802.1aq support over ietf evpn

[Cross-reference to related applications] This application claims priority from US Provisional Patent Application No. 61 / 645,431 filed May 10, 2012.

[Field of invention] Embodiments of the present invention relate to the field of computer networking. Specifically, the embodiment relates to the interworking of 802.1aq SPBM with the IETF EVPN.

The IEEE 802.1aq standard (hereafter referred to as 802.1aq), published in 2012, defines a routing solution for Ethernet that replaces traditional routing solutions, including the Spanning Tree Protocol. 802.1aq is also known as shortest path bridging or SPB. 802.1aq allows the creation of logical Ethernet networks on the native Ethernet infrastructure. 802.1aq employs a link-state protocol (ie, IS-IS (Intermediate System to Intermediate System)) to advertise the network topology and the logical network membership of nodes in the network. The link state information is used to calculate the shortest path tree from all bridges in the SPB region. The computation is done on each node in an independent and decentralized manner, and its own transfer table is generated to implement that part of the transfer of the entire area of the SPB.

The data is encapsulated at the edge nodes of the network that implements 802.1aq. This encapsulation can be within an 802.1ah frame or a tagged 802.1Q / p802.1ad frame. These frames are only forwarded to the other members of their respective logical networks. Unicast and multicast are also supported by 802.1aq. All such routing is done via the shortest symmetrical path. Multiple equal cost shortest paths are supported. The implementation of 802.1aq in networks simplifies the generation and configuration of various types of networks, including provider networks, corporate networks and cloud networks. The configuration is relatively simple and reduces the likelihood of errors, especially man-made configuration errors, compared to traditional routing solutions. 802.1aq also increases bandwidth and reliability through the improved use of network mesh topologies. By enabling all paths, higher utilization can be obtained through the use of multiple equal cost paths. It can also support improved convergence times and larger topologies.

EVPN is a new approach for performing L2VPN over MPLS using the BGP protocol. EVPN brings several improvements to previous L2VPN technology, one of which is the use of a BGP control plane for mirroring Ethernet FDBs across a set of supported MPLS PEs.

In MPLS (multiprotocol label swapping) provider edge switch (PE), a method for interworking the 802.1aq control plane with the BGP (Border Gateway Protocol) control plane of EVPN (Ethernet Virtual Private Network) is implemented. The above method selects a designated forwarder (DF) for a given B-VID (Backbone-Virtual Local Area Network Identifier) in the local PBBN (Provider Backbone Bridged Network), thereby identifying the B-VID. I-SID (I-Component Source Identifier) and MAC (Media Access Control) information of IS-IS (Intermediate System-Intermediate) System) It is decided whether to move the database uniquely from the BGP database to the BGP database and to move the I-SID and MAC information in the BGP database to the IS-IS database of the PE. The above method receives an IS-IS PDU (protocol data unit) including an IS-IS TLV (Type Length Value) in the above PE. It is determined whether the received IS-IS PDU data requires a configuration change as to which PE is the target of the B-VID and which is the DF. It is determined whether the PE is still a DF for one or more B-VIDs. NLRI (network layer reachability) related to the DF from the BGP database according to the fact that the PE is no longer the DF. information) is deleted, and the remote NLRI origin information about the DF is deleted from the IS-IS database. The DF-related NLRI is added from the IS-IS database to the BGP database in response to the PE becoming the DF for at least one B-VID, and the IS-IS database from the BGP database. A local DF-related remote SPBM (shortest path bridging MAC mode) sub-tenant is added to.

It is an MPLS (multiprotocol label swapping) provider edge switch (PE) for interworking the 802.1aq control plane with the BGP (Border Gateway Protocol) control plane of EVPN (Ethernet Virtual Private Network). The set of PEs to attach to the PBBN selects the designated forwarder (DF) for a given B-VID (Backbone-Virtual Local Area Network Identifier) in the local PBBN (Provider Backbone Bridged Network), thereby which PE IS-IS (Intermediate System-Intermediate) provides specific I-SID (I-Component Source Identifier) and MAC (Media Access Control) information for the above B-VID. System) It is decided whether to move the database uniquely from the BGP database to the BGP database and to move the I-SID and MAC information in the BGP database to the IS-IS database of the PE. The PE includes a control plane interworking function, an IS-IS module, and a processor configured to execute a BGP module. The BGP module is configured to manage the BGP database and communicate with the peer's BGP speaker, and the IS-IS module is an IS-IS PDU (protocol data) including an IS-IS TLV (Type Length Value). It is configured to receive unit) and manage the above IS-IS database. The control plane interworking function is received by the IS-IS. It is configured to determine if the PDU's data requires a DF configuration change. Also, the control plane interworking function determines if the PE is still a DF for one or more B-VIDs, and depending on the PE being no longer the DF, the DF from the BGP database. The relevant NLRI (network layer reachability information) is also configured to remove remote NLRI origin information about the DF from the IS-IS database. The control plane interworking function transfers the DF-related NLRI from the IS-IS database to the BGP database, depending on the PE becoming the DF for at least one B-VID. It is configured to add a local DF-related remote SPBM (shortest path bridging MAC mode) sub-tenant to the above IS-IS database.

The data plane transfer feature on MPLS PE makes certain modifications to Ethernet frames traversing from EVPN to PBBN. This is the addition or overwriting of B tag information to map I-SID traffic to the local configuration and the SP source ID in the MAC destination address of the multicast frame to reflect the value assigned to the local PE. Including overwriting.

The present invention is shown in the drawings of the accompanying drawings in an exemplary manner, but not limited, in which similar reference numerals point to similar elements. It should be noted that different references to "one" or "one" embodiment in the present disclosure do not necessarily refer to the same embodiment, and those references refer to at least one embodiment. Should be. Furthermore, if a particular feature, structure, or property is described in relation to one embodiment, is it explicitly stated that such feature, structure, or property is applied in connection with another embodiment? It should be presented that it is within the knowledge of those skilled in the art, whether or not it is.

<figref num="1">FIG. 5 is a diagram of one embodiment of an exemplary PBB and SPBM-PBBN EVPN network that implements 802.1aq support over EVPN.</figref><figref num="2">It is a figure of one embodiment of the process of converting a frame from PBBN to EVPN and from EVPN to PBBN.</figref><figref num="3">It is a diagram of one embodiment of PE that implements 802.1aq on EVPN.</figref><figref num="4">It is a flowchart of one embodiment of the process for handling the IS-IS TLV received by the PE. This process is started in response to the reception of an IS-IS PDU such as TLV.</figref><figref num="5">It is a flowchart of one embodiment of the process for BGP NLRI processing.</figref><figref num="6">An example of a network element that can be used to implement one embodiment of the present invention is shown.</figref>

In the following description, many specific details will be described. However, it should be understood that embodiments of the present invention can be practiced without their specific details. Other examples do not detail well-known circuits, structures, and techniques so as not to obscure the understanding of this description. However, it will be appreciated by those skilled in the art that the present invention can be practiced without such specific details. The description contained herein will allow one of ordinary skill in the art to implement appropriate functionality without undue experimentation.

The operation of the flow diagram will be described with reference to exemplary structural embodiments shown in the diagram. However, the operation of the flow diagram can be performed by structural embodiments of the invention other than those discussed with reference to the drawings, and embodiments discussed with reference to the drawings refer to the flow diagram. It should be understood that it is possible to perform actions different from those discussed in.

The illustrated technique can be implemented using code and data stored and executed on one or more electronic devices (eg, end stations, network elements, or similar devices). Such electronic devices are machine readable, such as non-temporary machine readable or computer readable storage media (eg, magnetic disks, optical discs, random access memory, read-only memory, flash memory devices, phase change memory). Or use computer-readable media to store and / or communicate codes and data (internally and / or with other electronic devices over networks). In addition, such electronic devices typically include one or more storage devices, user input / output devices (eg, keyboards, touch screens, and / or displays), and one or more other devices such as network connections. Contains a set of one or more processors attached to a component of. The connection between the set of processors and other components is typically done through one or more buses and bridges (also known as bus controllers). A storage device represents one or more non-transitory machine-readable or computer-readable storage media, and a non-temporary machine-readable or computer-readable communication medium. Thus, a storage device for a given electronic device typically stores code and / or data for execution on one or more sets of processors for that electronic device. Of course, one or more parts of one embodiment of the invention may be implemented using different combinations of software, firmware and / or hardware.

As used herein, a network element (eg, router, switch, bridge, etc.) is communicably interconnected with other equipment on the network (eg, other network element, end station, etc.). A network device that includes hardware and software. Some network elements are "multi-service network elements", which are multiple networking functions (eg, routing, bridging, switching, Layer 2 integration, session border control, multi-casting), and / or subscription. Provide support for network operations) and / or support for multiple application services (eg, data, audio, and video). Subscriber end stations (eg servers, workstations, laptops, palmtops, mobile phones, smartphones, multimedia phones, VOIP (Voice Over Internet) Protocol) phones, portable media players, GPS units, game systems, set-top boxes (STBs, etc.) are content / services provided on the Internet and / or virtual private networks (VPNs) that are overlaid on the Internet. Access the content / services provided above. Their content and / or services are typically provided by a service provider or one or more end stations belonging to the content provider (eg, server end stations) or by end stations participating in peer-to-peer services and public web pages. Includes (eg, free content, storefront, search services, etc.), private web pages (eg, web pages that provide email services, accessed by username / password, etc.), corporate networks on VPN, IPTV, etc. obtain. Typically, subscriber end stations are connected to edge network elements (eg, through customer premises equipment (eg, wired or wirelessly) connected to an access network), and those edge network elements are (eg, one or more). Connected to other end stations (eg, server end stations) through core network elements to other edge network elements.

The following abbreviations are used here and are provided for reference: BCB --Backbone Core Bridge, BEB --Backbone Edge Bridge, BGP --Border Gateway Protocol, BU --Broadcast / Unknown, CE --Customer Edge, C-MAC --Customer / Client MAC Address, CP --Control Plane, DF --Designated Forwarder, ECT --Equal Cost Tree, ESI --Ethernet Segment Identifier, EVI --E-VPN Instance, EVN --EVPN Virtual Node, EVPN --Ethernet VPN, I-SID --I Component Service ID, IS-IS --Intermediate Service Intermediate Service, ISIS-SPB --IS-IS as extended for SPB, LAG --Link Aggregation Group, MAC --Media Access Conrol, PE --MPLS Edge Switch, MPLS --Multi-Protocol Label Switching, MP2MP --Multipoint to Multipoint, MVPN --Multicast VPN, NLRI --Network Layer Reachability Information, OUI --Organizationally Unique ID, PBB --Provider Backbone Bridge, PBB-PE --Co-located BEB and PE, PBBN --Provider Backbone Bridged Network, P2MP --Point to Multipoint, P2P --Point to Point, RD --Route Distinguisher, RPFC- Reverse Path Forwarding Check, RT --Route Target, SPB --Shortest Path Bridging, SPBM --Shortest Path Bridging MAC Mode, TLV --Type Length Value, VID --VLAN ID, VLAN --Virtual Local Area Network, VPN --Virtual Private Network.

Embodiments of the present invention provide methods and systems for avoiding the above drawbacks of prior art. The disadvantage is that the Provider Backbone Bridging Ethernet Virtual Private Network (PBB EVPN) approach does not consider the possibility or implementation of the Provider Backbone Bridging Network (PBBN), and the prior art is attached in a complex manner. It also includes, under normal circumstances, no guarantee of symmetric congruence across core networks, which is an important requirement for PBBN. Thus, the necessary procedures and requirements for implementing PBB EVPN have not been properly or fully defined in the past.

Embodiments of the present invention overcome the above drawbacks of prior art. In an embodiment of the invention, the backbone edge bridge encodes a backbone media access control (B-MAC) address and a service instance identifier (I-SID) in the MAC advertising route information of network layer reachability information (NLRI). That, to provide. In addition, the Multiprotocol Label Switching (MPLS) labels offered in this NLRI are common to all BEB / I-SID NLRIs that share a common backbone virtual LAN identifier (B-VID) in the subtending PBBN. is there. It provides a mechanism for guessing the B-VID locally and reconstructing multipath symmetry congruence for traffic as it passes from EVPN to the opposite PBBN. It also allows decoupling of multipath designs between the various SPBM-PBBNs interconnected with EVPN.

[Overview of IEEE802.1aq] IEEE802.1aq uses a link-state protocol to control the forwarding of Ethernet frames over the network. One link-state protocol, IS-IS (Intermediate System to Intermediate System), is used in 802.1aq networks to advertise both network topology and logical network membership.

802.1aq has two modes of operation. The first mode for VLAN (Virtual Local Area Network) based networks is called the Shortest Path Bridging VLAN ID (SPBV). The second mode for MAC-based networks is called Shortest Path Bridging MAC (SPBM). Both SPBV and SPBM networks can support more than one set of equal cost transfer trees (ECT sets) at the same time in the data plane. An ECT set is typically associated with multiple shortest path VLAN identifiers (SPVIDs), forming an SPVID set for SPBVs, and one-to-one with backbone VLAN IDs (B-VIDs) for SPBMs. ..

According to 802.1aq's MAC mode, network elements within a provider network have different paths across the network with different frames mapped to different B-VIDs with the same destination address (called "multipath instances"). It is configured to perform multipath forwarding of data traffic separated by B-VID so that it can be forwarded above. The customer data frame associated with the service is encapsulated with a header with a separate service identifier (I-SID) and B-VID according to 802.1aq. This separation allows services to be scaled independently of the network topology. Therefore, the B-VID can be used exclusively as an identifier for a multipath instance. The I-SID identifies the specific service to be provided by the multipath instance identified by the B-VID. The actual routing of multipath instances in 802.1aq is determined by tie-breaking based on the system ID of each node.

EVPN is an Ethernet VPN protocol solution over MPLS that uses BGP to spread VPN and MAC information and uses MPLS as the transport protocol. At least one embodiment determines the amount of state, topology information, node nicknames and B-MACS leaking from BGP to the IS-IS control plane of each opposing SPBM-PBBN (through the need to know filtering). By minimizing it, we pursue the interconnection of the opposing 802.1aq networks (called SPBM-PBBN) while operatingly separating the SPBM-PBBN. These embodiments also aim to isolate the network design with a separate number of multipathing instances in each opposite SPBM-PBBN. These embodiments also minimize the amount of configuration required. The embodiment provides a mechanism and procedure for appropriately resolving the presence of B-MAC in more than one B-VID once the B-VID context has been removed. The embodiment is BEB It also allows B-MAC to be present in multiple B-VIDs within the MSTP control network (802.11ah) and within the IS-IS (802.1aq) control Ethernet network.

Traditional EVPN implementations do not have a PBBN attached to any PE, and there are no Ethernet networked components in the backbone. Therefore, B-VID does not exist in these conventional systems, and its function is null. There is no requirement for a symmetric congruent multicast tree, and EVPN itself is not congruent across MPLS networks. Therefore, in past networks, the unit of load distribution that is configured is the I-SID. EVPN is devoting a lot of energy to "active-active" multi-chassis uplinks (also known as multi-chassis link aggregation groups (MC-LAGs)).

The traditional designated forwarder (DF) approach assigns an I-SID to a specific PE. This requires all B-MACs that support the I-SID to transit towards that particular PE. However, the B-MAC can be associated with other I-SIDs that have different PEs to pass through to the core network, i.e., those I-SIDs are associated with different DFs. Therefore, B-MAC appears to be associated with multiple PEs within the same B-VID range. This creates what is considered a duplicate MAC problem, which is inconsistent with the implementation of Ethernet switching hardware. This issue forces all I-SIDs that have a common BEB to share a common DF, and enforces unsustainable requirements across additions, moves and changes within the network, and thus ultimately Will eventually cause a failure in the network because it is constructed in a way that is inconsistent with the capabilities of the data plane.

Explicit pinning of DFs for I-SID / B-MAC tuples causes a lot of configuration overhead. A rule is needed to allocate all I-SIDs associated with the B-VID for the BEB in the same PE, which introduces a great deal of operational complexity. Separate the representation of B-MAC from the I-SID allocation type DF. EVPN was introduced for B-MAC only, but because it generated a non-joint unicast / multicast tree, SPBM RPFC should be confused and fail for most multicast traffic. Calculation of the shortest path to the virtual node of EVPN was introduced to select the PE that advertises the I-SID / B-MAC tuple. However, the problem with this is the change in the shortest path in ISIS-SPB, resulting in the withdrawal and re-advertising of much information.

In the embodiment, interworking between BGP and SPBM IS-IS is envisioned. EVPN PE can convert between BGP / MPLS on the EVPN side and standard 802.1aq on the PBBN side. In embodiments, it is possible to run the PBB-EVPN model with SPBM-EVPN. In such an embodiment, the co-located BEB / PE is a larger SPBM. You can fully participate in the PBBN set. The embodiment provides a solution for mapping PBNs tagged with existing S tags to EVPN instances. In one embodiment, the B-VID is subject to allocation. Providing allocations to individual I-SIDs results in a large number of configurations, which can result in DFs being configured such that B-MACs appear on multiple PEs at the same B-VID within the same PBBN. It shouldn't have worked. Rather, the above embodiment provides that PE is selected as DF for a given B-VID within the PBBN. A given PE is designated as a transit device for the set of I-SIDs associated with a given B-VID. The PE inherits the B-VID designation and advertises the PE as a DF for all I-SIDs advertised by SPBM-PBBN within that segment for that B-VID (note that given I). -SID value exists only in a single B-VID at a given point in time in SPBM-PBBN). The PE may use a common MPLS label for B-VIDs for all I-SID advertisements (advertised with "Type + 2" NLRI). Therefore, it is advertised in the BGP NLRI information element specific to SPBM and PBB and structured to include: RD (route distinguisher) important for BGP; Ethernet segment identifier (for local SPBM-PBBN). Unique identifier of); Ethernet tag ID including I-SID and Tx / Rx attribute, Tx / Rx attribute is the logical OR of all BEBs in BGPN registered for that I-SID; I- The BEB's B-MAC address associated with the SID; and the MPLS label value to use when encapsulating the frame for that PE.

In the above embodiment, the PE logically looks like a BEB in the PBBN. When a segment connected to another EVPN advertises the I-SID corresponding to the I-SID registered in the local PBBN and the PE is the DF for the B-VID associated with the I-SID. , The PE advertises the target of the remote I-SID to ISIS-SPB, otherwise the PE is silent. This process keeps the global I-SID table outside of a separate ISIS-SPB instance.

To enable this functionality, a configuration is specified. The PE is configured as a BEB and has a node nickname (SP source ID) for multicast along with the ISIS system ID. The choice of DF is further described below. The configuration of the DF selection can, in other embodiments, be by other means as long as the selection can be determined by decentralized means, whether algorithmic or preset, the requirement is: Each PE can independently and separately select the (in isolation) DF and get the same solution as the other nodes. A segment ID is assigned to each PBBN. A route target (RT) is assigned to a set of PBBNs in a larger multi-site VPN.

In the interworking operation of the control plane between ISIS-SPB and BGP, NLRI data elements are received by the PE from other BGP speakers. The PE has already been determined to be a DF for B-VID by the DF selection procedure described below. PE learns local I-SID / B-VID bindings in PBBN from IS-IS. PE accepts SPBM-PBB NLRI advertisements received from BGP. If the PE does not have local knowledge of the I-SID (no object registered in the local PBBN), the PE simply retains the BGP information for future use. If the PE is a DF for the I-SID (the I-SID is associated with the target B-VID for which the PE is a DF), the PE adds an entry to the EVPN mapping table and the B_MAC / I_SID Indicates the label stack to be used, which is the MPLS label from the SPBM-PBB NLRI ad and the RD (route) advertised in the SPBM-PBB NLRI ad. Includes FEC label for destination). The PE contains a B-MAC, a list of I-SIDs associated with that B-MAC in a BGP table, and a base VID (B-VID associated with an I-SID learned from ISIS-SPB). Organize a "SPBM service identifier and unicast address sub-TLV" and add it to the IS-IS database. This process may include adding an I-SID to an existing IS-IS sub-TLV as a means of updating the IS-IS database.

In the interworking operation of the control plane for SPBM, the service identifier and unicast address sub-TLV are received by the PE from other IS-IS speakers in the PBBN. The PE checks if it is the DF for the B-VID in the sub TLV. If the PE is a DF for the B-VID, the PE organizes a BGP NLRI for each I-SID listed within the TLV and not previously seen, where the RD is the IP address of the PE. The segment ID is the segment ID for the local PBBN, the tag ID = I-SID and the tx / rx attribute of the I-SID, and the MAC address is the B-MAC address from the sub TLV. Note that the NLRI schema is based on evolving standards and can take into account alternative NLRI coding, such as advertising multiple I-SID registrations per B-VID, which adds separate elements. And for deletion, suggest modification of existing BGP database records.

In data plane operation for EVPN interworking with PBBN, every MPLS unicast packet coming from the EVPN MPLS network to the PE has a B-VID to overwrite in the frame. This can be inferred from the MPLS label (if it is uniquely advertised for all I-SIDs in a single B-VID) or from the I-SID. All MPLS multicast packets arriving from EVPN to PE have a B-VID inferred from the I-SID, because the label is managed by the source rather than by the destination. Therefore, the role of guessing B-VID is also not assigned to the label. All packets with an Ethernet multicast destination address (DA) have an OUI portion of the DA-MAC, which is overridden with the local SPBM node (nodal) nickname (SP source ID) for the PE and SPBM. If in PBBN, the local address bit is set. If the PE is PBB-PE, the standard OUI for the 802.1ah I-SID multicast tree is included and the local address bits are clear.

In data plane operation for interworking with PBBN's EVPN, the unicast frame arriving at the PE has a B-MAC / I-SID tuple and looks to determine the MPLS label stack to apply to the packet. Up is done. Frames with a multicast DA have an I-SID, are looked up, and depend on what multicast capabilities are integrated into the EVPN implementation, with the registered objects in the I-SID. It is either replicated locally to each PE or mapped to the multicast group corresponding to the target I-SID community.

If there is a DF change for B-VID, other PEs need to know when the DF failed for B-VID. This notification should be quick, but only in response to a failure. This news can be inferred by having a means of correlating the RD for the segment ID advertised in BGP with the system ID advertised in IS-IS. When a node that is a backup DF for VID confirms that the segment ID and RD associated with B-VID have disappeared from IS-IS and performs loop avoidance handshaking with its own peer (common ISIS database digest). ), It is safe for the node to know that the state has left the PBBN and assume the role of the DF for the VID. Other defenders are guaranteed to be offline to avoid duality or looping. NLRI originates from the facing BGP sub-TLV and heads for the facing PBBN.

The embodiment provides a solution for ensuring that the B-MAC within a given B-VID appears in only one PE within the PBBN. This ensures that there is no transfer ambiguity. The embodiment functions so that the PE can direct traffic to the appropriate peer PE. Algorithmic allocation by B-VID DF as part of normal CP interworking behavior minimizes the amount of potential PE configuration. Another allocation by B-VID DF makes BGP EVPN free from "churn" and routing instability in individual PBBNs. The embodiment is consistent with the existing PBB-EVPN model and does not require complicated interworking. Embodiments can be interworked with complete loop avoidance.

FIG. 1 is a diagram of one embodiment of an exemplary PBB and SPBM-PBBN EVPN network that implements 802.1aq support over EVPN. The network can include any number of customer edge equipment (CE) nodes 101, which are devices that connect a local area network (LAN) or a similar set of customer devices to the SPBM-PBBN103. The CE101 can be any type of network router, switch, bridge or similar device for interconnecting networks.

SPBM-PBBN103 is a set of network devices such as routers or switches that form a provider backbone network that implements the shortest path bridging MAC mode. This network can be controlled by entities such as Internet service providers and similar entities. The SPBN-PBBN may be connected to any number of other SPBN-PBBN105s, CE109s (via BEB111), or similar networks or devices on an IP / MPaS113 network or similar wide area network. These networks can be interfaced through any number of PE115A-D. From now on, the modification of PE115A to D to support 802.1aq on EVPN within SPBM-PBBN will be further described. The network illustrated in Figure 1 has been simplified for clarity. Those skilled in the art can have any number of CE101, 109, PBBN103, 105 and PE115A-D in which any given PBBN with IP / MPLS network 113 through any number of PE115A-D. You will understand that you can connect.

The embodiment incorporates control plane interworking within the PE that maps the ISIS-SPB information element to EVPN NLRI information and vice versa. Associated with this is the procedure for configuring PE transfer behavior so that any number of EVPNs facing SPBM-PBBN can be interconnected on any topology or without multipathing dependencies. is there. This model also allows PBB-PE to communicate seamlessly with multiple SPB-PBBNs. The model can be extended to support future 802.1Qbp standards, allowing seamless interworking between 802.1ah, .1aq and .1Qbp, as well as supporting opposing 802.1ad-based PBNs. Model.

A BGP route target (a BGP identifier used to identify a particular community of interest in a BGP instance) identifies the set of SPB-PBBN and BEB-PE that can be communicated with. BGP acts as a common repository of I-SID attachment points for opposite sets of PE / PBBN, so to speak, it is a set of PBB-PE and SPBM-PBBN interconnected via EVPN. It takes the form of a tuple of B-MAC address / I-SID / Tx-Rx attributes stored in the PE's local BGP database. CP's interworking feature filters leaking I-SID information in the BGP database to local PBB-PE implementations or local ISIS-SPB routing instances within each PBBN based on locally registered targets. To do. As used herein, leakage refers to selective filtering of which BGP information is moved to the local IS-IS database. If the PBBN does not have a BEB registration target within the I-SID, information about that I-SID from other PBBNs or PBB-PEs will not be leaked to the local ISIS-SPB routing system.

Each PBBN 103, 105 is managed to have an associated Ethernet segment ID (ESI) associated with it. For each B-VID in SPBM-PBBN103, 105, a single PE115A ~ D is selected as the designated forwarder for the B-VID. PE115A-D may be DFs for more than one B-VID. This may be through a setting or an algorithmic means, and an algorithmic is a preferred embodiment. In some embodiments, the network minimizes disruption in BGP-EVPN (ie, the data load caused by BGP messaging and similar activities that reconfigure the network to utilize different PEs as DFs). It is configured to ensure that changes in the designated forwarder are required only in the case of PE115A-D failures or disconnections from BGPN103, 105 or MPLS network 113.

FIG. 2 is a diagram of one embodiment of processing frames during interworking between PBBN and EVPN. When the SPBM frame reaches the PE and is determined by EVPN to go to a reachable destination, the DA B-MAC and I-SID are looked up in the interworking table and the MPLS label stack to prepend to the frame. Is determined. At this point, we begin to refer to the frame as a packet. This will typically be, but not limited to, VPN and PSN labels. The packet is then forwarded accordingly. Packets arriving from EVPN will have a local B-VID, either by guessing from the MPLS label at the bottom of the label stack or via an I-SID lookup. The MPLS information is removed from the front of the Ethernet frame and the B-VID field in the frame is updated. If the frame has a multicast DA B-MAC, the OUI field is overwritten with the SP source ID and the local bits are updated according to the local value before the frame is forwarded.

[Control plane interworking from EVPN to ISIS-SPB] When the PE receives the BGP NLRI containing the new information, it checks whether the I-SID in the Ethernet tag ID is mapped to the target B-VID that is selected as the DF. If there is no BEB in the SPB-PBBN that advertised the target in the I-SID, the PE is not locally associated with any B-VID and is therefore not targeted. If the I-SID is a local target for SPBM-PBBN and the PE is a DF for that I-SID, then the SPBM service identifier and unicast address sub-TLV are built for advertising to IS-IS. / Updated and added to the local IS-IS database.

The NLRI information from BGP advertised to ISIS-SPB is also used to populate the transfer table locally and is indexed by the B-MAC / I-SID pointing to the label stack imposed on the SPBM frame. Attached. The last label is the label offered at NLRI.

[Data plane interworking from SPBM-PBBN or PBB-PE to EVPN] When the PE receives a frame from SPBM-PBBN with the target B-VID that is its own DF, it should look up the B-MAC / I-SID information and add it to the frame for transfer in EVPN. Determine the label stack. The PE adds label information to the frame and forwards the resulting MPLS packet.

[Data plane interworking from EVPN to SPBM-PBBN] Upon receiving the packet from the EVPN, the PE can infer the B-VID to be overwritten in the SPBM from the I-SID or by other means (such as via the last label in the MPLS stack). If the frame has a local multicast DA, the PE overwrites the nickname in the frame with the local nickname (SP source ID).

[Data plane interworking from EVPN to PBB-PE] PBB-PE does not have the concept of PBBN or B-VID which is substantially opposite, and therefore does not require frame processing. PBB-PE needs to accept SPBM-encoded multicast DAs as if they were 802.1ah-encoded multicast DAs. The only information of interest is that it is a multicast frame and the I-SID is encoded in the lower 24 bits.

FIG. 3 is a diagram of one embodiment of PE that implements 802.1aq over EVPN. PE115 connects to PBBN103 through one interface and to IP / MPLS network 115 through a second interface. PE includes IS-IS module 301, control plane (CP) interworking function 305, BGP module 307, IS-IS database 309 and BGP database 311.

The IS-IS module 301 sends and receives IS-IS protocol data units (PDUs) on the PBBN 103 to maintain topological and similar network information to allow the transfer of data packets on the PBBN 103. The BGP module 307 also sends and receives BGP PDUs and / or NLRIs on the IP / MPLS network interface to maintain topological and similar network information for the IP / MPLS network 115.

The CP interworking function exchanges information between IS-IS module 301 and BGP module 307 to enable proper data transfer and implementation of 802.1aq over EVPN. Control plane interworking for ISIS-SPB to EVPN is used when the PE receives the SPBM service identifier and the unicast address sub-TLV as part of the ISIS-SPB MT capability TLV, which PE owns. Check if is the DF for the B-VID in the sub TLV. If the PE is the DF, there is new or altered information and a MAC advertising route NLRI is generated for each new I-SID in the sub-TLV. RD (Route Distinguisher) is set to that of PE. ESI is set to that of PBBN. The Ethernet tag ID contains the I-SID (including the Tx / Rx attributes). The coding of I-SID information follows Table 1 below.

<tables num="1"><img id="000002" he="57" wi="134" file="JP6234440B2_D0001.tif" img-format="tif" img-content="drawing" /></tables>

The management of IS-IS TLV and BGP NLRI is further discussed here with respect to Figures 4 and 5.

The composition of PE can occur in the commissioning of PBB. The PE is configured with a route target for the service instance, and the service stance is defined as a set of PBBN and PBB-PE to be interconnected by EVPN. PE is also the unique ESI for SPBM-PBBN, the nickname of the node used for construction by the algorithm of the multicast DA address (SP source ID); the B-VID used in PBBN and the multipath to be used. Consists of a set of algorithm IDs and an RD to be used for a particular ESI. It is typically encoded as type 1 according to RFC4364.

The following is an exemplary DF selection process implemented by each PE. In one embodiment, the PE self-appoints the role of DF for B-VID for a given PBBN. This process is implemented so that the PE notes the set of RDs associated with the ESI. For each B-VID in the PBBN, the PE XORs the associated ECT mask (see Section 12 of RFC6329) in the quota subfield of the set of RDs and ranks the set of PEs by the quota subfield. .. If the assignment number subfield for the local PE is the smallest value in the set, then the PE is the DF for that B-VID. It should be noted that PE needs to reassess the role of DF whenever RD is added or disappears from ESI for RT. In other embodiments, any DF selection process implemented via a configured table of possible scenarios or algorithms can be utilized. The DF selection process can ensure that all PEs attached to the ESI have some of the load split between them, and the distributed DF selection process is relevant for a given Ethernet segment. It will produce the same results across a set of PEs that implement the DF selection process.

Figure 4 is a flow chart of one embodiment of the process of interworking an 802.1aq control plane with an EVPN BGP control plane, which technique is used to select a designated forwarder for a given B-VID within the local PBBN. The PE uniquely moves specific I-SID and MAC information about the B-VID from IS-IS to BGP and is associated with that particular PE (from local I-SID to B-VID mapping). Determine which I-SID and MAC information in BGP is leaked from BGP to IS-IS.

The method is initiated when the PE receives an IS-IS PDU containing an IS-IS TLV (block 401). The received IS-IS PDU data is processed in the normal IS-IS protocol procedure (block 403). The received IS-IS PDU data is checked to see if it requires a change in the current configuration of the designated forwarder (DF) (block 405) (eg, there is a change in the RD in the ESI for RT). If a change is required, a determination is made as to whether the PE is still a DF for one or more B-VIDs (block 413). If no change is required, it is determined whether the PE at that time is DF for any B-VID (block 407).

In the case of a change, if the PE at that time is no longer a DF for at least one B-VID, then all local DF associated NLRI (network layer reachability information) is from the BGP database. Deleted and all remote NLRI origin information for the DF from the IS-IS database (NLRI sourced) information) is deleted (block 415). The IS-IS and BGP protocols then synchronize their respective peers with the database. In the case of a change and the PE at that time is not a DF, or after all NLRIs have been removed and the PE is a DF, the PE has become a DF for one or more B-VIDs. Is checked (block 407). If PE is not DF, the process is complete. If the PE becomes a DF, all local DF-related NLRIs will be added from the IS-IS database to the BGP database. Also, all DF-related NLRIs with remote SPBM subtenants from BGP are added to the IS-IS database (block 419).

Once the DF status confirmation is completed, it is checked whether the PE at that time is the DF for any B-VID. If it is not a DF for any B-VID, the process ends. If the PE is a DF for at least one B-VID, a check is made to see if the received SPBM service identifier and unicast address sub-TLV associated with that B-VID has changed (block 409). If it has not changed, the process is complete. If the above data changes, local information from the IS-IS database is added to the NLRI in the BGP database (block 411). Then, the process ends. This process may be resumed in response to the reception of other IS-IS PDUs or TLVs.

FIG. 5 is a flowchart of one embodiment of the process for BGP NLRI processing. In one embodiment, processing begins upon receipt of BGP NLRI (block 501), and BGP NLRI is processed according to the normal BGP protocol (block 503). A check is made to see if new information has been received with NLRI (block 505). If no new information has been received, the process is complete. If new information is received, it is checked whether the PE at that time is a DF for a specific I-SID. If the PE is not a DF for the I-SID (block 507), the remote information is added to the IS-IS database and the process completes (block 509).

Similarly, in a scenario where a PE has become the selected DF for B-VID in a running network, the IS-IS database is processed to build NLRI information associated with the new role of the PE. Will.

If the BGP database has NLRI information about the I-SID and it is the first instance of the subject's registration in the I-SID from SPB-PBBN, the NLRI information with that tag is processed by the PE. An updated set of SPBM service identifiers and unicast address sub-TLVs to be advertised is constructed.

The ISIS-SPB information also keeps the local table at that time indexed by I-SID to indicate the B-VID associated for processing frames received from EVPN. Only one entry in the table is allowed if the I-SID is associated with more than one B-VID.

[Flow order] If multicast by I-SID is implemented through PE replication, a stable network will preserve the frame order between known unicast and BU traffic (eg, no race condition). Will). This cannot be guaranteed if MVPN is used (individual multicast trees may not forward the same path as unicast traffic).

[Transit] PEs that do not need to participate in tandem calculations can use the IS-IS overload bit to eliminate the SPBM tandem path and act as a pure interworking platform.

FIG. 6 shows an example of a network element that can be used to implement one embodiment of the present invention. The network element 610 may be any PE or similar device described above.

As shown in FIG. 6, network element 610 includes a data plane that includes a switching fabric 630, multiple data cards 635, receive (Rx) interface 640, transmit (Tx) interface 650, and I / O port 655. Rx and Tx interfaces 640 and 650 interface with links in the network through I / O port 655. If the network element is an edge node, I / O port 655 also includes multiple user-facing ports to provide communication to and from the outside of the network. The data card 635 performs functions on the data received on interfaces 640 and 650, and the switching fabric 660 switches the data between the data card and the I / O card.

Network element 610 also includes a control plane, which includes one or more network processors 615 that contain control logic configured to handle the routing, forwarding, and processing of data traffic. Network processor 615 runs a split tiebreaker for spanning tree route selection, calculates and installs the forwarding state for spanning tree, calculates the SPF tree in the event of a link failure, and FDB626 for data forwarding. It is also configured to input data to. Other processes may also be implemented within the control logic.

The network element 610 also includes a memory 620 that stores the FDB 626 and the topology database 622. The topology database 622 stores a representation of a network model or similar network topology, including the link state of the network. FDB626 stores the forwarding states of network element 610 in one or more forwarding tables, which indicate where to forward incoming traffic to network element 610.

In one embodiment, network element 610 may be coupled to management system 680. In one embodiment, the management system 680 includes one or more processors 660 that are attached to memory 670. The processor 660 contains the logic for configuring the operation of the system ID and network element 610, which is the non-blocking of the network to update the system ID and thereby distribute the work in the network and to a subset of the spanning tree. Includes priority assignments, such as maintaining characteristics at least for those spanning trees. In one embodiment, the management system 680 may perform a system management function of calculating a transfer table for each node and downloading the transfer table to the node. The system management function is optional (as shown by the dashed line), and in an alternative embodiment, the distributed routing system performs the above calculations so that each node calculates its own forwarding table. You may.

Various embodiments of the present invention may be implemented using various combinations of software, firmware, and / or hardware. Thus, the illustrated technique can be implemented using code and data stored and executed in one or more electronic devices (eg, end stations, network elements). Such electronic devices include non-transitory computer-readable storage media (eg, magnetic disks, optical disks, random access memory, read-only memory, flash memory devices, phase-change memory) and temporary computer-readable transmission. Computer-readable media such as media (eg, electrical, optical, acoustic, or other forms of propagating signals-carriers, infrared signals, digital signals, etc.) are used to store codes and data Communicate (internally and / or with other electronic devices over the network). In addition, such electronic devices typically include one or more storage devices (non-temporary machine-readable storage media), user input / output devices (eg, keyboards, touch screens, and / or displays). , As well as one or more sets of processors linked to one or more other components, such as network connections. The connection between the set of processors and other components is typically done through one or more buses and bridges (also known as bus controllers). Thus, a storage device for a given electronic device typically stores code and / or data for execution on one or more sets of processors for that electronic device.

Flow diagrams in the drawings show a particular sequence of actions performed by certain embodiments of the invention, but it should be understood that such order is exemplary (eg, alternative embodiments). May perform actions in different order, combine certain actions, duplicate certain actions, etc.).

Although the present invention has been described with respect to some embodiments, the present invention is not limited to the above-described embodiments, and it may be practiced with modifications and modifications within the ideas and scope of the appended claims. Those skilled in the art will recognize that they can. Therefore, this description should be taken as an example, not a limitation.