Cisco Gets GRIP on Network Availability
The set of new features, being called the Cisco Globally Resilient IP, or GRIP, portfolio, is part of Ciscos push to what it calls "resiliency," or the ability for a network to recover from a failure. CEO and President John Chambers stressed the importance of resiliency in both the network and among businesses generally during a keynote last week at NetWorld+Interop in Las Vegas. As networks converge on IP, more and more traffic and applications, from voice to storage area networks, are relying on IP network availability.
"Industry has more and more riding on the IP networks being available, and at Cisco were working on making those network available for all [types] of application needs," said Charles Goldberg, product line manager for GRIP, in San Jose, Calif.
By incorporating the GRIP features into the IOS, Cisco is offering a higher level of recovery from potential network failures, officials said. For instance, among the most important new features are the two standards-based protocols of Nonstop Forwarding and Stateful Switchover. When combined, they offer the ability for near-zero packet loss as a router switches to a redundant route processor, Goldberg said.
Stateful Switchover maintains state information, such as what the router is connected to and the type of connection, and activates a switchover in a two-processor router where one processor is active and another is on standby. Nonstop Forwarding allows for packets to continue flowing even during the switchover.
Cisco initially is focusing that part of the upgrade at higher-end service provider-oriented routers. It will be available for Cisco 12000, 10000 and 7500 Series Internet Routers in June. Availability for more enterprise-class routers, such as the Cisco 6500 and 7600, will follow within a year, Goldberg said. An interim step, called Route Processor Redundancy Plus, will ship in June for the enterprise-class routers and will allow for fast switchover between route processors but will not include storage of state information, he said.
Also focused on enterprises is the ability for sub-second convergence of multicast with unicast when routing tables must be reconverged after a new router is added or is coming back online. Called Multicast Sub-Second Convergence, the feature also will be available in June.
An additional upgrade aimed at enterprises edge premises is scheduled for the second half of the year to improve the uptime for IP services in the event of network faults. The Stateful NAT (Network Address Translation) and Stateful IPsec (IP Security) features allow in a two-router configuration, where one is acting as a backup, for the routers to maintain state information between the routers so IP services, such as virtual private networks, dont fail during a switchover, Goldberg said. On top of those two features, Cisco is planning to add its Gateway Load Balancing Protocol so that multiple routers on separate access WANs can load balance across the WANs to make use of what today is often standby bandwidth and potentially double the available bandwidth, he said.
On the service provider front, Cisco is building in MPLS (Multiprotocol Label Switching) Fast Reroute-Node Protection to be available in June. The feature enables an MPLS tunnel to be routed around a failed node, or router, within 50 milliseconds. Before, Ciscos IOS simply allowed for rerouting in case of failed links between routers, Goldberg said.
Across the entire network, Cisco is adding in June IP Event Dampening to police skittish links and BGP (Border Gateway Protocol) Convergence Optimization to more quickly converge BGP tables. Following those in the second half of the year will be Incremental SPF (Shortest Path First) Optimization to simplify and speed router recalculations of routes in the network.