EP1206859A1 - Verteilter router zur etikettvermittlung - Google Patents
Verteilter router zur etikettvermittlungInfo
- Publication number
- EP1206859A1 EP1206859A1 EP00939871A EP00939871A EP1206859A1 EP 1206859 A1 EP1206859 A1 EP 1206859A1 EP 00939871 A EP00939871 A EP 00939871A EP 00939871 A EP00939871 A EP 00939871A EP 1206859 A1 EP1206859 A1 EP 1206859A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- label
- router
- gateway
- switching
- egress
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/26—Route discovery packet
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/44—Distributed routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/50—Routing or path finding of packets in data switching networks using label swapping, e.g. multi-protocol label switch [MPLS]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/02—Network architectures or network communication protocols for network security for separating internal from external traffic, e.g. firewalls
- H04L63/0272—Virtual private networks
Definitions
- This application relates to data communication networks, and in particular, to interfaces between MPLS networks and non-MPLS networks.
- a network typically includes a collection of routers interconnected with each other. From a data packet's point of view, each router is a stepping stone toward its destination. To traverse the network from its source to its destination, a data packet engages in a sequence of "hops" from one router to the next.
- Each router in a network is a device having an input side with at least one input connected to an adjacent router and an output side with several outputs, each of which is connected to an adjacent router.
- the router receives a data packet through its input and forwards the data packet to an adjacent router through one of its several outputs.
- the function of a router is to decide, for each data packet arriving at its input, which of the several adjacent routers it is to relay the data packet to.
- a router accomplishes this by examining a header associated with the data packet to determine that data packet's destination. Using a look-up table, the router then determines which of the adjacent routers on its output side is the appropriate router for the data packet's next hop toward its destination.
- each router In order for a network to operate correctly, it is preferable that each router understand the information contained in the header associated with the data packet. This can create difficulties when, in order to reach its destination, the data packet traverses a network in which the routers are unable to recognize the information contained within the header. This can occur when a router understands only a protocol other than that used to generate the header.
- An MPLS (“multi-protocol label switching”) network is based on the recognition that information for relaying the data packet on its next hop need be understood only by two adjacent routers.
- each data packet is encapsulated by a label.
- Each router on an MPLS network operates by receiving the labeled data packet through its input side and determining, on the basis of information in the label, which router to relay the data packet to on its output side. Then, the router switches the label with a new label containing information that can be understood by the next router to receive the data packet.
- the label is thus of only local significance and is typically switched to a new label before every hop. Because a router on an MPLS network switches the label associated with the data packet, such a router is referred to as a "label-switching router" or "LSR.”
- an MPLS network need not recognize or understand the header of the data packet.
- the LSRs need only recognize labels.
- a data packet can readily traverse networks in which the routers are incapable of recognizing its header. All that is necessary is that each router in the network be a label switching router. It is for this reason that an MPLS network is a "multi-protocol" network.
- a distributed LSR includes an ingress gateway logically positioned between a first edge LSR and the non-MPLS network.
- This ingress gateway receives a label request from the first edge LSR and instructs a multicast agent to relay this label request from the ingress gateway, across a non-MPLS network, to a plurality of egress gateways.
- the multicast agent also relays responses, if any, from the egress gateways, across the non-MPLS network, and back to the ingress gateway.
- the multicast agent maintains at least one table specifying the egress gateways corresponding to a particular virtual private network.
- a designated gateway which is typically selected from the egress gateways on the basis of whether a connection can be made between that egress gateway and an address specified in the label request message.
- the designated gateway is logically positioned between the second edge LSR and the non-MPLS network.
- the non-MPLS network can be a conventional packet switching network such as an IP network.
- the distributed LSR preferably includes an ingress IP router logically positioned between the ingress gateway and the multicast agent and an egress IP router logically positioned between the egress gateway and the IP network.
- the ingress gateway preferably includes a timer for detecting the lapse of a predetermined interval following transmission of the label request.
- the distributed LSR also includes confirmation means for sending a confirmation message to the egress gateways.
- This confirmation message specifies the designated gateway for the communication path.
- the confirmation message can be sent to all egress gateways responding to the label request message or only to those egress gateways that are not to become the designated gateway.
- the destination requested in a label request message may be accessible by more than output edge router.
- the distributed LSR preferably includes selection means for selecting a designated gateway from a plurality of responses from the plurality of egress gateways. This selection can be made on a first-come-first-served basis, in which case the designated gateway is the first egress gateway to respond, or by determining which of the responding egress gateways is the optimal egress gateway.
- the distributed LSR receives a label request from a first edge LSR and multicasts that label request across the non-MPLS network to a plurality of egress gateways. At least one egress gateway then generates a response indicating that there exists a communication path between a second edge LSR and that responding egress gateway. Upon receipt of this response, the ingress gateway then designates the responding egress gateway to be the designated gateway for transmission of data to the second edge LSR.
- first and second LSRs are associated with NPNs
- the ingress gateway waits only a predetermined interval for responses before assuming that communication is not currently available. This ensures that the ingress gateway will not wait an interminable period for a response.
- the ingress gateway preferably sends a confirmation message to the egress gateways.
- This confirmation message can be sent to all egress gateways or all egress gateways other than the designated gateway.
- the invention thus provides a method and system for interfacing a non- MPLS network with an MPLS network in a gradual and controlled manner without the need to replace each router in the non-MPLS network with an LSR. This is accomplished by having the entire non-MPLS network appear to the MPLS network as a single, virtual, LSR. This is a significant advantage because the conversion of a potentially large number of routers in a non-MPLS network requires a period during which the network will be non-operational.
- the single, virtual LSR of the invention is a distributed LSR because it performs the functions of a conventional LSR through the cooperation of several routers.
- the distributed LSR accepts a labeled data packet at its input port, switches the label on the data packet, and sends the data packet, with its new label, out a selected output port.
- the distributed LSR of the invention does not rely on a label information base to accomplish label switching. Instead, an ingress gateway in the distributed LSR instructs a multicasting agent to multicast a label request, across a non-MPLS network, to each of a plurality of egress gateways. If a route to a destination from a particular egress gateway is available, that egress gateway then sends a label mapping message back to the ingress gateway.
- the ingress router selects one of the egress routers to complete the LSP (label-switching protocol).
- the distributed LSR then creates a tunnel within the non-MPLS network for communication between the ingress router and the egress router.
- FIG. 1 is a schematic diagram of the architecture of a distributed LSR according to the invention.
- FIG. 2 is the distributed LSR of FIG. 1 but with two output edge LSRs communicating with the same output MPLS network;
- FIG. 3 is a flow chart of the process carried out by the distributed LSR of FIG. 1.
- a distributed label-switching router (LSR) 10 interfaces a non-MPLS network 12 with both an input MPLS network 14 and with a plurality of output MPLS networks 16.
- the input MPLS network 14 and the plurality of output MPLS networks 16 form two parts of a virtual private network (VPN).
- the non-MPLS network is an IP network 12, however other packet switching networks can be substituted for the IP network 12.
- the IP network 12 is a public network that links these two parts of the VPN.
- data preferably travels along a secure path when traversing the IP network 12. This is achieved by establishing a tunnel across the IP network 12.
- the distributed LSR 10 dynamically establishes such tunnels on an as-needed basis.
- the input MPLS network 14 connects to the distributed LSR 10 through an input edge LSR 18.
- the output MPLS networks 16 connect to the distributed LSR 10 through a plurality of output edge LSRs 20.
- Both the input edge LSR 18 and the output edge LSRs 20 are MPLS egress routers ranning LDP and operating in downstream on-demand mode with conservative label retention. Both the input edge LSR 18 and the output edge LSRs 20 perform the ingress and egress functions specified by the conventional MPLS architecture.
- the distributed LSR 10 uses MPLS signaling rather than conventional frame relay signaling to minimize the extent of the routing infonnation exchanged between ingress router and the egress router. This results in a simpler and more scalable network.
- Each edge LSR shown in FIG. 1 communicates with the distributed LSR 10 tlirough a frame-relay interface. However, the edge LSRs can also communication with the distributed LSR through other WAN interfaces.
- the distributed LSR 10 interfaces with the input MPLS network 14 tlirough an ingress router 22 connected to the input edge LSR 18.
- An ingress gateway 24 logically positioned between the input edge LSR 18 and a first IP router 26 provides the interface between the ingress router 22 and the input edge LSR 18.
- the ingress router 22 interfaces with the IP network 12 that provides communication paths between the ingress router 22 and a plurality of egress routers 30,
- FIG. 1 shows a first egress router 32 in communication with a first output network 34 tlirough a first output edge LSR 36.
- FIG. 1 shows second and third egress routers 38, 40 in communication with second and third output networks 42, 44 through second and third output edge LSRs 46, 48 respectively.
- the plurality of egress routers 30 can include any number of constituent egress routers.
- the structures of the egress routers 32, 38, 40 mirror that of the ingress router 22.
- the first egress router 32 includes a first egress gateway 50 logically positioned between a second IP router 52 and the first output edge LSR 36.
- the first egress gateway 50 interfaces with the first output edge LSR 36 while the second IP router 52 interfaces with the IP Network 12.
- the second and third egress routers 38, 40 include second and third egress gateways 54, 56 logically positioned between the third and fourth IP routers 58, 60 and the second and third output edge LSRs 46, 8 respectively.
- the second and third egress routers 38, 40 interface with the IP network 12 through the third and fourth IP routers 58, 60 respectively and with the second and third output edge LSRs 46, 48 through the second and third egress gateways 54, 56 respectively.
- the input edge LSR 18 and the ingress gateway 24 establish an LDP (label distribution protocol) session between each other using conventional discovery and initialization procedures. This is equivalent to setting up an LDP session between two LSRs in conventional MPLS networking. Little or no routing information is exchanged between the ingress gateway 24 and the input edge LSR 18 during establishment of an LDP session. It is sufficient that the input edge LSR 18 use the ingress gateway 24 as a default router or default next hop for addresses within the NPN. This implies that to resolve an address within the VPN, the input edge LSR 18 sends the ingress gateway 24 a label request message specifying a destination address.
- LDP label distribution protocol
- the ingress gateway 24 communicates the label request message to the multicast agent 28.
- the multicast agent 28 For each VPN serviced by the distributed LSR 10, the multicast agent 28 maintains a router table 62 listing all egress routers associated with that VPN. As a result, changes to the topology of a particular VPN are easily implemented by adding and deleting egress routers from the appropriate router table 62.
- the multicast agent 28 relays the label request to the all the egress routers in its router table 62, which in the example of FIG. 1 includes the first, second, and third egress routers 32, 28, 40.
- the first, second, and third egress routers 32, 28, 40 then forward the label request to the first, second, and third output edge LSRs 36, 46, 48, each of which is has established an LDP session with its corresponding egress gateway 50, 54, 56,
- Each of the output edge LSRs 36, 46, 48 includes a routing table listing all destinations accessible by that edge LSR.
- Each of the first, second, and third output edge LSRs 36, 46, 48 determines, independently of the others, whether the destination address specified in the label request message is within its respective routing table. If one of the output edge LSRs, for example the first output edge LSR 36, determines that the destination address is in its routing table, it sends a label mapping message back to the first egress gateway 50 to which it is connected. The first egress gateway 50 then forwards the label mapping message across the IP network 12 to the ingress gateway 24.
- the ingress gateway 24 Upon receipt of a label mapping message, the ingress gateway 24 designates the responding egress gateway (in this case the first egress gateway 50) to be the designated gateway for the ensuing label switched path (LSP) between the ingress gateway 24 and the responding output edge LSR 36.
- the first egress gateway 50 which is now the designated gateway, then establishes a tunnel across the IP network 12 between itself and the ingress gateway 24. This establishes an LSP path between the input edge LSR 18 and the first output edge LSR 36.
- the resulting LSP path includes three distinct path segments: (1) a segment from the input edge LSR 18 to the ingress gateway 24, which is labeled in normal MPLS fashion; (2) a segment from the ingress gateway 24 to the first egress gateway 50, in which traffic is carried in an IP tunnel; and (3) a segment from the first egress gateway 50 to the first output edge LSR 36, which is labeled in normal MPLS fashion.
- the ingress gateway 24 upon receipt of a label mapping message from any one of the egress gateways 50, 54, 56, instructs the multicast agent 28 to relay to all the egress gateways 50, 54, 56 a confirmation message containing the identity of the designated gateway.
- Each egress gateway 50, 54, 56 checks the identity of the designated gateway specified in this confirmation message. If the identity of the designated gateway differs from that of the egress gateway, the egress gateway releases the LSP at its interface with its corresponding output edge LSR. If the output edge LSR connected to the egress gateway has already responded, this is achieved by sending a label abort notification message to the output edge LSR.
- the ingress gateway 24 can instruct the multicast agent 28 to send a message only to those egress gateways that are not the designated gateway. This message instructs those egress gateways to abort any LSP established in response to the original label request message.
- the ingress gateway 24 includes a timer 64 that begins counting down upon transmission of the label request message. If, at the expiration of a predetermined interval the ingress gateway 24 fails to receive any response from the egress gateways 50, 54, 56, the ingress gateway 24 sets the address of the designated gateway to be a null address and sends a confirmation message, as described above.
- a particular destination address may be in two different router tables associated with two different output edge routers.
- the ingress gateway 24 may receive two label mapping messages.
- a VPN topology of the type shown in FIG. 2 is both uncommon and undesirable, it is preferable that the ingress gateway 24 be provided with a mechanism for designating an egress gateway for the LDP session from a plurality of responding egress gateways. This choice can be made randomly, for example on a first-come-first-served basis. Alternatively, the choice can be made on the basis of which egress gateway is perceived to be optimal. For example, if one of the two responding egress gateways is notoriously busy, the designated egress gateway might be the less busy egress gateway.
- the process of establishing a communication path between the input edge LSR and the output edge LSR begins with the receipt 66, by the ingress gateway, of a label request message from the input edge LSR.
- the ingress gateway then instructs 68 the multicast agent to multicast the label request message across the IP network to the egress gateways.
- the input edge LSR sets 70_a timer and waits 72 for a response.
- the ingress gateway checks 74 the timer to see if a waiting interval has elapsed. If the waiting interval has not yet elapsed, the ingress gateway continues 72 to wait for a response. If the waiting interval has elapsed and no response has been received from an egress gateway, the ingress gateway sets 76 the designated gateway to be a null gateway. A confirmation message specifying this designated gateway is then multicast 80 to the egress gateways.
- each egress gateway receives 82 the multicast label request message and establishes 84 an LDP session with output edge router to determine whether the destination specified in that message is accessible to it. If it is not, the egress gateway does nothing 86. If it is, then the egress gateway sends 88 a label mapping message back to the ingress gateway.
- the ingress gateway Upon receipt 90 of the label mapping message, the ingress gateway sets 92 the egress gateway that sent the label mapping message to be the designated gateway. A confirmation message specifying the designated gateway is then multicast 80 to the egress gateways.
- the ingress gateway selects a designated gateway, either on a first-come-first-served basis or by selecting the optimal egress gateway to be the designated gateway.
- Each egress gateway then receives 94 the confirmation message and determines 96 whether it is to become the designated gateway. If the egress gateway determines that it is to become the designated gateway, it proceeds to establish 98 a tunnel across the IP network between itself and the ingress gateway. If the egress gateway determines that it is not to become the designated gateway, it proceeds to tear down 100 the LDP session, if any, established between itself and an output edge router.
- the ingress and egress gateways cooperate with a multicasting agent in the manner set forth above, encapsulates a non-MPLS network so that, from the point of view of an edge LSR, it appears to be no more than another LSR. This is achieved without the burden and expense of having to convert each router within the non-MPLS network into an LSR.
- the distributed LSR dynamically establishes tunnels through the non-MPLS network on an as-needed basis, thereby simplifying the task of adding and removing customers from a NPN.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/US2000/016361 WO2001097460A1 (en) | 2000-06-14 | 2000-06-14 | Distributed label switching router |
Publications (1)
Publication Number | Publication Date |
---|---|
EP1206859A1 true EP1206859A1 (de) | 2002-05-22 |
Family
ID=21741489
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP00939871A Withdrawn EP1206859A1 (de) | 2000-06-14 | 2000-06-14 | Verteilter router zur etikettvermittlung |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP1206859A1 (de) |
CA (1) | CA2363732A1 (de) |
WO (1) | WO2001097460A1 (de) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006002543A1 (en) * | 2004-07-05 | 2006-01-12 | Versa Power Systems, Ltd. | Flexible high temperature cables |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2397195B (en) * | 2000-10-30 | 2004-09-22 | Siemens Plc | End-to-end voice over ip streams for telephone calls established via legacy switching systems |
KR100441891B1 (ko) * | 2002-06-15 | 2004-07-27 | 한국전자통신연구원 | 이종의 다중 프로토콜 라벨 스위칭 네트워크 장비를 원격제어하기 위한 구성 정보 자료구조 및 관리장치, 관리방법 |
KR100884184B1 (ko) * | 2002-06-27 | 2009-02-17 | 주식회사 케이티 | Mpls vpn에서의 멀티캐스트 트리 설정/해지 방법및 멀티캐스팅 서비스 제공 방법 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3384319B2 (ja) * | 1998-03-27 | 2003-03-10 | 日本電気株式会社 | 仮想専用網構築システム |
-
2000
- 2000-06-14 CA CA002363732A patent/CA2363732A1/en not_active Abandoned
- 2000-06-14 EP EP00939871A patent/EP1206859A1/de not_active Withdrawn
- 2000-06-14 WO PCT/US2000/016361 patent/WO2001097460A1/en not_active Application Discontinuation
Non-Patent Citations (1)
Title |
---|
See references of WO0197460A1 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006002543A1 (en) * | 2004-07-05 | 2006-01-12 | Versa Power Systems, Ltd. | Flexible high temperature cables |
Also Published As
Publication number | Publication date |
---|---|
CA2363732A1 (en) | 2001-12-20 |
WO2001097460A1 (en) | 2001-12-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6947428B1 (en) | Distributed label switching router | |
EP1563644B1 (de) | System und verfahren zur verbindung von heterogenen layer-2 vpn anwendungen | |
EP1213881B1 (de) | System und Verfahren zum Herstellen eines Kommunikationsweges auf einer ATM Plattform | |
EP1678873B1 (de) | Autodiscovery für virtuelle netzwerke | |
US8151000B1 (en) | Transparently providing layer two (L2) services across intermediate computer networks | |
US7440438B2 (en) | Refresh and filtering mechanisms for LDP based VPLS and L2VPN solutions | |
US20030053450A1 (en) | Layer 2-VPN relay system | |
US20030133412A1 (en) | VLAN to MPLS mapping: method to establish end-to-end traffic path spanning enterprise local area networks and a global network | |
US20060168279A1 (en) | Apparatus and method for providing multi protocol label switching (MPLS)-based virtual private network (VPN) | |
US8279881B2 (en) | Method and system for route updating | |
US20050232263A1 (en) | Communication control apparatus, communication network and method of updating packet transfer control information | |
CN102739501A (zh) | 二三层虚拟私有网络中的报文转发方法和系统 | |
US20150146573A1 (en) | Apparatus and method for layer-2 and layer-3 vpn discovery | |
EP2640014B1 (de) | Verfahren und netzwerkvorrichtung zur verteilung von mpls-etiketten | |
CN101415005B (zh) | 一种实现业务转发的方法、系统和设备 | |
JP2000341294A (ja) | パケット中継装置 | |
CN100502343C (zh) | 多协议标签交换虚拟专用网相互通信的方法 | |
CA2376690A1 (en) | System and method for loop avoidance in multi-protocol label switching | |
WO2001097460A1 (en) | Distributed label switching router | |
Cisco | Troubleshooting Tag and MPLS Switching Connections | |
CN100559755C (zh) | 自动发现虚拟网络的方法和系统 | |
Cisco | Troubleshooting Tag and MLPS Switching Connections | |
CN107547350B (zh) | 流量处理方法及装置 | |
KR100431207B1 (ko) | 엠피엘에스(mpls)기반망에서의 엑스트라넷아이피-브이피엔(ip-vpn)서비스 제공 방법 | |
JP2002290441A (ja) | Ip−vpnルータおよびip−vpn用パケット転送路自動設定方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 20020620 |
|
RBV | Designated contracting states (corrected) |
Designated state(s): DE FR GB |
|
GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
GRAS | Grant fee paid |
Free format text: ORIGINAL CODE: EPIDOSNIGR3 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION HAS BEEN WITHDRAWN |
|
18W | Application withdrawn |
Effective date: 20051209 |