SURVIVABILITY IN IP-OVER-WDM NETWORKS (2)

Shared Integrated Protection Scheme Unlike the conventional approach of designing a static logical topology where the traffic demand is known a priori, we have dynamically constructed the logical connections between IP routers

Shared Integrated Protection Scheme

Shared Integrated Protection Scheme If no existing lightpath is available, the bandwidth requirement is checked to see whether a direct lightpath is needed If necessary, RWA process is invoked If it is necessary to set up a direct lighpath, a path selection algorithm is run on the TE database for the IP layer topology

Shared Integrated Protection Scheme Here the proposed approach supports QoS via providing differentiated reliability services, where αf(k) represents the fraction of traffic that needs to be protected

Shared Integrated Protection Scheme In addition, when the traffic volume is light in a lightpath (lower than a certain threshold), the network reconfiguration will take action and move the traffic to other lightpaths. The unloaded lightpath will now be released via a proper signaling process

Dedicated Integrated Protection Scheme Two-variants of integrated protection schemes : IP-over-WDM protection scheme 1 Only allows the primary path to traverse a single-hop route, but the secondary path can use two-hop routing IP-over-WDM protection scheme 2 Both allowed to be multiple hops on the logical topology

Dedicated Integrated Protection Scheme Difference with shared protection scheme No SRLG information is needed, and the protection paths do not share any spare resources. Require less complex signaling and management and can achieve rapid restoration

Fault Detection and Notification The first task for all survivability mechanisms is to detect the fault Fault detection can be done by performance monitoring to detect WDM layer impairments or by the IP layer via link-probing mechanisms

Fault Detection and Notification Upon detecting the failure, alarms will be generated to initiate the recovery process To avoid multiple alarms stemming from the same failure, LMP (link management protocol) provides a set of diagnostic mechanisms to localize the failure location through a Failure indication signal notification message

Signaling Protocol Mechanism The restoration protocol has to support the following operations Secondary path setup Dedicated protection or Shared protection Secondary path reservation Shared protection Error handling The restoration signaling protocol has to have the capacity to report errors during path setup process

Signaling Protocol Mechanism Reversion function The reverting capacity means that after failure has been fixed, the traffic is moved back to the primary path Reversion will be very critical for shared protection schemes Path deletion operation

Survivability in future IP-over-WDM Since all optical packet switches are capable of switching at various granularities，they effectively bridge the gap between IP packet-level switching and optical circuit-based wavelength-level switching Such switches can function as LSR and OXC，making them theoretically perfect for at least IP over packet switching

Survivability in future IP-over-WDM Whereas in IP layer protection schemes，the ongoing packets to the affected node are buffered by electronic RAM，in optical packet networks no such equivalent optical memory is available

Survivability in future IP-over-WDM Optical burst switching is an alternative approach between circuit switching and packet switching The burst will cut through the intermediate node without being buffered Fig15.20 demonstrates an example of how OBS can handle faults by using deflection routing

Survivability in future IP-over-WDM

Survivability in future IP-over-WDM With OBS，one primary path can be protected by n secondary paths，each of which carries only a fraction of the working traffic (bursts) Overall，the major problem with all optical packet switching is that enabling technologies are still in their infancy