JP2006173704A - Path control system, path control method, and program for the method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a path control system in a packet switched network for transmitting a packet having the same destination address to a destination via a different transmission path, and to provide a path control method and a program for the path control method. <P>SOLUTION: The path control system 10 comprises a means 40 for securing an unused IP address in a network by allowing an edge node 26 to transmit a request for transmitting a packet in a specified transmission path, a means 28 for selecting a packet to be transmitted in the specified transmission path, a means 36 for analyzing the selected packet and substituting it with an unused IP address for securing a destination IP address, and a means 38 for transmitting the substituted packet to other specified nodes. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a routing technology for packets transmitted via a network, and more specifically, a route in a packet switching network for transmitting packets having the same destination address to a destination via different transmission routes. The present invention relates to a control system, a path control method, and a program therefor.

  Conventionally, a hop-by-hop packet switching network using a packet format such as IPv4 or IPv6 adds a header indicating a source address and a destination address to data transmitted and received between computers, and encapsulates the packet. And send it to the network. A network device such as each relay computer (router) determines a node (next hop) to be transmitted next from its destination address. In the hop-by-hop type, each node determines the next hop by looking up the route table of each node. In the route table held by the node, a pair of a destination address and a next hop address corresponding to the destination address is registered, and the node can look up the route table and know the next hop from the destination address. . Further, such a route table is updated by a static management method or a dynamic management method in response to a change in the network configuration. However, in the above-described conventional packet switching network, all packets having the same destination address pass through the same route in principle.

  On the other hand, in recent years, information transmitted and received using a communication infrastructure such as the Internet tends to increase more and more, and information transmitted and received is often given high information added value. For this reason, in order to guarantee quality such as traffic distribution, bandwidth guarantee, delay guarantee, etc. in the packet switching network, the source address, destination port number, and source port number are not used for routing using only the destination address. It may be preferable to specify a specific set of packets (packet flow or flow) indicated by, etc., and communicate the specified packets separately from other packets by specifying the desired transmission path. .

  However, as described above, in a conventional hop-by-hop packet switching network in which a route table is provided for each router, all packets having the same destination address go through the same route and cannot pass through the specified route. It is impossible, and it is impossible to specify a route by specifying a packet.

  In the past, for example, the technology described below has been proposed for routing for a specific packet flow, and generally a method that combines MPLS technology and explicit routing, Generally used.

Patent Publication No. 2002-516042, entitled “Improved method and apparatus for dynamically shifting between packet routing and switching in a transmission network” MPLS: IETF (Internet Engineering Task Force) RFC3031.txt (http://www.ietf.org/rfc/rfc3031.txt) RSVP-TE: IETF (Internet Engineering Task Force) RFC3209.txt (http://www.ietf.org/rfc/rfc3209.txt) Type of Service (ToS): IETF (Internet Engineering Task Force) RFC1349.txt (http://www.ietf.org/rfc/rfc1349.txt)

  In Patent Document 1, a list of routers that do not pass through forwarding in a route table among nodes belonging to a route that the packet is to be explicitly passed through is designated in the packet. Therefore, as the number of routers that need to be specified for not passing through increases, the packet length greatly changes depending on the route, and the overhead for the communication bandwidth becomes difficult to manage. For this reason, the method of Patent Document 1 causes a disadvantage that the scalability of transmission processing cannot be guaranteed. In addition, a method of providing a different transmission route by preparing a list of routers that do not pass through and selecting them can be considered, but that alone cannot provide information on packet quality control. Another method is to increase the overhead for the network device, such as filtering all packets to determine whether or not the device itself is designated as a node that does not pass, and requiring a route selection process. Cause problems.

  Non-Patent Documents 1 and 2 assign a label defined in a data link layer such as ATM or frame relay to a packet flow, or a label defined on a data link layer such as Ethernet (registered trademark). This is a method for identifying a packet flow based on this. These methods can be applied to packet flows of various granularities, and can have information on quality control when transferred to a label. However, in order to use these methods, in addition to the conventional route table, all routers in the network are required to have a new label forwarding table, and a new management target is added to the administrator. Cause the burden of In addition, as information processing means at the data link layer level also used systemically, it is necessary to select one that can use a label and replace it as necessary.

  In addition, as a technique for limiting the granularity of transaction nodes in a packet flow, a technique described in Non-Patent Document 3 can be cited. The method disclosed in Non-Patent Document 3 is a method of changing a packet transfer method based on ToS information stored in an IP header. According to the method described in Non-Patent Document 3, a route can be changed depending on ToS information. However, such a method cannot realize explicit route control with high flexibility for various types of packet flows.

  As described above, according to the conventional technique, in order to transmit a packet flow along an explicit route, the data link is changed to a data link that can be used, and a new forwarding table is added to perform management. There is a disadvantage that it is necessary to provide an address list separately from the routing table, which may be a large overhead for bandwidth, and further, the types of packet flows that can be routed are limited.

  That is, the present invention (1) does not require a new forwarding table, (2) can identify and handle packet flows of various granularities, (3) and without making a large-scale change of the network infrastructure It is an object of the present invention to provide a path control system, a path control method, and a program therefor that are easy to migrate.

  In order to solve the above problems, the present inventors have intensively studied, and as a result, by using the route control of the present invention, a highly flexible hop-by-hop packet exchange is possible. This has been found and the present invention has been achieved. That is, in the present invention, an unused IP address Z is used for routing in the address space used for routing control. More specifically, in the present invention, the unused IP address Z is an IP address that is not used by the egress node of the network, and is acquired when the routing control of the present invention is executed. It means an IP address that is used exclusively during the route control.

  In the present invention, the ingress node identifies an egress node from a corresponding route table for a packet that is to be passed through a specific path, and sends a request packet that notifies the egress node to perform the path control of the present invention. Issue. Upon receiving this request packet, the egress node obtains and occupies an unused IP address Z that can be used at the egress node, encapsulates the unused IP address Z obtained in the response packet, and enters it together with the response packet. Reply to the node. Upon receiving the response packet, the relay node and the relay node that transmit to the gateway node create an entry corresponding to the new hop address in the routing table, and use the routing of the present invention other than the routing table. Realize without.

  The ingress node receives the response packet from the egress node and acquires an unused IP address. Thereafter, the ingress node writes an unused address to the packet for routing according to the present invention and sends it to the network. When the egress node receives a packet including an unused address in the header, it replaces the unused IP address with the original destination address using the correspondence table held by the egress node, or the original destination address held by the packet. Replace the unused IP address with, and transmit the packet to the correct destination node.

  In the present invention, the storage of the original destination address is designated from the ingress node to the egress node by converting the address at the egress node, storing in the packet to be stored in the IP packet, forming a new IP header at the ingress node It is possible to perform the packet transmission by tunnel method along the route and delete the new IP header.

That is, according to the present invention, there is provided a flow identification and flow level routing control system for performing routing control of packets transmitted in a whip-by-whipping type packet switching network in the network, the system including the network And a plurality of nodes connected to the network, and at least a first node of the nodes is
Means for transmitting a request for transmitting a packet on a designated transmission path, and securing an unused IP address in the network;
Means for selecting a packet to be transmitted on the designated transmission path;
Means for analyzing the selected packet and replacing a destination IP address with the reserved IP address reserved;
A routing control system including means for sending the replaced packet to another designated node can be provided.

  The plurality of nodes include a plurality of edge nodes connected to a user node and a core node that relays the packet, and the first node constitutes the edge node, and the edge node And the core node includes a data structure storing an entry associating the reserved unused IP address with the IP address of the next hop node for routing the unused IP address.

  The sending means describes the packet transmitted through the designated route, describing the reserved unused IP address in the destination IP address field of the IP header, and the destination IP address behind the payload or the IP header. Means for encapsulating the IP option field or the routing header, or adding a new header including the unused IP address.

  The sending means replaces the packet transmitted through the designated route by replacing a first predetermined byte of a destination IP address of the packet with a second predetermined byte of the unused IP address. Means for generating as an address and encapsulating the packet.

  At least one edge node of the plurality of nodes receives the packet including the unused IP address and replaces a second predetermined byte of the unused IP address with the first predetermined byte. Data structure for.

According to the second configuration of the present invention, routing of packets transmitted in a whip-by-whip type packet switching network is performed in the network including a network and a plurality of nodes connected to the network. A flow identification and flow level route control method to be performed, wherein the route control method performs a first information processing apparatus,
A function unit for transmitting a request for transmitting a packet through a designated transmission path and securing an unused IP address in the network;
A functional means for selecting a packet to be transmitted on the designated transmission path;
Functional means for analyzing the selected packet and replacing a reserved IP address with the reserved IP address,
A path control method can be provided in which the replaced packet is configured with functional means for transmitting the packet to another designated node.

The plurality of nodes include a plurality of edge nodes connected to a user node and a core node that relays the packet, and the first node constitutes the edge node;
In the method, the edge node and the core node have an entry corresponding to the unused IP address and the IP address of the next hop node for routing the reserved unused IP address. Functional means for generating a stored data structure is included.

  The sending function means describes the packet transmitted through the designated route by describing the reserved unused IP address in the destination IP address field of the IP header, and the destination IP address behind the payload, or IP It includes functional means encapsulated in an IP option field of a header or a routing header, or means for adding a new header including the unused IP address.

  The sending function means replaces the first predetermined byte of the destination IP address of the packet with the second predetermined byte of the unused IP address in the packet transmitted through the designated route. Function means for generating as a destination address and encapsulating the packet in the packet is included.

  At least one edge node of the plurality of nodes receives the packet including the unused IP address and sends a second predetermined byte of the unused IP address to the first node of the first node. Functional means for generating a data structure for substituting the predetermined bytes.

According to the third configuration of the present invention, path control of packets transmitted in a whip-by-whip type packet switching network within the network including a network and a plurality of nodes connected to the network is performed. A program for causing an information processing apparatus to execute a flow identification and flow level path control method to be performed,
A function unit for transmitting a request for transmitting a packet through a designated transmission path and securing an unused IP address in the network;
A functional means for selecting a packet to be transmitted on the designated transmission path;
Means for analyzing the selected packet and replacing a destination IP address with the reserved IP address reserved;
Means for generating a data structure storing an entry that associates the unused IP address with the IP address of the next hop node;
And a function unit configured to look up the data structure of the replaced packet to a next hop node and send the packet to a next hop address.

According to a fourth configuration of the present invention, there is provided a flow identification and flow level routing control system for identifying and controlling a flow of a packet transmitted in a whip-by-whipping type packet switching network in a network. The system includes the network and a plurality of nodes connected to the network, and includes at least a first node and a second node of the nodes, and the second node includes:
Means for receiving a request for transmitting a packet from the first node through a designated transmission path and securing an unused IP address in the network;
Means for notifying an unused IP address secured to the first node in the designated transmission path;
Analyzing means for receiving the second unused IP address converted from the unused IP address and the destination IP address by a predetermined protocol or a packet including the unused IP address and analyzing the packet;
Including
When the analyzing unit detects a second unused IP address in the packet, the analyzing unit reproduces the original destination IP address from the second unused IP address, and the analyzing unit obtains the unused IP address. When detected, a routing control system can be provided that reads the destination IP address described in the packet and reproduces the original destination IP address of the packet.

  According to the present invention, a routing control system, a routing control method, and a routing control system that enable flexible routing using only an existing routing table without using a forwarding table at the data link layer level, and therefore Can be provided.

  The present invention will be described below with reference to embodiments shown in the drawings, but the present invention is not limited to the embodiments described below.

  FIG. 1 shows an embodiment of a routing system 10 used by the hop-by-hop packet switching network of the present invention. The routing system shown in FIG. 1 is essentially capable of performing conventional routing and routing according to the present invention, and FIG. 1 shows an implementation in which packets are transmitted by conventional routing. The form is shown. A packet switching system 10 shown in FIG. 1 is composed of a plurality of nodes, and is an ISP (Internet Service Provider) node, for example, where a node P, a node Q, a node R, and a node S are connected to a network. A and node B are user nodes 12 and 14, respectively. For example, when the user node 12 transmits a packet to the user node 14 without any particular designation, the node is hopped in the order of node P → node Q → node S → user node 14 to The packet is transmitted from 12 to the user node 14. Node P, node Q, node R, and node S hold route tables 16 to 22, respectively. As shown in FIG. 1, the route tables 16 to 22 store the destination IP address Dst and the hop node NHop corresponding to Dst in association with each other.

  For example, when the node P receives a packet whose destination IP address is addressed to the user node 14, the node P looks up the routing table 16, selects the node Q as the next hop node, and hops the packet toward the node Q. In addition, when the node Q receives the packet from the node P, the node Q determines the destination IP address, selects the node S as the next hop node, and transmits the received packet to the node S. The packet from the user node 12 is transmitted to

  The user node 12 and the user node 14 are configured as a personal computer, a mail server, or a router, and the user nodes 12 and 14 are connected to a LAN or WAN to which a plurality of computers are connected. It can be shown as a representative of the network equipment to be configured. Nodes P and S connected to the user node 12 and the user node 14 are referred to as network equipment that receives a packet from the user node and transmits it to a destination, for example, an edge node. And referred to as the exit node. Node R and node Q are network devices such as routers, switches or computers having a function of relaying packets, and are referred to as core nodes hereinafter. However, even the core node can include an edge router function. The nodes P, Q, R, and S are connected to each other by a communication network to form a network 24 such as a so-called Internet. In the present invention, the network 24 can use a network using a communication protocol such as UDP (User Datagram Protocol) in addition to a protocol at a network layer / transport layer level such as a TCP / IP protocol.

  FIG. 2 is a diagram showing an embodiment when a packet is transmitted through a designated route according to the present invention. As shown in FIG. 2, the route control system 10 of the present invention can arbitrarily designate a transmission route other than that shown in FIG. That is, in the embodiment of the present invention, a packet sent from the user node 12 is received by the node P, and then transmitted to the user node 14 via the node R, unlike FIG. . In this case, in the present invention, the node P specifies an egress node for the user node 14 specified in advance from the route table, and transmits a request packet for performing the route control of the present invention to the node S. To do. Thereafter, a response packet from the node S is transmitted, and the node P and the node R respectively acquire an unused IP address Z in the node S. For example, in the node P, a new entry {Z, R} in the route table Node R adds a new entry {Z, S} to the route table.

  The user node 12 sends a packet to be transmitted to the user node 14 in a normal format to the node P. The node P creates a packet including the unused IP address Z of the node S acquired in advance, looks up the route table, and dispatches the packet to the network 24. When node R corresponding to the specified route receives the packet, it looks up the route table to obtain the next hop address of the entry containing the unused IP address Z, and hops the packet to node S. ing. On the other hand, when the node S receives a packet including the unused IP address Z as a destination address, the node S performs a process of converting the packet to an original destination IP address corresponding to the packet, and transmits the packet to the user node 14. Note that the symbol indicated by “*” indicates processing for returning to the original address.

  FIG. 3 is a schematic diagram showing the functional configuration of the edge node of the present invention. The edge node 26 shown in FIG. 3 can have the same configuration for both the entry node and the exit node, and is controlled by a central processing unit (CPU) (not shown). The configuration of the edge node 26 shown in FIG. 3 will be further described. The edge node 26 classifies a packet for performing routing according to the present invention and a packet to be transmitted through a normal route. And a route table 30 and an address pool management means 32. When receiving a packet from the outside, the edge node 26 first determines whether or not the packet needs to be route-controlled according to the present invention by the packet classification unit 28. This determination can be made by, for example, registering the source IP address of the packet flow in an appropriate storage device of the edge node 26 such as the packet classification means 28 in advance. The other functional configuration of the packet classification means 28 is constructed using conventional functional means such as IntServ / DiffServ.

  In the present invention, the route table 30 is dynamically updated by RIP or the like, for example, and an entry registered in association with the IP address of the next hop node corresponding to the destination IP address, and generated according to the present invention. It includes an entry in which the IP address Z of the designated next hop node corresponding to the unused IP address is registered. Further, the address pool management means 32 selects an unused IP address Z from the address pool 34 among the IP addresses that can be used by the edge node 26 and registers it in the routing table.

  Note that the address pool 34 and the address pool management means 32 keep all unused IP addresses Z that can be used by the edge node 26, for example, all the IP addresses assigned to the edge node 26 in the address pool 34. It is also possible to configure so that a flag is set for an already used IP address, an IP address without a flag is searched, and the use of the IP address Z is instructed. In another embodiment of the present invention, the address pool 34 sets a flag directly indicating that it is an unused IP address, and an unused IP address is directly used by using the flag-set IP address. Can be obtained as further. In another embodiment of the present invention, only unused IP addresses can be stored in the address pool 34.

  The edge node 26 further includes an input flow processing means 36, an output flow processing means 38, and a signaling processing means 40. The input flow processing means 36 processes the packet processed according to the route control of the present invention, and looks up the route table 30 so as to transmit the packet to the destination node via the designated route. Perform mapping. When the edge node 26 is an egress node, the output flow processing means 38 converts the unused IP address Z into the original destination IP address and returns to the user node 14 having the original destination IP address. Send the packet. When the edge node 26 is an ingress node, the output flow processing means 38 sends the mapped packet to the next hop node according to the routing at the data link layer level and the physical layer level such as the MAC address. And dispatch the packet.

  Further, the signaling processing means 40 identifies the corresponding egress node from the route table at the ingress node, generates a request packet, and transmits the generated request packet to the egress node. Further, the egress node generates a response packet including an unused IP address Z corresponding to the request packet, and performs a process of returning the response packet to the ingress node.

  FIG. 4 shows a schematic flowchart of the route control method of the present invention. As shown in FIG. 4, in the routing control method of the present invention, the ingress node receives a packet from the user node in step S400. In step S402, it is determined whether or not the incoming packet is a packet flow that has been determined as an explicit routing target. If the packet is an explicit routing packet (yes), the process proceeds to step S404. To create a packet containing an unused IP address Z. If it is not a packet for performing explicit routing (no), the packet is transmitted according to the routing table without processing the original packet.

  The packet including the unused IP address Z created in step S404 is transmitted to the next hop node determined by referring to the route table in step S408. Thereafter, in step S410, the packet is restored to the original packet at the egress node, and then transmitted to the original destination user node.

  FIG. 5 is a flowchart showing processing executed by the ingress node and the core node of the present invention. The ingress node of the present invention issues a request packet to the egress node using the RSVP_PATH command defined in the RSVP-TE protocol in order to enable the routing of the present invention. As a signaling function of the present invention, although lacking flexibility, it is possible to use a static setting method, SNMP, or the like.

  Regarding the RSVP-TE protocol used in the present invention, for example, David O. Awduche et. Al., “RSVP-TE: extensions to RSVP for LSP tunnels.”, RFC 3209, Internet Engineering Task Force, December 2001, “Constraint -based LSP setup using LDP. ”, RFC 3212, Internet Engineering Task Force, January 2002, S. Blake, et. al.,“ An architecture for differentiated service. ”, RFC 2475, Internet Engineering Task Force, December 1998,“ Resource reservation protocol (RSVP) -version 1 functional specification. ”, RFC 2205, Internet Engineering Task Force, September 1997, P. Pan, et. Al.,“ Fast reroute extensions to rsvp-te for lsp tunnels. ”, Internet- Details of the protocol are published in draft, Internet Engineering Task Force, May 2004. The packet format used in the present invention will be described later in more detail.

  Hereinafter, in more detail, the process of the ingress node shown in FIG. 5 issues a request packet for performing path control of the present invention to the egress node using RSVP_PATH in step S500. Next, in step S502, the unused IP address Z transmitted from the egress node is transmitted through the route designated by the ingress node. If the response packet passes through each core node in step S502, the core node analyzes the received packet and determines that it is a response packet to the request packet in step S500. The unused IP address Z included in this area is associated with the IP address of the previous hop node and stored as an entry in the routing table. In step S504, the ingress node acquires an unused IP address Z. If the ingress node determines that the packet is a response packet to signaling, the ingress node adds the unused IP address Z to its route table as an entry associated with the designated next hop node IP address.

  FIG. 6 is a flowchart showing processing executed by the egress node of the present invention. The egress node of the present invention receives the RSVP_PATH transmitted from the ingress node in step S600 as a request packet. Next, in step S602, the egress node searches the address table to obtain an unused IP address Z. At the same time, the egress node reserves an unused IP address Z by setting a flag or the like at least during execution of the routing control of the present invention, and occupies this as an explicit routing IP address. In step S604, the egress node creates a RESV packet as a response packet, and returns the generated RESV packet to the ingress node via the route described in PATH. As described with reference to FIG. 5, the core node writes the unused IP packet Z and the IP address of the previous hop node in association with each other in the route table when the RSVP packet passes.

  FIG. 7 is a diagram showing a signaling (request / response) transaction between the nodes described in FIGS. 5 and 6 together with a data structure used for signaling. The RSVP_PATH sent from the node P will be described in detail. The RSVP_PATH includes an explicit routing object ERO that specifies a column (list) of passing nodes, a PHOP that is a history identifier that records a node of the immediately preceding hop, LRO (Label Request Object) command for requesting a reply to the transmission source with an unused IP address as a label and a destination IP address while notifying that the route control of the invention (Flow Mapped Explicit Host Routing: FMHR or FMEHR) Including. The RSVP_PATH packet passes through the node R and reaches the node S according to the route described in the PATH without looking up the route table. When the RSVP_PATH packet reaches the node S, the node S assigns an unused IP address, for example, abc.cde.fgh.1 / 32 (/ 32 indicates the number of bits of the subnet mask), and the RESV packet It is added as a label and sent back to node P via the same route as RSVP_PATH. Note that this unused IP address Z includes private addresses defined as 10.0.0.0 to 10.255.255.255, 127.0.0.0 to 127.255.255.255, 192.0.0.0 to 192.255.255.255, IPv6 site local addresses, etc. Can be used. However, other addresses may be used as long as other routes are not affected.

  After the transaction in FIG. 7 is completed, the ingress node acquires abc.def.ghi / 32 as an unused IP address Z. However, if the unused IP address Z received without any processing is sent to the node S as it is, any packet from the node P can be sent if there is no corresponding routing table on the node S side. It will be discarded. For this reason, in the present invention, in order to enable accurate routing using the unused IP address Z, in a specific embodiment, for the node S, the node S is an edge node, and the unused IP address Z It is identified that the process of performing address conversion from the address Z is executed. As a method for this, in the present invention, an egress node storage type address translation system, an intra-packet storage type address translation system, or a tunnel system address translation system can be used.

  FIG. 8 is a diagram showing an embodiment of an egress node storage type address conversion method, which is a so-called egress node storage type address conversion method in which an egress node generates an accurate destination address. In this method, the destination address D is completely rewritten to an unused address Z at the ingress node, and the destination IP address is not transmitted together with the packet. Referring to FIG. 8A, the process of acquiring an unused IP address will be described. The node P, which is an ingress node, sets RSVP_PATH to the network, RSVP_PATH w / LRO (IP / FMHR, 1.2.3.0/24 ) Including RSVP_PATH message. Here, the IP address 1.2.3.0/24 is the destination IP address of the user node 14 to which the packet is to be transmitted.

  This response packet RSVP_PATH reaches the node S via the node Q. When the egress node receives this request packet, it obtains an unused IP address Z equal to the prefix length secured in the range such as abc.def.ghi from the address pool, and uses the selected unused IP address prefix. In this particular embodiment, the IP address of abc.cde.fgh.0 / 24 is created by replacing the upper byte 0/24 of the destination IP address. This IP address is included in the Label as an unused IP address used for path control of the present invention, and returned to the node P.

  Node P obtains an unused IP address, abc.cde.fgh.0 / 24, as Label, but if this is the case, node P will send a reply to abc.cde.fgh.0 / 24. Although node S is reached, there is no routing table in node S, so it is discarded, and node S is not the target destination node B, so it can send a packet to node B. Can not.

  Therefore, the node P stores the prefix part of abc.cde.fgh in an appropriate storage device and uses the routing of the present invention addressed to the node P as shown in FIG. 8B. Set the destination IP address of the packet to abc.cde.fgh.5. . . And then send it to the network. The egress node looks up the routing table using the received packet prefix as a key, determines that it is the egress node, reads the destination IP address prefix from the storage device, and sets abc.cde.fgh as Replace with 1.2.3. In the case where processing from a plurality of ingress nodes is executed in the present invention, the output flow control means described in FIG. 3 is converted in accordance with the unused IP address prefix as shown in FIG. Look up a table containing the prefix of the destination IP address Z to be. Thereafter, the same processing as described above is executed, an appropriate prefix is read, the IP address of the packet transmitted from the node P is restored, and the packet is accurately transmitted to the node B. The IP address shown in FIG. 8 is described in IPv4, but the same conversion can be applied to IPv6.

  FIG. 10 shows another address translation method that can be used in the present invention. FIG. 10 (a) shows an intra-packet storage type address conversion method for storing a destination IP address in a packet. FIG. 10 (b) shows an ingress node where the ingress node IP address + the node S IP address. Shows a tunnel address translation method that adds a new IP header to the packet. In the embodiment of FIG. 10A, in the packet 50, the payload 54 and the original destination IP address 56 are added to the rear part of the IP header including the transmission source address 52 in addition to the unused IP address Z.

  In the embodiment shown in FIG. 10B, another IP header 60 including an unused IP address and an IP address of the ingress node is added to the packet 58, and the subsequent packet 62 is stored. By doing so, a pseudo tunnel method is provided. In the embodiment shown in FIG. 10, since the original destination IP address is transmitted from the ingress node, it is not necessary to convert the IP address as in the egress node preserving type packet switching method. Will increase the amount of bandwidth consumed. In the present invention, which of these embodiments is used to perform address translation can be appropriately selected in consideration of the capacity of the router and the bandwidth of traffic to be used.

  FIG. 11 is a diagram showing packet conversion processing for realizing hop-by-hop when a destination IP address is included in a packet in the present invention. Note that the packet shown in FIG. 11 is an embodiment showing an IPv4 packet format. FIG. 10 shows an embodiment in which the route is controlled in the direction indicated by the arrow from the node A as the user node to the node B as the destination. Referring to FIG. 11, when the present invention is performed using the intra-packet storage type address translation method, a hop-by-hop flow path is designated for the original packet 70 transmitted from the user node. A packet 76 in which the IP address is added after Payload is generated.

  The packet 70 includes an IP header 72, version information (V), header length (HL), priority (T), packet length L + n (n is 4 bytes in IPv4), The checksum (SUM), the source IP address (S), and the destination IP address (D) are included. Further, behind the IP header 72, Payload 74, which is data to be transmitted, is described. The packet 76 generated in the present invention includes an IP header 78 and a payload 80, and an unused IP address Z is described in the IP header 78 in the area of the destination IP address. In addition, the packet 76 has a destination IP address D added by 4 bytes. In this respect, it is simpler than the RIP and IGRP packet formats used in conventional routing methods, is compact in software, can be configured to be installed on an existing router, and is directly Route control can be performed.

  When the node P receives the packet 70 from the node A, the node P determines the contents of the transmission source address S and determines whether or not the value is within the range for which the routing of the present invention should be executed. As a result, when it is determined that the routing of the present invention is performed, the original destination IP address D is replaced with the unused IP address Z sent from the node S in advance, and further the original destination IP address stored in the memory Add D after the payload. At this stage, the data length L is increased by 4 bytes, and the checksum is recalculated. Thereafter, the routing table is looked up, and the packet is transmitted to the node Q assigned to the unused IP address. In the present invention, when the network is composed of a plurality of layers, the unused IP addresses Z1 and Z2 can be assigned to the plurality of layers. When a plurality of unused IP addresses are used, the packet length is more generally increased by (nx4) bytes when n unused IP addresses Z1 to Zn are used.

  In FIG. 11, the packet 76 is routed from the node P → the node Q → the node S according to the route table shown in FIG. When the packet 76 reaches the node S, the packet 76 is restored to the packet 70 which is the original transmission packet in the node S, and is transmitted to the node B which is the user node. When each packet is changed, the checksum is recalculated.

  FIG. 12 is a diagram showing an embodiment in the case where route control is performed via a plurality of hierarchies in the present invention. In FIG. 12, the network 24 shown in FIGS. 1 and 2 is separated into a plurality of layers, and unused IP addresses Z1 and Z2 are assigned to each layer. In the embodiment shown in FIG. 12, the first layer is composed of two sub-networks, and a common unused IP address Z1 is assigned. The second hierarchy is composed of one subnetwork, and Z2 is assigned as an unused IP address. In transmission of a packet from the user node 12 to the user node 14, first, an unused IP address Z1 assigned in the first layer is encapsulated and sent to S1, which is an egress node. To the ingress node P2 of the hierarchy, the destination address is converted into an unused IP address Z2, and the destination IP address is changed to the unused IP address Z1 to the user node 14 via the egress node S2. Replaced and sent. The present invention can perform hierarchical flow identification and path control in a large-scale network by using a packet data format and a path control method described later.

  FIG. 13 is a diagram showing an embodiment when a packet of the intra-packet storage type address translation method of the present invention that can be used in the multi-layered network shown in FIG. 12 is described in the IPv6 format. In this example, the prefix for the upper 4 bytes is replaced with an unused address space, but any number of bytes may be used. The packet 84 shown in FIG. 13 includes an IPv6 header 86 that can include version information (V), unused IP address prefixes Z2 and Z1, and an IPv6Payload 88. When this IPv6 packet is received by the node P1, an unused IP address prefix Z1 assigned by the egress node corresponding to the first layer is set as the destination IP address, and an IPv6 packet 90 is generated. When the ingress node in the second hierarchy is reached, the ingress node overwrites the destination IP address with the unused IP address Z2 allocated in the second hierarchy, and performs routing control according to the present invention in the second hierarchy. An IPv6 packet 92 for performing is created. Thereafter, when another ingress node P3 in the first hierarchy is reached, it is overwritten again with the unused IP address Z1, the IPv6 packet 90 is reproduced, and is transmitted again through the first hierarchy by the routing of the present invention. After that, when the IPv6 packet 90 reaches the egress node in the first hierarchy, the unused IP address Z1 is overwritten with the original destination IP address D, and the original packet should be delivered to the user node Node B. The packet has been restored.

  FIG. 14 is a diagram showing an embodiment of a packet in the IPv4 format of the source route control type address translation method using the IP source loose source routing option as the intra-packet storage type address translation method of the present invention. In the embodiment in the IPv4 format of the source route control type address translation method used in the present invention, when the ingress node receives the packet 96, the packet for executing the route control of the present invention is identified, and the destination of the IP header 98 is determined. The IP address area is rewritten with the unused IP address Z1, and the destination address D is written into the IP option area to generate the packet 94. Further, the IP option of the packet 94 shown in FIG. 14 includes an unused IP address Z1 and a destination IP address pointer information description area 100 required at an appropriate stage, and an address information description area 102 describing an address field. Is provided.

  In the 4-byte pointer information description area 100, a pointer that points to the MSB position of the destination IP address and a pointer that points to the MSB of the unused IP address in the second layer are arranged. In the embodiment shown in FIG. 14, the IP option pointer information description area 100 is further added with 4 bytes of the pointer information description area and 4 bytes × bytes corresponding to the IP address of the number of layers. The packet 104 transmitted by the routing of the present invention has a total packet length that is 12 bytes longer than the original packet.

  When the packet 96 reaches the ingress node in the first hierarchy, a packet 94 is generated. In the packet 94, the unused IP address Z1 replaces the destination IP address D, and the destination IP address D is written in the address information description area 102 formed at the end of the IP option, and the corresponding pointer information description area 100 is written. The value of is also updated accordingly. Thereafter, when the packet 94 reaches the entry node P2 of the second hierarchy, the packet 104 is generated. In the packet 104, the unused IP address Z2 is written in the destination IP address area, and the unused IP address Z1 is stored in the address information description area 102 together with the destination IP address D. When the packet reaches the entry node P3 of the first layer again, the packet 94 is reproduced again. When the packet reaches the final exit node, the original destination IP address D in the IP option is overwritten at the position of the destination IP address. At the same time, all contents of the IP option are deleted and restored as the original transmission packet 96.

  FIG. 15 is a diagram showing still another embodiment of a packet used in the present invention. In FIG. 15, the original destination IP address and the unused IP address Z1 in the first layer are described using a 24-byte IPv6 routing header 112 in the IPv6 format packet. At the entry node of the second hierarchy, the unused IP address Z2 necessary for the next routing is acquired, the routing of the present invention is executed, and the unused IP address Z1 is again entered at the entry node P3 of the first hierarchy. The first-layer egress node S3 obtains the original destination address D, overwrites Z1, and deletes the contents of the routing header from the packet to the node B that is the user node. Play packets that can be transmitted.

  In the present invention, the tunnel method is not described in detail. However, since the address to be assigned to the new IP header may be a host address, an unused IP address is used in both cases of IPv4 and IPv6. There are no particular restrictions. In each case described above, when an error occurs, the ICMP packet returns an error message to an appropriate edge node, and the packet can be reproduced or a subsequent operation can be determined.

  A program for causing the information processing apparatus to execute the path control method of the present invention is described using a non-object-oriented or object-oriented programming language such as C language, C ++ language, Java (registered trademark), or the like. And can be distributed as storage media or transmission media such as CD, MO, CD-RW, DVD.

  In addition, the present invention has been described by taking a hop-by-whip type packet switching network under protocols such as IPv4 and IPv6 as an example. However, in addition to the above, the present invention uses network layer protocols such as CLNP and IPX. The present invention can be easily applied to a hop-by-hop packet switching network to be used. Further, the present invention provides a packet including a predetermined unused IP address instead of identifying each packet flow in a system that provides services such as bandwidth guarantee and priority transfer corresponding to the flow to which the packet belongs for each packet. The present invention can also be applied to a transaction identification system or a QoS guarantee system such as a bandwidth guarantee in a system for identifying a flow destination IP address and providing a service.

  Although the present invention has been described above, as described above, the present invention can increase an additional element in IP network management that allows the end host to confirm and approve the IP layer path as compared with the conventional MPLS method. It is possible to migrate from a conventional IP network without significantly changing the network equipment and configuration, while maintaining an appropriate bandwidth overhead, and further, a routing system, a routing method and a routing control system excellent in hierarchical stackability A program can be provided. That is, according to the present invention, (1) without requiring large-scale hardware resources such as a new forwarding table, (2) packet flows of various granularities can be identified and handled, (3) from a conventional routing system It is possible to provide a path control system, a path control method, and a program therefor that are easy to migrate. This application is based on the content of the doctoral dissertation of Keio University Graduate School of Media and Governance.

The figure which showed embodiment of the route control system 10 which the hop-by-hop type packet switching network of this invention uses (conventional route control). The figure which showed embodiment of the route control system 10 which the hop-by-hop type packet switching network of this invention uses (route control of this invention). Schematic which showed the function structure of the edge node of this invention. 3 is a schematic flowchart of a route control method according to the present invention. The flowchart which showed the process which the ingress node and core node of this invention perform. The flowchart which showed the process which the exit node of this invention performs. The figure which showed the signaling transaction between each node demonstrated in FIG. 5 and FIG. 6 with the data structure used for signaling. The figure which showed embodiment of the egress node preservation | save type address translation system. The figure which showed the table containing the prefix of the unused IP address which can be included in an exit router, and the prefix of the destination IP address which should be converted correspondingly. The figure which showed embodiment of the address translation system which uses the storage type address translation system in a packet and a tunnel system. The figure which showed embodiment of the hop by hop type packet conversion process in the case of including a destination IP address in a packet. 1 is a schematic diagram of a system that performs route control via a plurality of network hierarchies. The figure which showed embodiment of the packet conversion process for implement | achieving the hop by hop in the case of including a destination IP address in a packet. The figure which showed embodiment of the packet conversion process for implement | achieving the hop by hop in the case of including a destination IP address in a packet. The figure which showed embodiment of the packet conversion process for implement | achieving the hop by hop in the case of including a destination IP address in a packet.

Explanation of symbols

10: Routing system 12 ... User node (Node A)
14: User node (Node B)
16, 18, 20, 22 ... route table 24 ... network 26 ... edge node 28 ... packet classification means 30 ... route table 32 ... address pool management means 34 ... address pool 36 ... inlet flow processing means 38 ... outlet flow processing means 40: Signaling (request / response) processing means

Claims (12)

A flow identification and flow level routing control system for identifying and controlling a packet flow transmitted by a whip-by-whipping packet switching network in a network, the system being connected to the network and the network A plurality of nodes, and at least a first node of the nodes is
Means for transmitting a request for transmitting a packet on a designated transmission path, and securing an unused IP address in the network;
Means for selecting a packet to be transmitted on the designated transmission path;
Means for analyzing the selected packet and replacing a destination IP address with the reserved IP address reserved;
A routing control system comprising: means for sending the replaced packet to another designated node.   The plurality of nodes include a plurality of edge nodes connected to a user node and a core node that relays the packet, and the first node constitutes the edge node, and the edge node And the core node includes a data structure storing an entry associating the reserved unused IP address with the IP address of the next hop node for routing the unused IP address. Item 4. The routing control system according to Item 1.   The sending means describes the packet transmitted through the designated route, describing the reserved unused IP address in the destination IP address field of the IP header, and the destination IP address behind the payload or the IP header. The routing control system according to claim 1, further comprising: means for encapsulating and generating the IP option field or routing header, or adding a new header including the unused IP address.   The sending means replaces the packet transmitted through the designated route by replacing a first predetermined byte of a destination IP address of the packet with a second predetermined byte of the unused IP address. The routing control system according to claim 1, further comprising means for generating an address and encapsulating the packet.   At least one edge node of the plurality of nodes receives the packet including the unused IP address and replaces a second predetermined byte of the unused IP address with the first predetermined byte. 5. The routing system of claim 4, including a data structure for: A flow identification and flow level route control method for performing route control of a packet transmitted in a whip-by-whip type packet switching network in the network including a network and a plurality of nodes connected to the network In the route control method, the first information processing apparatus is used.
A function unit for transmitting a request for transmitting a packet through a designated transmission path and securing an unused IP address in the network;
A functional means for selecting a packet to be transmitted on the designated transmission path;
Functional means for analyzing the selected packet and replacing a reserved IP address with the reserved IP address,
A path control method, comprising: functional means for sending the replaced packet to another designated node. The plurality of nodes include a plurality of edge nodes connected to a user node and a core node that relays the packet, and the first node constitutes the edge node;
In the method, the edge node and the core node have an entry corresponding to the unused IP address and the IP address of the next hop node for routing the reserved unused IP address. The path control method according to claim 6, comprising functional means for generating a stored data structure.   The sending function unit describes the reserved unused IP address in the destination IP address field of the IP header, and the destination IP address in the rear of the payload or the IP header. The route control method according to claim 6 or 7, further comprising functional means for encapsulating and generating in an IP option field or routing header, or functional means for adding a new header including the unused IP address. .   The sending function means replaces the first predetermined byte of the destination IP address of the packet with the second predetermined byte of the unused IP address in the packet transmitted through the designated route. The route control method according to claim 6, comprising functional means for generating a destination address and encapsulating the packet in the packet.   At least one edge node of the plurality of nodes receives the packet including the unused IP address and sends a second predetermined byte of the unused IP address to the first node of the first node. The path control method according to claim 9, comprising functional means for generating a data structure for replacement with a predetermined byte. A flow identification and flow level path control method for performing path control of packets transmitted in a whip-by-whip type packet switching network in the network including a network and a plurality of nodes connected to the network A program for causing an information processing device to execute, wherein the program performs an information processing device,
A function unit for transmitting a request for transmitting a packet through a designated transmission path and securing an unused IP address in the network;
A functional means for selecting a packet to be transmitted on the designated transmission path;
Means for analyzing the selected packet and replacing a destination IP address with the reserved IP address reserved;
Means for generating a data structure storing an entry that associates the unused IP address with the IP address of the next hop node;
A function unit configured to look up the data structure of the replaced packet to a next-hop node and send the packet to a next-hop address. A flow identification and flow level routing system for identifying and controlling a flow of a packet transmitted in a whip-by-whipping type packet switching network in the network, the system including the network and the network A plurality of connected nodes, including at least a first node and a second node of the nodes, wherein the second node is:
Means for receiving a request for transmitting a packet from the first node through a designated transmission path and securing an unused IP address in the network;
Means for notifying an unused IP address secured to the first node in the designated transmission path;
Analyzing means for receiving the second unused IP address converted from the unused IP address and the destination IP address by a predetermined protocol or a packet including the unused IP address and analyzing the packet;
Including
When the analyzing unit detects a second unused IP address in the packet, the analyzing unit reproduces the original destination IP address from the second unused IP address, and the analyzing unit obtains the unused IP address. A routing control system that, when detected, reads a destination IP address described in the packet and reproduces the original destination IP address of the packet.

Images