Abstract:
Techniques for reducing the latency in performing a failover from a protected connection to its backup connection when a network event is detected affecting the protected connection. In an MPLS network, techniques are provided for failing a protected LSP to a backup LSP in a fast and efficient manner. In one embodiment, the faster failover is facilitated by storing protected LSPs and their backup LSPs information in the data plane, such as locally on a linecard.
Abstract:
Techniques for reducing the latency in performing a failover from a protected connection to its backup connection when a network event is detected affecting the protected connection. In an MPLS network, techniques are provided for failing a protected LSP to a backup LSP in a fast and efficient manner. In one embodiment, the faster failover is facilitated by storing protected LSPs and their backup LSPs information in the data plane, such as locally on a linecard.
Abstract:
Techniques for dynamically establishing and configuring a liveliness detection session for a local repair connection of a connection. In one set of embodiments, a signaling message is received at a node configured to maintain a local repair connection for a connection. A configuration parameter is identified in the signaling message, where the configuration parameter pertains to a liveliness detection mechanism or protocol. A liveliness detection session is then established for the local repair connection based on the configuration parameter.
Abstract:
Techniques for configuring a local repair connection for a protected connection including determining a path for the local repair connection. The path traversed by a local repair connection starts at a node in the path associated with the protected connection and ends at a merge point node in the path associated with the protected connection that is downstream from the start node. In one embodiment, the merge point node may even be more than two hops downstream from the start node in the path associated with the protected connection. The local repair path may include zero or more nodes that are not included in the path associated with the protected connection. Techniques are also described for optimizing the path associated with a local repair connection.
Abstract:
A method and system for processing active AFP sessions by a multithreaded file system process. The file server is coupled to entities on a network, wherein an open AFP session is created when the file server and an entity on the network establish communication, and an active AFP session is created when an AFP session has a pending request. The method and system comprises monitoring the file server for active AFP sessions. When a first active AFP session having at least one AFP request becomes available, the method and system assigns a first one of the threads to the first active AFP session for processing. The method and system preempts the processing of the first active AFP session by the first one of the threads when additional active AFP sessions are available so that the first one of the threads becomes available to process additional active AFP sessions, whereby equal processing time is provided to each one of the active AFP sessions.
Abstract:
Techniques for computing a path for a local repair connection to be used to protect a connection traversing an original path from an ingress node to an egress node. The computed path originates at a node (start node) in the original path and terminates at another node (end node) in the original path that is downstream from the start node. A Constraint Shortest Path First (CSPF) algorithm may be used to compute the path. The computed path is such that it satisfies one or more constraints and does not traverse a path from a first node in the original path to a second node in the original path, wherein the first and second nodes are upstream from the start node in the original path and the second node is downstream from the first node in the original path. A local repair connection may then be signaled using the computed path.
Abstract:
A system and method for enabling a file server to service a plurality of physical networks, wherein each of the networks has a network protocol of the same type. The system and method comprises a single instance of a filing protocol, and an object-oriented class for implementing a session protocol. The system and method further comprises means for invoking a plurality of session protocol instances, where each one of the plurality of session protocol instances corresponds to one of the plurality of physical networks. Each one of the plurality of session protocol instances services the corresponding physical network and communicates with the single instance of the filing protocol, which enables the single instance of the filing protocol to service the plurality of physical networks.
Abstract:
Techniques for configuring a local repair connection for a protected connection including determining a path for the local repair connection. The path traversed by a local repair connection starts at a node in the path associated with the protected connection and ends at a merge point node in the path associated with the protected connection that is downstream from the start node. In one embodiment, the merge point node may even be more than two hops downstream from the start node in the path associated with the protected connection. The local repair path may include zero or more nodes that are not included in the path associated with the protected connection. Techniques are also described for optimizing the path associated with a local repair connection.
Abstract:
Techniques for computing a path for a local repair connection to be used to protect a connection traversing an original path from an ingress node to an egress node. The computed path originates at a node (start node) in the original path and terminates at another node (end node) in the original path that is downstream from the start node. A Constraint Shortest Path First (CSPF) algorithm may be used to compute the path. The computed path is such that it satisfies one or more constraints and does not traverse a path from a first node in the original path to a second node in the original path, wherein the first and second nodes are upstream from the start node in the original path and the second node is downstream from the first node in the original path. A local repair connection may then be signaled using the computed path.
Abstract:
Techniques for finding an optimized local repair path that may be used to signal a local repair connection for a protected connection. The optimized local repair path starts at a node in the path associated with the protected connection and ends at a merge point node in the path associated with the protected connection that is downstream from the start node. Various techniques may be used for finding an optimized local repair path.