Abstract:
Techniques are described for performing inline NAT functions in a forwarding element of a mobile gateway router or other device in which subscriber sessions of a mobile access network are distributed across a plurality of session management cards. The session management cards pre-allocate a public network address and port range for subscribers at the time a network connection is established in response to connection request prior to receiving any data traffic associated with the subscriber. NAT profiles are programmed into hardware forwarding elements of the mobile gateway router for inline NAT when routing subscriber traffic for the mobile access network.
Abstract:
Techniques are described for selecting an alternate path for end-to-end service data traffic that traverses multi-homed routers that provide the service to customer networks. For example, as described herein, a router that is a member of a first multi-homing set connected to a layer two (L2) network with one of a plurality of first access links. The router advertises a status of one of the first access links to a second multi-homing set connected to the first multi-homing set with one or more core links. A core link database stores advertised status information for access links of the first and second multi-homing set. Upon a link failure, a path selector selects a core link to transport service data traffic and directs a switch module to switch to active a status a first access links that connects to a router in the first multi-homing set connected to the selected core link.
Abstract:
In general, techniques are described for decentralizing handling of subscriber sessions within a gateway device of a mobile network. A mobile network gateway comprises a data plane having a plurality of forwarding components to receive session requests from a mobile service provider network in which the mobile network gateway resides. A control plane comprises a plurality of distributed subscriber management service units coupled by a switch fabric to the data plane. Each of the subscriber management service units serve as anchors for communication sessions for mobile devices that are accessing one or more packet data network by the mobile service provider network. A request delegation module within each of the forwarding components directs the session requests to the subscriber management service units unit to provide management services for the sessions requested by the mobile device.
Abstract:
In general, techniques are described for aggregating, within a network device, internal forwarding routes for multiple control protocols and allocating next hops for the routes among individual service units of a decentralized control plane for the network device. The techniques may also include aggregating internal forwarding routes for data protocols and allocating next hops for the routes among individual forwarding units of a decentralized data plane for the network device. In one example, a mobile gateway includes a plurality of subscriber management service units that present a uniform interface to nodes within a mobile service provider network. An allocation manager apportions a control protocol session identifier namespace into a plurality of contiguous, non-overlapping protocol session identifier ranges and allocates the ranges among the service units. The service units execute the control protocol by utilizing respective allocated ranges, which the aggregate internal forwarding routes use to identify the associated service units.
Abstract:
In general, techniques are described for performing scalable layer two (L2) learning in computer networks. A network device that includes interfaces and a control unit may implement these techniques. The control unit stores a L2 learning table having entries that are each associated with a service tag identifying a service virtual local area network. In response to receiving a packet that includes a service tag, the interfaces access the L2 learning table using the service tag to determine whether any of the entries of the L2 learning table are associated with the service tag. When none of the entries are associated with the service tag, the L2 learning module updates the L2 learning table to create a new entry defining an association between the one of the interfaces that received the packet and the service tag.
Abstract:
Techniques are described for providing point-to-multipoint (P2MP) Ethernet service in a L2 network. Routers providing the Ethernet service allow an administrator to classify local attachment circuits as either “leaf” attachment circuits or “root” attachment circuits to define a tree-like architecture for forwarding Ethernet frames within a VPLS domain. Based on the classifications, each of router constructs flood domains, referred to herein as mesh groups, that control switching behavior between attachment circuits and pseudowires that transport the L2 communications through the VPLS domain. The routers utilize the mesh groups when switching L2 communications to enforce the requirements of E-TREE service or other L2 services in which L2 traffic is constrained within the L2 VPN to tree-like connectivity.
Abstract:
In general, techniques are described for encoding port ranges. In one example, a method includes generating an encoded value that represents a specified port range including a first element storing an identifier that identifies a frequently occurring port range stored in an associative data structure of most frequently occurring port ranges, a second element storing an index that represents a dynamically-learned port range specifying at least a part of the specified port range, the dynamically-learned port range represented in a tree-structure of dynamically-learned port ranges and identified by the index, and applying, by a forwarding plane of the computing device, one or more filters associated with the encoded value to a packet that specifies a port included in the specified port range.
Abstract:
Methods, apparatus, and products for routing frames in a network using bridge identifiers, wherein the network includes a plurality of bridge nodes. At least one of the bridge nodes operates as an ingress bridge node through which frames are received into the network. At least one of the bridge nodes operates as an egress bridge node through which frames are transmitted out of the network. One of the bridge nodes receives, from the ingress bridge node, a frame for transmission to a destination node. The destination node connects to the network through the egress bridge node. The frame includes an ingress bridge identifier and an egress bridge identifier. The bridge that received the frame then routes the frame to the egress bridge node through which the destination node connects to the network in dependence upon the ingress bridge identifier and the egress bridge identifier included in the frame.
Abstract:
Methods, apparatus, and products for routing frames in a shortest path computer network for a multi-homed legacy bridge, wherein the network includes a plurality of bridges. At least two of the plurality of bridges operate as edge bridges through which the frames ingress and egress the network. A first edge bridge identifies a legacy bridge nickname for a legacy bridge connected to the network through the first edge bridge and a second edge bridge using active-active link aggregation. The first bridge receives a frame from the legacy bridge and determines, in dependence upon the frame's destination node address, an egress bridge nickname for a third bridge through which a destination node connects to the network. The first bridge then adds the legacy bridge nickname and the egress bridge nickname to the frame and routes the frame to the third bridge in dependence upon the egress bridge nickname.
Abstract:
Methods, apparatus, and products are disclosed for routing frames in a TRILL network using service VLAN identifiers by: receiving a frame from an ingress bridge node for transmission through the TRILL network to a destination node that connects to the TRILL network through an egress node, the received frame including a customer VLAN identifier, a service VLAN identifier uniquely assigned to the ingress bridge node, and a destination node address for the destination node, the received frame not having mac-in-mac encapsulation; adding, in dependence upon the service VLAN identifier and the destination node address, a TRILL header conforming to the TRILL protocol, the TRILL header including an ingress bridge nickname and an egress bridge nickname; and routing, to the egress bridge node through which the destination node connects to the network, the frame in dependence upon the ingress bridge nickname and the egress bridge nickname.