140 and another server 110 in order to control communication in the network.

Therefore, the server cloud network 1 of the communication nodes is configured to satisfy the following three points.
(1) A cloud in which all servers 110 are virtualized (hereinafter referred to as server cloud 100) is constructed, and the connection for the server 110 in the server cloud 100 is an existing router that accommodates the terminal 160 used by the user. This is performed using a router network 150 including 140 and a link. In other words, a dedicated network is not newly constructed to make a cloud.
(2) A representative address S is set in the server cloud 100.
(3) All routers 140 are connected to the server cloud 100 using the representative address S.

Here, server virtualization will be described.
Instead of assigning one OS (operating system) or application to one server (processor) as a means for efficiently using the server, for example, one server (processor) is used using software called a hypervisor. Is used to operate a plurality of OSs and applications by virtually appearing to a plurality of servers (processors) having the same specifications. With this technology, by operating different types of OS on a single server, or by operating multiple programs with low processor usage efficiency, processing performed on multiple servers can be performed efficiently on a single server. Can be processed well. It is also possible to virtualize all of a plurality of servers, combine them into one, and make it appear as one virtually large server.

  That is, the server cloud network 1 of the communication node is provided by the server 110 that was a pair of routers 140 without being aware of the server 110 that is a pair of each router 140 as in the current network of FIG. The service can be received from the server cloud 100 (for example, SaaS: Software as a Service). Note that SaaS does not mean that the operator of the server cloud 100 provides only the resources of the server 110, but the operator sets up an OS or application on the server resource 100, and the application is sent to the user via the network. A service that can be used.

  Next, a configuration example of the server cloud 100 in the present embodiment will be described with reference to FIG. FIG. 3A shows a case where one load balancer is provided, and FIG. 3B shows a case where the load balancer is provided in a hierarchical structure. 3A and 3B, the terminal 160 is not shown.

  In FIG. 3A, the server cloud 100 includes a single load balancer 130 and a plurality of servers 110. The load balancer 130 has a function of determining a server 110 that executes response processing for a packet received from the router 140. Although only one load balancer 130 is shown in FIG. 3A, a plurality of units may be shown as one by applying the server virtualization technology described above.

  The load balancer 130 is set with the representative address S and receives a packet from the router 140. Then, the load balancer 130 mechanically allocates the servers 110 in charge in the order of processing, and determines which server 110 in the server cloud 100 is in charge of processing. In the determination of the server 110, the load balancer 130 may allocate processing to the server 110 having a low processing load in consideration of the processing load of each server 110. In addition, the load balancer 130 transmits the received packet to the server 110 that is in charge of processing.

The server 110 in the server cloud 100 transmits the response information corresponding to the received packet using the address of the router 140 that accommodates the terminal 160 as the address to be notified to the terminal 160. Note that the address used as the destination of the response information may be added by the router 140 instead of the server 110.
When the terminal 160 accesses the server cloud 100, the router 140 converts the destination of the packet received from the terminal 160 (the address of the router 140) into the representative address S of the server cloud 100 and receives the received packet. Is transmitted to the server cloud 100. For the conversion, for example, NAT (Network Address Translation) technology may be used.

  Since the server 110 in the server cloud 100 stores the registration information of the terminal 160, it can execute a response process for a packet received from any terminal 160.

  Next, a configuration example in the case where the load balancer is provided in a hierarchical structure will be described with reference to FIG. FIG. 3B differs from FIG. 3A in that load balancers 132 are further provided in a hierarchical manner between the load balancer 131 and the server 110 to which representative addresses are set. However, the function of the load balancers 131 and 132 is to determine the server 110 that executes a response process for the packet received from the router 140. The load balancer 132 is for distributing the load of the load balancer 131. Note that wiring between the load balancers 131 and 132 is newly implemented. 3B shows the case where the number of hierarchies is two, the number of hierarchies may be three or more. Since the configuration other than the difference is the same as that of FIG. 3A, the same reference numerals are given and detailed description is omitted.

  In FIG. 3B, the address that can be seen outside the server cloud 100 is the representative address S of the load balancer 131. Therefore, when the terminal 160 accesses the server 110, the router 140 converts the destination of the packet received from the terminal 160 (the address of the router 140) into the representative address S of the server cloud 100, and converts the received packet What is necessary is just to transmit to the server cloud 100. Note that the load balancer 131 and the load balancer 132 may be combined to realize the function of one load balancer as shown in FIG.

Next, a configuration example in which the server cloud 100 of this embodiment is provided for each region will be described with reference to FIG. FIG. 4 shows a case where there are two regions (groups). For example, the network administrator is different between the east region and the west region. Note that the description of the terminal 160 is omitted.
In FIG. 4, the router 140 is arranged to belong to any region (group). In addition, a server cloud 100 is configured for each region (group). The configuration in the server cloud 100 is the same as the configuration shown in FIG. However, the load balancer 130 of each server cloud 100 sets different representative addresses S1 and S2.

When the terminal 160 installed in the west region accesses the server cloud 100, the router 140 converts the destination of the packet received from the terminal 160 (the address of the router 140) into the representative address S1 of the load balancer 130. The received packet is transmitted to the server cloud 100. Further, when the terminal 160 installed in the east region accesses the server cloud 100, the router 140 sets the destination of the packet received from the terminal 160 (the address of the router 140) to the representative address S2 of the load balancer 130. The converted packet is transmitted to the server cloud 100.
In addition, although FIG. 4 demonstrated the case where there exist two areas (groups), even if it divides into three or more areas (groups), what is necessary is just to set a different representative address for every server cloud 100. FIG.

Next, a configuration example of the router 140 used in the server cloud network 1 of the communication node according to the present embodiment will be described with reference to FIG. 5 (see FIG. 3A as appropriate).
The router 140 includes a processing unit 410, a communication unit 420, and a storage unit 430.
The processing unit 410 is configured by a CPU (Central Processing Unit) and a main memory (not shown), and develops an application program stored in the storage unit 430 in the main memory to realize at least the function of the address conversion unit 411. . In addition, the processing unit 410 performs control such as data transmission / reception in the communication unit 420 and the storage unit 430. However, description of functions such as routing control and congestion control in the processing unit 410 is omitted.
The communication unit 420 is an interface connected to the network, and transmits / receives packets to / from the terminal 160, the other router 140, and the load balancer 130.
The storage unit 430 stores at least the application program and the representative address 431 of the server cloud 100.

Next, an example of a processing sequence at the time of terminal registration in the SIP server in the server cloud network 1 of the communication node according to the present embodiment will be described with reference to FIG.
When communicating between the terminals 160A and 160B using SIP, the terminals 160A and 160B perform registration (RESISTER) in advance with the server 110 to which the terminals 160A and 160B should connect before performing communication. FIG. 6 shows an example of the registration processing sequence.

  Terminals 160A and 160B acquire the addresses of routers 140A and 140B, respectively, by DHCP. Then, the router 140A converts the destination address of the packet received from the terminal 160A into the representative address S of the server cloud 100, and transmits the packet to the server cloud 100. Further, the router 140B converts the destination address of the packet received from the terminal 160B into the representative address S of the server cloud 100, and transmits the packet to the server cloud 100.

In step S501, the terminal 160A transmits a packet for registration (RESISTER) to the router 140A (address: R1).
In step S502, the router 140A converts the destination address R1 of the received packet into the representative address S of the server cloud 100.
In step S503, the router 140A transmits the registration (RESISTER) packet received from the terminal 160A to the server cloud 100 (address: S).
In step S504, the server cloud 100 transmits ACK to the terminal 160A (address: UA-A) as response information to the reception of the registration (RESISTER) packet.

In step S511, the terminal 160B transmits a registration (RESISTER) packet to the router 140B (address: R2).
In step S512, the router 140B converts the destination address R2 of the received packet into the representative address S of the server cloud 100.
In step S513, the router 140B transmits the registration (RESISTER) packet received from the terminal 160B to the server cloud 100 (address: S).
In step S514, the server cloud 100 transmits ACK to the terminal 160B (address: UA-B) as response information to the reception of the registration (RESISTER) packet.

Next, an example of a processing sequence for session setting after terminal registration will be described with reference to FIG.
In step S601, the terminal 160A transmits an INVITE request to the router 140A (address: R1).
In step S602, the router 140A receives the INVITE request, and converts the destination address R1 of the received packet into the representative address S of the server cloud 100.
In step S603, the router 140A transmits the received INVITE request to the server cloud 100 (address: S).

In step S604, the server cloud 100 transmits an INVITE request to the terminal 160B (address: UA-B).
In step S605, the server cloud 100 receives the INVITE request and returns a response (Trying) indicating that processing is in progress to the terminal 160A (address: UA-A).

In step S606, the terminal 160B transmits a response (Ringing) for notifying the transmission source that the call is being made to the router 140B (address: R2).
In step S607, the router 140B converts the destination address R2 of the received response (Ringing) into the representative address S of the server cloud 100.
In step S608, the router 140B transmits the received response (Ringing) to the server cloud 100 (address: S).
In step S609, the server cloud 100 transmits a response (Ringing) to the transmission source terminal 160A (address: UA-A).

In step S610, terminal 160B returns a response (OK) for notifying the transmission source that the INVITE request has been received, to router 140B (address: R2).
In step S611, the router 140B converts the destination address R2 of the received response (OK) into the representative address S of the server cloud 100.
In step S612, the router 140B transmits the received response (OK) to the server cloud 100 (address: S).
In step S613, the server cloud 100 transmits a response (OK) to the terminal 160A (address: UA-A) that is the transmission source.

In step S614, terminal 160A transmits to terminal 160B (address: UA-B) a response (ACK) for confirming that the response (OK) has been received.
In step S615, terminal 160A and terminal 160B establish a session between them (Media Session) and exchange communication information with each other.

  As described above, the server cloud network 1 of the communication node according to the present embodiment includes the server cloud 100 by virtualizing all the servers 110, and the connection for the server 110 in the server cloud 100 is connected to the router 140 that accommodates the terminal 160. Connection is made using a router network 150 configured with links. That is, a dedicated network is not newly constructed for the server 110 to be clouded. Further, the server cloud 100 is provided with a load balancer 130 having a function of distributing the processing load, and a representative address S is set in the load balancer 130. Therefore, all the routers 140 are connected to the server cloud 100 using the representative address S. And since the server 110 in the server cloud 100 can distribute the processing load by the load balancer 130, the server cloud network 1 of the communication node can efficiently use the server resources. Moreover, since it is not necessary to construct a new network for the server in the server cloud 100, the cost can be reduced.

DESCRIPTION OF SYMBOLS 1 Server node network of communication nodes 100 Server cloud 110 Servers 130, 131, 132 Load balancer 140 Router 150 Router network 160 Terminal 410 Processing unit 411 Address conversion unit 420 Communication unit 430 Storage unit 431, S, S1, S2 Representative address

Claims (3)

A server cloud network of communication nodes configured with a plurality of servers that control communication in the network and a plurality of routers that accommodate links as communication nodes,
A server cloud that functions as one server by virtualizing a plurality of the servers, and sets one representative address as a destination used when communicating with the router;
A communication node comprising a router connected to a line or a link to which a terminal used by a user is connected, or the link, and a router network that connects the terminal and the server in a communicable manner. Server cloud network. The server cloud is
A load balancer that allocates processing to the server;
2. The communication node server cloud network according to claim 1, wherein the representative address is set in the load balancer. The router belongs to one of a plurality of groups;
The server cloud is provided for each group,
2. The server cloud network of communication nodes according to claim 1, wherein each of the server clouds sets a different representative address.

Images