JP2011166384A - Computer system and communication method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve fault resistance of a computer system using open flow technique. <P>SOLUTION: The computer system includes: a plurality of controllers 1 each of which calculates communication routes and indicates the setting of flow entry with a priority 147 added thereto to a switch on the communication route; and the switches 2 each determining whether the setting of the flow entry is permitted or not in accordance with the priority 147, and performing a relay operation corresponding to the flow entry set in itself with respect to a reception packet meeting the flow entry set in itself. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a computer system and a communication method, and more particularly to a computer system redundancy technology and a communication method using open flow (programmable flow) technology.

  Conventionally, the determination of the route from the packet source to the destination and the packet transfer processing have been performed by a plurality of switches on the route. In recent years, in a large-scale network such as a data center, a change in the network configuration has always occurred due to a stop of a device due to a failure or a new addition of a device for expanding the scale. For this reason, it has become necessary to have the flexibility to determine an appropriate route in response to changes in the network configuration. However, since the route determination processing program in the switch cannot be changed from the outside, the entire network cannot be controlled and managed centrally.

  On the other hand, in the computer network, a technique (open flow) for centrally controlling the transfer operation of each switch by an external controller has been proposed by the OpenFlow Consortium (see Non-Patent Document 1). A network switch (hereinafter referred to as a programmable flow switch (PFS)) that supports this technology holds detailed information such as protocol type and port number in a flow table, and can control the flow and collect statistical information. it can.

  The configuration and operation of a computer system using OpenFlow will be described with reference to FIG. Referring to FIG. 1, a computer system according to a related technique of the present invention includes a programmable flow controller 100 (hereinafter referred to as PFC 100), a plurality of programmable switches 102-1 to 102-n (hereinafter referred to as PFSs 102-1 to 102-n). And a host group 300 having a plurality of host computers 103-1 to 103-i (hereinafter referred to as hosts 103-1 to 103-i). However, n and i are natural numbers of 2 or more. Hereinafter, the PFSs 102-1 to 102-n will be collectively referred to as PFS102, and the hosts 103-1 to 103-i may be collectively referred to as the host 103.

  The PFC 100 sets a communication path between the hosts 103 and a transfer operation (relay operation) for the PFS 102 on the path. At this time, the PFC 100 sets a flow entry in which a rule for specifying a flow (packet data) and an action for defining an operation for the flow are associated with each other in a flow table held by the PFS 102. The PFS 102 on the communication path determines the transfer destination of the received packet data according to the flow entry set by the PFC 100, and performs transfer processing. As a result, the host 103 can transmit and receive packet data to and from other hosts 103 using the communication path set by the PFC 100. That is, in a computer system using OpenFlow, the PFC 100 that sets a communication path and the PFS 102 that performs transfer processing are separated, and communication of the entire system can be controlled and managed centrally.

  Referring to FIG. 1, when packet transmission is performed from the host 103-1 to the host 103-i, the PFS 102-1 refers to transmission destination information (header information) in the packet received from the host 103-1, and the PFS 102 -1 An entry that matches the header information is searched from the flow table held inside. The contents of entries set in the flow table are defined in Non-Patent Document 1, for example.

  When the entry for the received packet data is not described in the flow table, the PFS 102-1 transfers the packet data (hereinafter referred to as the first packet) or the header information of the first packet to the PFC 101. The PFC 101 that has received the first packet from the PFS 102-1 determines the path 400 based on information such as a transmission source host and a transmission destination host included in the packet.

  The PFC 101 instructs all PFSs 102 on the path 400 to set a flow entry that defines a packet transfer destination (issues a flow table update instruction). The PFS 102 on the path 400 updates the flow table managed by itself in response to the flow table update instruction. Thereafter, the PFS 102 starts forwarding the packet according to the updated flow table, so that the packet reaches the destination host 103-i via the path 400 determined by the PFC 101.

  On the other hand, in a system in which a transfer route is set by a switch, a technique for selecting and using an appropriate route according to priority from multiplexed routes is disclosed in, for example, Japanese Patent Application Laid-Open No. 2006-333135 (see Patent Document 1) and Japanese Patent Application Laid-Open No. 2000. -324154 (see Patent Document 2) and JP-A-4-215349 (see Patent Document 3).

JP 2006-333135 A JP 2000-324154 A JP-A-4-215349

OpenFlow Switch Specification Version 0.9.0 (Wire Protocol 0x98) July 20, 2009

  When any failure occurs in the PFC 100, the PFC 100 cannot instruct the PFS 102 on the route to update the flow table. In this case, the PFS 102 that has received the first packet continues to wait for a flow table update instruction from the PFC 101 until the timeout time set by itself, and cannot perform packet communication between the hosts 103 during that time. Further, unless the PFC 100 is recovered from a failure, transfer control corresponding to the first packet cannot be performed.

  For this reason, it is conceivable to prepare a plurality of PFCs 100 from the viewpoint of fault tolerance. However, a method of constructing a communication path using a plurality of PFCs (updating a flow table for each PFS) has not been established yet. For this reason, there is a problem that a computer system using the open flow technology has low resistance to a failure occurring in the PFC.

  Accordingly, an object of the present invention is to improve the fault tolerance of a computer system using open flow technology.

  Another object of the present invention is to perform early construction of a communication path according to a first packet and transfer processing of packet data even when a failure occurs in the programmable flow controller.

  In order to solve the above problems, the present invention employs the means described below. In the description of technical matters constituting the means, in order to clarify the correspondence between the description of [Claims] and the description of [Mode for Carrying Out the Invention] The number / symbol used in [Form] is added. However, the added numbers and symbols should not be used to limit the technical scope of the invention described in [Claims].

  A computer system according to the present invention includes a plurality of controllers (1) each of which calculates a communication path and instructs a switch (2) on the communication path to set a flow entry with priority (147) added thereto, A switch (2) that determines whether or not to permit setting of a flow entry in accordance with the priority (147), and performs a relay operation according to the flow entry that is set for the received packet that matches the flow entry that is set for itself It comprises.

  In the communication method according to the present invention, each of the plurality of controllers (1) calculates a communication path, and sets a flow entry with priority (147) added to the switch (2) on the communication path. A step of instructing, a step in which the switch (2) determines whether or not to permit setting of the flow entry according to the priority (147), and a step in which the switch (2) matches the flow entry set in itself. And performing a relay operation according to the flow entry set in itself.

  According to the present invention, it is possible to improve fault tolerance of a computer system using the open flow technology.

  Even when a failure occurs in the programmable flow controller, the construction of the communication path according to the first packet and the packet data transfer process can be executed at an early stage.

FIG. 1 is a diagram showing an example of the configuration of a computer system according to the prior art. FIG. 2 is a diagram showing a configuration in the embodiment of the computer system according to the present invention. FIG. 3 is a diagram showing a configuration in an embodiment of a programmable flow controller according to the present invention. FIG. 4 is a diagram showing an example of a flow table held by the programmable flow controller according to the present invention. FIG. 5 is a diagram showing an example of topology information held by the programmable flow controller according to the present invention. FIG. 6 is a diagram showing an example of communication path information held by the programmable flow controller according to the present invention. FIG. 7 is a diagram showing a configuration of the programmable flow switch according to the embodiment of the present invention. FIG. 8 is a diagram showing an example of a flow table held by the programmable flow switch according to the present invention. FIG. 9 is a diagram for explaining programmable flow control according to the present invention. FIG. 10 is a sequence diagram showing operations of communication path construction processing (flow table update processing) and packet transfer processing in the embodiment of the computer system according to the present invention. FIG. 11 is a flowchart showing the operation of the flow entry setting process (flow table update process) in the programmable flow switch according to the present invention. FIGS. 12A and 12B are diagrams showing an example of the flow table updated in the programmable flow switch according to the present invention. FIG. 13 is a diagram illustrating an example of communication paths calculated by a plurality of programmable flow controllers according to the present invention. FIG. 14A is a diagram illustrating an example of a communication path construction process and a packet transfer process when a failure occurs in a plurality of programmable flow controllers. FIG. 14B is a diagram illustrating an example of communication path construction processing and packet transfer processing when a failure occurs in a plurality of programmable flow controllers. FIG. 15A is a diagram illustrating another example of a communication path construction process and a packet transfer process when a failure occurs in a plurality of programmable flow controllers. FIG. 15B is a diagram illustrating another example of the communication path construction process and the packet transfer process when a failure occurs in a plurality of programmable flow controllers. FIG. 16 is a flowchart showing the operation of the communication path construction processing after the failure recovery of the programmable flow controller according to the present invention.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the drawings, the same or similar reference numerals indicate the same, similar, or equivalent components.

(Computer system configuration)
The configuration of the computer system according to the present invention will be described with reference to FIG. FIG. 2 is a diagram showing a configuration in the embodiment of the computer system according to the present invention. The computer system according to the present invention performs communication path construction and packet data transfer control using OpenFlow. Referring to FIG. 2, the computer system according to the present invention includes a programmable flow controller 1-1 to 1-m (hereinafter referred to as PFC 1-1 to 1-m), a plurality of programmable switches 2-1 to 2-n ( Hereinafter, a switch group 20 having PFS 2-1 to 2-n) and a host group 30 having a plurality of host computers 3-1 to 3-i (hereinafter referred to as hosts 3-1 to 3-i) are provided. To do. However, m, n, and i are natural numbers of 2 or more. When the PFCs 1-1 to 1-m are collectively referred to without distinction, they are referred to as PFC1, and when the PFSs 2-1 to 2-n are collectively referred to without distinction, they are referred to as PFS2, and the hosts 3-1 to 3-i Are collectively referred to as “host 3”.

  The host 3 is a computer device including a CPU, a main storage device, and an external storage device (not shown), and communicates with other hosts 3 by executing a program stored in the external storage device. Communication between the hosts 3 is performed via the switch group 20. The host 3 realizes a function exemplified in a Web server, a file server, an application server, a client terminal, or the like according to a program to be executed. For example, when the host 3 functions as a Web server, an HTML document or image data in a storage device (not shown) is transferred to another host 3 (example: client terminal) in accordance with a request from the other host 3 (example: client terminal). Forward to.

  The PFC 1 includes a switch control unit 11 that controls communication path packet transfer processing related to packet transfer in the system using the open flow technology. Open flow technology refers to a technology in which a controller (here PFC1) performs path control and node control by setting multi-layer and per-flow path information in PFS2 in accordance with a routing policy (flow entry: flow + action). (For details, see Non-Patent Document 1.) As a result, the route control function is separated from the routers and switches, and optimal routing and traffic management are possible through centralized control by the controller. PFS2 to which the open flow technology is applied handles communication as a flow of END2END, not as a unit of packet or frame like a conventional router or switch.

  The PFC 1 controls the operation of the switch and the node (for example, packet data relay operation) by setting a flow entry (rule + action) in the flow table 23 held by the PFS 2. The PFC 1 according to the present invention sets the flow entry to which the priority of the setting route (setting controller) is added to the PFS 2 on the route.

  Details of the configuration of the PFC 1 will be described with reference to FIG. FIG. 3 is a diagram showing a configuration of the PFC 1 according to the present invention. The PFC 1 is preferably realized by a computer including a CPU and a storage device. In the PFC 1, each function of the switch control unit 11, the flow management unit 12, and the flow generation unit 13 illustrated in FIG. 3 is realized by a CPU (not shown) executing a program stored in a storage device. The PFC 1 includes a flow table 14, topology information 15, and communication path information 16 that are stored in a storage device (not shown).

  The switch control unit 11 sets or deletes a flow entry (rule + action) for each PFS 2 according to the flow table 14. At this time, the switch control unit 11 associates the priority with the flow entry (rule + action information) and sets it in the flow table 23 of the PFS2. In addition, the update of the flow table for the PFS 2 (flow entry setting or deletion) is performed in order from the destination device side of the flow to the transmission source device. The PFS 2 refers to the set flow entry and executes an action (for example, relay or discard of packet data) corresponding to the rule according to the header information of the received packet. Details of the rules, actions, and priorities will be described later.

  FIG. 4 is a diagram illustrating an example of the configuration of the flow table 14 held by the PFC 1. Referring to FIG. 4, the flow table 14 includes a flow identifier 141 for specifying a flow entry, an identifier (target device 142) for identifying a setting target (PFS2) of the flow entry, path information 143, a rule 144, Action information 145, setting information 146, and priority 147 are set in association with each other. In the flow table 14, flow entries (rule 144 + action information 145) generated for all the PFSs 2 to be controlled by the PFC 1 are set. The flow table 14 may define how to handle communication, such as QoS and encryption information for each flow.

  The rule 144 defines, for example, combinations of addresses and identifiers of layers 1 to 4 of the OSI (Open Systems Interconnection) reference model included in header information in TCP / IP packet data. For example, each combination of a layer 1 physical port, a layer 2 MAC address, a layer 3 IP address, a layer 4 port number, and a VLAN tag (VLAN id) shown in FIG. The VLAN tag may be given a priority (VLAN Priority).

  Here, identifiers such as port numbers and addresses set in the rule 144 may be set within a predetermined range. In addition, it is preferable that the destination 144 and the address of the transmission source are distinguished and set as the rule 144. For example, the range of the MAC destination address, the range of the destination port number that identifies the connection destination application, and the range of the transmission source port number that identifies the connection source application are set as the rule 144. Furthermore, an identifier for specifying the data transfer protocol may be set as the rule 144.

  The action information 145 defines a method for processing TCP / IP packet data, for example. For example, information indicating whether or not the received packet data is to be relayed and the transmission destination in the case of relaying are set. The action information 145 may be set with information instructing to copy or discard the packet data.

  The route information 143 is information for specifying a route to which the flow entry (rule 144 + action information 145) is applied. This is an identifier associated with communication path information 16 described later.

  The setting information 146 includes information (“set” or “not set”) indicating whether or not the flow entry (rule 144 + action information 145) is currently set in the PFS2 on the communication path. Since the setting information 146 is associated with the target device 142 and the route information 143, it can be confirmed whether or not a flow entry is set for the communication route, and a flow entry is set for each PFS2 on the communication route. It can be confirmed whether or not.

  A priority 147 associated with each flow entry is set in the flow table 14 according to the present invention. The priority 147 may be set in advance by the user, or may be designated (set) by the PFS 2 that has received the first packet. Further, the priority 147 may be fixedly set to a value determined for each PFC 1 or may be changed according to the situation. For example, the PFS 2 may change the priority 147 specified for each of the PFCs 1-1 to 1-m each time the first packet is received. Alternatively, when 1 to m is set as the priority 147 for the PFCs 1-1 to 1-m (“m” is the lowest priority), a low value “m” is set for the PFC 1 having a large number of failure occurrences. When the priority 147 is set and the failure occurrence rate becomes low, the priority 147 to be set is raised to the upper level. However, when changing the priority 147 according to the situation, it is preferable to provide an integrated device in the system that monitors all the PFCs 1-1 to 1-m and can set the priority 147 by the user.

  Further, when the PFS 2 that has received the first packet sets the priority 147 for the PFC 1, the PFS 2 gives a high priority 147 to the PFC 1 (main controller) that mainly manages itself, and the PFC 1 that manages itself as a spare ( It is preferable to set a low priority for the sub-controller. For example, the OFC 1 that mainly manages other switch groups is selected as the sub-controller.

  The flow management unit 12 attaches the flow identifier 141 to the flow (rule 144 + action information 145) generated by the flow generation unit 13 and records it in the storage device. At this time, the identifier of the communication route to which the flow entry is applied (route information 143), the identifier of the PFS2 to which the flow entry is applied (target device 142), and the priority 147 are attached to the flow entry (rule 144 + action information 145). Recorded. When the priority 147 is notified from the PFS 2 together with the first packet, the flow management unit 12 sets the priority 147 in the flow table 14 in association with the flow entry generated by the flow generation unit 13. .

  Also, the flow management unit 12 refers to the flow table 14 and extracts the flow entry (rule 144 + action information 145) corresponding to the header information of the first packet and the priority 147 associated therewith, and performs switch control. Notification to the unit 11. The switch control unit 11 sets the notified flow entry (rule 144 + action information 145) and priority 147 in the target device 142 (PFS2) associated with the flow entry.

  The flow generation unit 13 calculates a communication path from the header information of the first packet notified from the PFS 2, and generates a flow entry (rule 144 + action information 145) to be set in the PFS 2 on the communication path. Specifically, the flow generation unit 13 specifies the transmission source host 3 and the destination host 3 from the header information of the first packet, calculates the communication path using the topology information 15, and calculates the calculation result as the communication path information 16. To the storage device. Here, the host 3 serving as the end point of the communication path, the PFS 2 on the communication path, and the connection relationship between them are set as the communication path information 16. Further, the flow generation unit 13 sets a flow entry (rule 144 + action information 145) to be set in the PFS 2 on the communication path based on the communication path information 16.

  FIG. 5 is a diagram showing an example of the topology information 15 held by the PFC 1 according to the present invention. The topology information 15 includes information related to the connection status of the PFS 2, the host 3, and the like. Specifically, as the topology information 15, the device identifier 151 that identifies the PFS 2 or the host 3 is associated with the port number 152 or the port connection destination information 153 of the device and recorded in the storage device. The port connection destination information 153 includes a connection type (switch / node / external network) that identifies a connection partner, information that identifies a connection destination (switch ID in the case of PFS2, MAC address in the case of a host, external network (example: Internet) ) Includes an external network ID).

  FIG. 6 is a diagram showing an example of communication path information held by the PFC 1 according to the present invention. The communication path information 16 is information for specifying a communication path. Specifically, as the communication path information 16, end point information 161 that designates the host 3 or an external network interface (not shown) as an end point, passing switch information 162 that designates a pair group of PFS 2 and a port to pass, and accompanying information 163. Are associated and recorded in the storage device. For example, when the communication path is a path connecting the hosts 3, the MAC address of each host 3 is recorded as the end point information 161. The passing switch information 162 includes an identifier of the PFS 2 provided on the communication path between the end points indicated by the end point information 161. The passing switch information 162 may include information for associating the flow entry (rule 144 + action information 145) set in the PFS2 with the PFS2. The accompanying information 163 includes information related to PFS2 (passage switch) on the route after the end point is changed.

  With the configuration as described above, the PFC 1 according to the present invention generates a flow entry for transferring the packet in response to the reception notification of the first packet from the PFS 2, and is notified from the generated flow entry and the PFS 2. The priority 147 (or preset priority 147) is set to the calculated PFS2 on the communication path.

  FIG. 7 is a diagram showing a configuration of the programmable flow switch according to the embodiment of the present invention. The PFS 2 determines a received packet processing method (action) according to the flow table 23 set (updated) by the PFC 1. The PFS 2 includes a transfer processing unit 21 and a flow setting unit 22. The transfer processing unit 21 and the flow setting unit 22 may be configured by hardware or may be realized by software executed by the CPU.

  A flow table 23 as shown in FIG. 8 is set in the storage device of the PFS2. The flow setting unit 22 sets the flow entry (rule 144 + action information 145) and priority 147 acquired from the PFC 1 in the flow table 23. Specifically, when the header information of the received packet does not match (or matches) the rule 144 recorded in the flow table 23, the flow setting unit 22 determines that the packet data is a first packet and receives the first packet. This is notified to PFC 1-1 to 1-m. At this time, the flow setting unit 22 assigns a priority 24 different for each PFC 1 to the first packet (or its header information), and transmits it to each of the PFCs 1-1 to 1-m. For example, the flow setting unit 22 transmits the first packet with the priority 24 “1” to the PFC 1-1 and the first packet with the priority 24 “m” lower than “1” to the PFC1-m. Send. In this case, it is preferable that the PFS 2 holds the priority 24 set in advance in the storage device.

  Then, the flow setting unit 22 sets the flow entry (rule 144 + action information 145) and priority 147 transmitted from each of the PFCs 1-1 to 1-m in response to the notification of the first packet in the flow table 23. Here, the transmitted priority 147 is a priority set (or assigned by PFS2) for each PFC1. When the same flow entry as the rule 144 transmitted from the PFC 1 exists in the flow table 23, the flow setting unit 22 compares the priority 147 transmitted from the PFC 1 with the priority 147 set in the flow table 23. Then, a flow entry with a high priority is set in the flow table 23. That is, the flow setting unit 22 determines whether or not the flow entry setting is permitted based on the transmitted priority 147 and the priority 147 set in the flow table 23 in response to the flow setting instruction from the PFC 1. Here, when the priority of the flow entry from the PFC 1 is high, the flow table 23 is overwritten by the flow entry. On the other hand, when the priority of the flow entry in the flow table 23 is high, the setting of the flow entry from the PFC 1 is rejected and the flow entry is discarded.

  In this way, since a flow entry with a high priority is preferentially overwritten and set, only one flow entry that matches the first packet is set in each of the PFSs 2-1 to 2-n.

  When the header information of the received packet matches (matches) the rule 144 recorded in the flow table 23, the packet data is transferred to another PFS 2 or the host 3 by the transfer processing unit 21. Specifically, the transfer processing unit 21 specifies action information 145 corresponding to the rule 144 that matches (or matches) the header information of the packet data. The transfer processing unit 21 transfers the packet data to the transfer destination node (PFS 2 or host 3) designated by the action information 145.

  Specifically, Rule 144: MAC source address (L2) is “A1 to A3”, IP destination address (L3) is “B1 to B3”, protocol is “http”, and destination port number (L4) is “C1”. ˜C3 ”and action information 145: The operation of the PFS 2-1 in which the flow entry in which“ Relay to PFS 2-2 ”is associated will be described. When packet data having a MAC source address (L2) of “A1”, an IP destination address (L3) of “B2”, a protocol of “http”, and a destination port number (L4) of “C3” is received, PFS2- 1 determines that the header information matches the rule 144 and transfers the received packet data to the PFS 2-2. On the other hand, when packet data having a MAC source address (L2) of “A5”, an IP destination address (L3) of “B2”, a protocol of “http”, and a destination port number (L4) of “C4” is received, The PFS 2-1 determines that the header information does not match the rule 144, notifies the PFC 1-1 to 1-m that the first packet has been received, and transmits the header information and the priority 24 to each PFC 1.

  It is preferable that the transfer processing of the first packet by the transfer processing unit 21 is performed with a flow entry set in the flow table 23 as a trigger. For example, when the transfer of the first packet is performed according to the flow entry with the priority 147 of “3” and then rewritten to the flow entry with the priority 147 of “1”, the next received packet is the flow entry after the rewrite. To the destination node according to

  As described above, in the computer system according to the present invention, a communication path for transferring a flow (packet) is constructed by a plurality of PFCs 3-1 to 3-i. Packet transfer is possible even if a failure occurs. Different priorities are assigned to each PFC 3, and the flow table of the PFS 2 is overwritten and updated by a flow entry having a higher priority. As a result, a plurality of action information 145 corresponding to the same rule 144 is not set in the PFS2, and one flow entry is always set for the flow. That is, even if a flow entry is set by a plurality of PFCs 3, there is always one flow entry set in PFS2, so that PFS2 can transfer a packet without malfunction.

  In the present invention, the flow table is updated in order from the PFS 2 on the destination host side toward the PFS 2 on the transmission source host side (PFS 2 that received the first packet). For this reason, even if the flow table 23 of the PFS 2 is overwritten by a flow entry having a high priority, the established communication path is always a path connecting the transmission source host and the destination host.

(Construction of communication path and packet transfer operation)
With reference to FIG. 10 to FIG. 12, the details of the communication path construction operation in the computer system according to the present invention will be described. Hereinafter, an operation of transmitting packet data from the host 3-1 to the host 3-i and establishing a communication path between them will be described. Here, it is assumed that the PFS 2-1 receives the first packet from the host 3-1.

  FIG. 10 is a sequence diagram showing operations of communication path construction processing (flow table update processing) and packet transfer processing in the embodiment of the computer system according to the present invention. Referring to FIG. 10, packet data addressed to host 3-i is transmitted from host 3-1, and is received by the nearest PFS 2-1 on host 3-1 side (step S101).

  The PFS 2-1 refers to the header information (transmission destination information) in the packet received from the host 3-1, and searches for the corresponding entry from the flow table 23 held in the PFS 2-1 (step S102). If there is a flow entry corresponding to the header information, the received packet is transferred to the transfer destination node based on the action information 145 defined by the entry (steps S102 No, S115). On the other hand, when there is no flow entry corresponding to the header information in the flow table 23, the PFS 2-1 determines that the first packet has been received, and notifies each of the PFCs 1-1 to 1-m that the first packet has been received (step S102 Yes). , S103, S104). At this time, the PFS 2-1 associates the first packet (or header information) and the priority 24 and transmits them to the PFCs 1-1 to 1-m. Here, as an example, the PFS 2-1 notifies the PFC 1-1 of the priority 24 “1”, the PFC 1-2 of the priority 24 “2”,..., The PFC 1-m of the priority 24 “m”. To do. However, m is the largest value from 1 to m, and the priority 24 having a small value is set as the high priority.

  Each of the PFCs 1-1 to 1-m that has received the reception notification of the first packet performs calculation of the communication path for the flow (packet), generation of the flow entry, and setting of the flow entry for the PFS2 (update of the flow table 23). It performs (steps S105-S114). Specifically, the PFC 1-1 that has received the first packet and the priority 24 from the PFS 2-1 has information defined as OpenFlow such as the MAC address and IP address of the transmission source host and the destination host included in the packet. A route is determined based on the above, and a flow entry to be set in the PFS 2 on the route is generated (step S105).

  The PFC 1-1 gives the priority 24 “1” notified from the PFS 2-1 to the generated flow entry, and updates the flow table 23 in order from the PFS 2-n on the destination host side (steps S106 to S109). . Specifically, the PFC 1-1 first transmits a flow entry with a priority 24 “1” to the PFS 2-n closest to the destination host 3-i (step S106). The PFS 2-n sets the transmitted flow entry in response to the flow table update instruction from the PFC 1-1 (step S107). Similarly, the PFC 1-1 goes back in the direction toward the transmission source on the calculated route, updates the flow table sequentially with respect to the previous PFS 2, and finally the PFS 2- closest to the transmission source host 3-1. One flow table is updated (steps S108 and S109).

  The flow entry update process is similarly performed for the other PFCs 1-2 to 1-m. For example, the PFC 1-m gives the priority 24 “m” notified from the PFS 2-1 to the generated flow entry, and updates the flow table 23 in order from the PFS 2-n on the destination host side (steps S110 to S110). S112). The PFC 1-m goes back on the calculated route toward the transmission source side, updates the flow table sequentially with respect to the previous PFS 2, and finally the PFS 2-1 closest to the transmission source host 3-1. The flow table is updated (steps S113 and S114).

  The process of updating the flow table for the PFS 2 by the PFC 1-1 to 1-m is performed independently without being synchronized. The status of the devices on the network may have changed due to the difference in the timing of receiving the first packet reception notification and the timing at the start of route calculation, and they do not always match. For this reason, the communication path calculated by each of the PFCs 1-1 to 1-m and the PFS 2 that is the setting target of the flow entry may be different. The process of updating the flow table 23 is also performed at a timing unrelated to the priority 147 set for each PFC2.

  However, since the PFS2 preferentially overwrites the flow table with a higher priority 24, even if a plurality of PFC1 updates the flow table without synchronization, one action information 145 corresponding to the rule 144 is obtained. can do.

  FIG. 11 is a flowchart showing the operation of the flow entry setting process (flow table update process) in the programmable flow switch according to the present invention. The details of the operation of the flow entry setting process (steps S107, S109, S112, S114) executed in PFS2 will be described with reference to FIG.

  The PFS 2 determines whether or not a flow entry that matches the rule 144 of the flow entry transmitted from the PFC 1 is set in the flow table 23 (step S201). If there is no corresponding entry in the flow table 23, the flow entry received from the PFC 1 is set in the flow table 23 (No in step S201, step S202). At this time, as shown in FIG. 12A, the flow entry (rule 144 + action information 145) transmitted from the PFC 1 and the priority 147 are associated with each other and set in the flow table 23.

  On the other hand, when there is an entry of the same rule 144 as the flow entry transmitted from the PFS 1 in the flow table 23, the comparison determination of the priority 147 is performed (steps S201 Yes and S203). Here, when the priority 147 of the flow entry transmitted from the PFC 1 is lower than the priority 147 set in the flow table 23, the PFS 2 ignores (rejects) the instruction to update the flow table from the PFC 1 (flow) (The table is not updated) (No in steps S203 and S204). On the other hand, when the priority 147 of the flow entry transmitted from the PFC 1 is higher than the priority 147 set in the flow table 23, the flow table is overwritten with the flow entry from the PFC 1, and the priority 147 is changed ( Step S203 Yes, S205).

  For example, when the PFC 1 in which the flow table 23 as shown in FIG. 12A is set is instructed by the PFC 1 to set the flow entry (priority “2”) for the packet A, the flow table 23 for the packet A Since the priority 147 of the flow entry already set to “1” is “1” (highest priority), the flow entry setting instruction (flow table update instruction) is ignored (rejected). On the other hand, when setting of the flow entry (priority “1”) for the packet C is instructed from the PFC 1, the priority 147 of the flow entry already set in the flow table 23 for the packet C is “3” and is notified. Since the priority is lower than “1”, the flow entry and the priority are changed as shown in FIG. In the example shown in FIG. 12A and FIG. 12B, when “Forward to PFS 2-5” is set as the action for packet B (rule), it is changed to “Forward to PFS 2-9”. The priority is changed from “3” to “1”.

  As described above, since the PFS 2 according to the present invention sets the flow entry by the PFC 1 having the higher priority 147 in the flow table 23 with priority, even if a plurality of PFCs 1 set the flow entries independently without being synchronized. , One transfer destination can be determined for the flow.

  Referring to FIG. 10, a PFS 2-1 (hereinafter referred to as a request source PFS 2-1) that requested setting of a flow entry in response to reception of a first packet has a flow entry corresponding to the first packet in its own flow table. If it is set to 23, it is determined that a communication path to the destination host 3-i has been constructed, and transfer of the first packet is started (step S115). Thereafter, the PFS 2 on the communication path transfers a packet conforming to the set rule 144 according to the set action information 145.

  The requesting PFS 2-1 preferably starts transferring the flow table at the same time that the flow entry for the first packet is set in its own flow table 23. In this case, the request source PFS 2-1 may transfer the first packet in accordance with the flow entry with the priority 147 lower than the priority “1”. For example, when the PFC 1-3 establishes a communication path before the PFC 1-1, the request source PFC 2-1 transfers the first packet according to the flow entry with the priority “3”. However, after the time has passed and the flow entry with the priority 147 “1” is set in the request source PFS 2-1 by the PFC 1-1, the packet is transferred via the communication path constructed by the PFC 1-1. It will be. As described above, even when a packet of the same rule is transferred, a flow entry having a higher priority is overwritten in the flow table 23 of the PFS2 as time elapses. And is transferred to the destination host 3-i. For this reason, the request source PFS 2-1 may wait for transmission of the first packet for a predetermined period after the flow table 23 of the request source PFS 2-1 is updated until the communication path is stabilized.

(Fault tolerance)
Next, fault tolerance in the computer system according to the present invention will be described with reference to FIGS. In the following, as an example, packet data is transferred from the host 3-1 to the host 3-2, the PFS 2-1 receives the first packet, and the three PFCs 1-1 to 1-3 construct a communication path. Will be described. The priority 147 “1” is assigned to the PFC 1-1, the priority 147 “2” is assigned to the PFC 1-2, the priority 147 “3” is assigned to the PFC 1-3, and the priority is “1”> “2”. “> 3”.

  FIG. 13 is a diagram illustrating an example of communication paths 40, 50, and 60 calculated by a plurality of PFCs 1-1 to 1-3 according to the present invention. Here, it is assumed that the communication path 50 is calculated by the PFC 1-1, the communication path 40 is calculated by the PFC 1-2, and the communication path 60 is calculated by the PFC 1-3. The communication path 40 is a path that passes through the PFSs 2-1, 2-2, 2-4, and 2-6, and the communication path 50 includes the PFSs 2-1, 2-5, 2-3, and 2-6. The communication path 60 is a path that passes through the PFS 2-1, 2-2, 2-3, 2-4, and 2-6.

  Under the preconditions described above, any of the PFCs 1-1 to 1-3 becomes inoperable during the construction of the communication path, and an instruction to update the flow table can be given to the PFS 2 on the communication path. The case where it disappears will be described.

  When there is no failure in the high priority PFC 1-1, even if there is a failure in the other PFC 1-2, 1-3, the hosts 3-1, 3-2 via the communication path 50 established by the PFC 1-1. In between, packet transfer is performed. For example, if a flow entry is set in the PFS 2 on the communication path 50 by the other PFC 1-2 and 1-3 before the PFC 1-1, the flow table 23 is a flow having a high priority “1”. Since the entry is overwritten, the communication path 50 is constructed without any problem. Further, even if the flow table is updated by other PFCs 1-2 and 1-3 after PFC 1-1, a high priority “1” flow entry is set in OFS2 which is shared with communication path 50. Therefore, the update of the flow table is ignored (rejected). For this reason, the communication path 50 constructed by the flow entry having the high priority “1” is not changed.

  In addition, when a failure occurs in the low priority PFC1-3, the flow entry is set by the high priority PFC1-1 or PFC1-2 regardless of the construction order of the communication path. A communication path is established. In the present invention, since the flow table update instruction is issued in order from the PFS 2-6 closest to the destination host 3-2, a communication path between the hosts can be constructed without any contradiction.

  Next, with reference to FIG. 14A and FIG. 14B, a description will be given of a communication path that is constructed when a PFC 1-1 with a high priority becomes inoperable during the construction of the communication path 50 preceding another PFC 1. .

  As shown in FIG. 14A, the PFC 1-1 updates the flow table 23 in order from the PFS 2-6 closest to the destination host 3-2 to the transmission source host 3-1, and the flow table 23 of the PFS 2-5 After the update, the flow table update process is stopped due to some failure. At this time, a flow entry having a priority “1” is set in the PFSs 2-5 and 2-6, and communication paths 51 and 52 to the PFSs 2-5 and 2-6 and the host 3-2 are established.

  After the start of update of the flow table by the PFC 1-1, the PFC 1-2 starts construction of the communication path 40 (update of the flow table). The PFC 1-2 updates the flow table 23 in order from the PFS 2-6 closest to the destination host 3-2 toward the transmission source host 3-1. At this time, since a flow entry with a high priority “1” is already set in the FS 2-6, the flow table with a priority “2” lower than this is not updated. Here, when there is no problem in the PFC 1-2 and the update of the flow table is completed up to the request source PFS2-1, the flow entry having the priority “2” is set in the PFSs 2-1 to 2-3. . That is, the communication path 41 is constructed between the host 3-1 and the PFS 2-6 by the PFC 1-2. On the other hand, since the communication path 51 is established by the PFC 1-1 between the PFS 2-6 and the host 3-2, the communication established by the PFC 1-2 is established between the host 3-1 and the host 3-2. It will be connected by the communication path 51 constructed | assembled by the path | route 41 and PFC1-1. Before and after these operations, the PFC 1-3 also updates the flow table. However, since a flow entry with a high priority is preferentially set, in the PFS 2 other than the PFS 2-2 on the communication path 60, A flow entry with priority “3” is not set.

  Even when the communication paths 40, 50, and 60 calculated for each PFC1 differ depending on the network environment, the PFS 2-6 closest to the transmission source and the request source PFS1-1 closest to the transmission source are all communication paths 40, 50. , 60. In the present invention, the flow table is updated in order from the destination host side to the source host side. For this reason, even if the update of the flow table by the high priority PFC1 (PFC1-1) is stopped in the middle, the communication path (communication path 41) constructed by the other low priority PFC1 (PFC1-2) The communication path (communication path 51) constructed by the high priority PFC1 (PFC1-1) is surely connected. From the above, according to the present invention, even if the PFC1 that sets a flow entry with a high priority 147 becomes inoperable during the construction of the communication path, the communication path can be supplemented by another PFC2. Packet transfer between the hosts 3 is guaranteed.

  Next, referring to FIG. 15A and FIG. 15B, when the PF 1-1 having a high priority starts construction of the communication path (update of the flow table) after other PFC1, and becomes inoperable during construction. A communication path construction operation will be described.

  The PFC 1-2 and PFC 1-3 update the flow table 23 in order from the PFS 2-6 closest to the destination host 3-2 to the source host 3-1. In this example, the PFC 1-1 starts building the communication path when the communication paths 42 and 61 are built as shown in FIG. 15A. As shown in FIG. 15A, when the PFC 1-1 starts to construct a communication path, the communication path 42 from the PFS 2-6 to the host 3-2 is constructed in the PFS 2-6 by the flow entry with the priority “2”. In the PFS 2-4, the communication path 61 is constructed between the PFS 2-5 and the PFS 2-4 by the flow entry having the priority “3”.

  Referring to FIG. 15B, the PFC 1-1 updates the flow table 23 in order from the PFS 2-6 closest to the destination host 3-2 to the transmission source host 3-1, and the flow table 23 of the PFS 2-5. After the update, the flow table update process is stopped due to some failure. At this time, a flow entry having a priority “1” is set in the PFSs 2-5 and 2-6, and communication paths 51 and 52 to the PFSs 2-5 and 2-6 and the host 3-2 are established.

  The PFC 1-2 continues to update the PFS2 flow table 23 on the communication path 40 from the PFS2-3 toward the transmission source host 3-1. Here, when there is no problem in the PFC 1-2 and the update of the flow table is completed up to the request source PFS2-1, the flow entry having the priority “2” is set in the PFSs 2-1 and 2-3. That is, the communication path 41 is constructed between the host 3-1 and the PFS 2-6 by the PFC 1-2. On the other hand, since the communication path 51 is established by the PFC 1-1 between the PFS 2-6 and the host 3-2, the communication established by the PFC 1-2 is established between the host 3-1 and the host 3-2. It will be connected by the communication path 51 constructed | assembled by the path | route 41 and PFC1-1. Before and after these operations, the PFC 1-3 also updates the flow table. However, since a flow entry with a high priority is preferentially set, in the PFS 2 other than the PFS 2-2 on the communication path 60, A flow entry with priority “3” is not set.

  Similarly to the above, even if the update of the flow table by the high-priority PFC1 (PFC1-1) that started construction of the communication path later stops, it is constructed by another low-priority PFC1 (PFC1-2) The communication path (communication path 41) is always connected to the communication path (communication path 51) constructed by the high priority PFC1 (PFC1-1). From the above, according to the present invention, even if the PFC1 that sets a flow entry with a high priority 147 becomes inoperable during the construction of the communication path, the communication path can be supplemented by another PFC2. Packet transfer between the hosts 3 is guaranteed.

  If the PFC 1-1 designated with the highest priority 147 is normal, the communication path 50 is always set after a predetermined time has elapsed by updating the flow table 23 (overwriting with a high priority flow entry). Built. In this case, since the flow can be controlled and managed by one PFC 1-1, the unified management of the flow proposed by the open flow technology can be realized.

  However, when an abnormality occurs in the PFC 1-1, a communication path that connects the hosts may be established by a plurality of PFCs 1 as described above. In this case, the flow is controlled by a plurality of PFCs 1, which is contrary to the “unified flow management by one PFC 1” which is the policy of the open flow technology.

  For this reason, when the PFC 1-1 notified of the highest priority 147 “1” is recovered from the failure, the communication path 50 is always reconfigured so that it does not communicate with the other PFC 1. Centralized management of flows by the PFC 1-1 can be realized.

  FIG. 16 is a flowchart showing the operation after the failure recovery of the PFC 1-1 notified of the priority 147 “1”. When the PFC 1-1 recovered from the failure returns, for example, by restarting, it confirms whether or not an instruction to update the flow table 23 has been issued to the PFS2 (request source PFS2-1) of the first packet transfer source (step S301). ). Specifically, after restarting, the PFC 1-1 searches the flow entry having the priority 147 set to “1” and its setting information 146 in the flow table 14 to determine the setting status of the corresponding entry in the PFS 1. Check.

  Here, when the setting of the flow entry for the request source PFS 2-1 is confirmed, the PFC 1-1 becomes inoperable after completion of the path construction, and is determined to have been restored by restart (steps S301 Yes and S302). On the other hand, if it is confirmed that no flow entry has been set for the request source PFS 2-1, the PFC 1-1 determines that it has become inoperable during path construction, and flows to the PFS 2 on the communication path 50. A table update instruction is issued (steps S301 No, S303). At this time, a new communication path may be calculated using the rule 144 set based on the first packet, and a flow entry (rule 144 + action information 145) set in the PFS2 on the communication path may be generated. . Normally, since the network environment after failure recovery has changed since the failure occurred, it is effective to set a new communication path and flow entry.

  By updating the flow table 23 of the PFS 2 of the entire communication path with the PFC 1-1 notified of the highest priority recovered from the failure in this way, the flow can be centrally managed by the PFC 1-1.

  Since the PFC 1 calculates a communication path that allows the packet to reach the destination host 3-2 from the transmission source host 3-1 fastest, a plurality of PFCs often determine the same communication path. For this reason, the method of constructing a communication path with a plurality of PFCs 1 is an effective measure from the viewpoint of fault tolerance. For this reason, in this invention, fault tolerance is improved by preparing several PFC1.

  Further, in the present invention, even if flow table update instructions from a plurality of PFCs 1 compete, the PFS 2 side determines whether or not to update (overwrite) the flow table on the basis of the priority. A communication path that guarantees Furthermore, since each of the plurality of PFCs 1 independently issues a flow table update instruction, there is no need to synchronize between controllers. For this reason, no waiting time for synchronization between the controls is required, so that it is possible to shorten the time to route construction. In other words, according to the present invention, it is possible to improve fault tolerance without increasing the time for establishing a communication path in a computer system using the open flow technology.

  The embodiment of the present invention has been described in detail above, but the specific configuration is not limited to the above-described embodiment, and changes within a scope not departing from the gist of the present invention are included in the present invention. . For example, the update control of the flow table according to the priority 147 may be performed not only for the construction of the communication path (setting of the transfer destination) but also for other control such as stop or end (discard) of the flow transfer. . Thereby, it is possible to execute instructions from a plurality of controllers without contradiction, and the fault tolerance is further improved.

  A part or all of the above embodiment can be described as in the following supplementary notes, but is not limited thereto.

(Appendix 1)
A plurality of controllers each for calculating a communication path and instructing a switch on the communication path to set a flow entry to which priority is added;
A switch that determines whether or not to permit setting of the flow entry according to the priority, and performs a relay operation according to the set flow entry for a received packet that matches the flow entry set for the computer. system.

(Appendix 2)
In the computer system according to attachment 1,
The switch overwrites the set flow entry by the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is higher than the priority of the flow entry set by itself. Computer system to configure.

(Appendix 3)
In the computer system according to appendix 1 or 2,
The switch rejects the setting of the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is lower than the priority of the flow entry set in the switch.

(Appendix 4)
In the computer system according to appendix 2 or 3,
Each of the plurality of controllers instructs the setting of a flow entry in order from the switch on the destination device side of the packet data to the switch of the transmission source device on the communication path.

(Appendix 5)
In the computer system according to any one of appendices 1 to 4,
The priority added to a flow entry by each of the plurality of controllers is a value that is different for each of the plurality of controllers.

(Appendix 6)
In the computer system according to any one of appendices 1 to 5,
When the switch receives packet data that does not match the flow entry set in itself, the switch notifies the plurality of controllers of the packet data and the priority,
Each of the plurality of controllers adds the notified priority to a flow entry created based on header information of the packet data, and instructs a switch on the communication path to set.

(Appendix 7)
A controller used in the computer system according to any one of appendices 1 to 6.

(Appendix 8)
A switch used in the computer system according to any one of appendices 1 to 6.

(Appendix 9)
Each of a plurality of controllers calculating a communication path and instructing a switch on the communication path to set a flow entry with a priority;
The switch determining whether or not to permit setting of the flow entry according to the priority;
A step in which the switch performs a relay operation in accordance with the set flow entry for a received packet that matches the flow entry set for the switch.

(Appendix 10)
In the communication method according to attachment 9,
When the priority of the flow entry transmitted from each of the plurality of controllers is higher than the priority of the flow entry set in the switch, the flow entry set in the switch is overwritten by the transmitted flow entry. A communication method further comprising the step of:

(Appendix 11)
In the communication method according to attachment 9 or 10,
The communication method further comprising the step of rejecting the setting of the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is lower than the priority of the flow entry set in the switch .

(Appendix 12)
In the communication method according to attachment 10 or 11,
The step of instructing the setting includes a step of instructing the setting of the flow entry in order from the switch on the destination device side of the packet data to the switch of the transmission source device on the communication path.

(Appendix 13)
In the communication method according to any one of appendices 9 to 12,
The communication method, wherein the priority added to the flow entry by each of the plurality of controllers is a different value for each of the plurality of controllers.

(Appendix 14)
In the communication method according to any one of appendices 9 to 13,
When the switch receives packet data that does not conform to the flow entry set in itself, the switch further includes a step of notifying the plurality of controllers of the packet data and priority,
The step of instructing the setting is performed by adding the notified priority to the flow entry created by each of the plurality of controllers based on the header information of the packet data, and setting the switch on the communication path. A computer system comprising the step of directing.

1, 1-1 to 1-m: Programmable flow controller (PFC)
2, 2-1 to 2-n: Programmable flow switch (PFS)
3, 3-1 to 3-i: Host computer (host)
11: Switch control unit 12: Flow management unit 13: Flow generation unit 14, 23: Flow table 15: Topology information 16: Communication path information 20: Switch group 21: Transfer processing unit 22: Flow setting unit 24, 147: Priority 141: Flow identifier 142: Target device 143: Route information 144: Rule 145: Action information 146: Setting information 151: Device identifier 152: Number of ports 153: Port connection destination information 161: End point information 162: Passing switch information 163: Accompanying information

Claims (10)

A plurality of controllers each for calculating a communication path and instructing a switch on the communication path to set a flow entry to which priority is added;
A switch that determines whether or not to permit setting of the flow entry according to the priority, and performs a relay operation according to the set flow entry for a received packet that matches the flow entry set for the computer. system. The computer system of claim 1,
The switch overwrites the set flow entry by the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is higher than the priority of the flow entry set by itself. Computer system to configure. The computer system according to claim 1 or 2,
The switch rejects the setting of the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is lower than the priority of the flow entry set in the switch. The computer system according to claim 2 or 3,
Each of the plurality of controllers instructs the setting of a flow entry in order from the switch on the destination device side of the packet data to the switch of the transmission source device on the communication path. The computer system according to any one of claims 1 to 4,
When the switch receives packet data that does not match the flow entry set in itself, the switch notifies the plurality of controllers of the packet data and the priority,
Each of the plurality of controllers adds the notified priority to a flow entry created based on header information of the packet data, and is a computer system for a switch on the communication path.   The controller used with the computer system of any one of Claim 1 to 5.   The switch used with the computer system of any one of Claim 1 to 5. Each of a plurality of controllers calculating a communication path and instructing a switch on the communication path to set a flow entry with a priority;
The switch determining whether or not to permit setting of the flow entry according to the priority;
A step in which the switch performs a relay operation in accordance with the set flow entry for a received packet that matches the flow entry set for the switch. The communication method according to claim 8, wherein
When the priority of the flow entry transmitted from each of the plurality of controllers is higher than the priority of the flow entry set in the switch, the flow entry set in the switch is overwritten by the transmitted flow entry. A communication method further comprising the step of: The communication method according to claim 8 or 9,
The communication method further comprising the step of rejecting the setting of the transmitted flow entry when the priority of the flow entry transmitted from each of the plurality of controllers is lower than the priority of the flow entry set in the switch .

Images