The Alcatel Carrier Environment Internet System (ACEIS) software that will be integrated into two of Alcatel's devices for the network core by the end of this year eliminates downtime that occurs when the devices are restarted. This frequently occurs when routing information becomes corrupted or routine software upgrades are made.
By allowing routers in the core of a data network to keep running non-stop through these restarts, service providers can make their networks available 99.999% of the time, according to Alcatel. This "five nines" availability is one of the key requirements carriers place on networks for phone calls and critical data functions such as financial transactions. With the increased availability, more service providers are likely to move voice, video and sensitive data services over to IP networks, according to Keith Allan, assistant vice-president of marketing in Alcatel's broadband networking division.
Many routers, including Alcatel's, already come with twin central processors so that one can take over while the other is updated. However, the processor that takes over may need to rebuild the current routing tables for the network, which can change frequently. That task often occupies the chip for between one minute and 30 minutes, according to Allan. ACEIS essentially allows the backup processor to work in parallel with the main chip so that it can take over without delay.
"You have zero outage because our second processor is fully up to date with everything that's going on in the network," Allan said.
ACEIS will be used with three key IP routing protocols: BGP (Border Gateway Protocol), IS-IS (Intermediate System to Intermediate System) and OSPF (Open Shortest Path First). The software will be integrated into the Alcatel 7770 Routing Core Platform and 7670 Routing Switch Platform by the end of this year. By that time it will be available as a software upgrade at no extra cost to customers with a maintenance contract that includes updated software, Allan said.
Alcatel, based in Paris, is currently focusing ACEIS on the core of the network because high-availability requirements are highest there, but it is examining where else to deploy the software, Allan said.
Other vendors have developed similar technologies and a group within the IETF (Internet Engineering Task Force) is studying a possible standard called Hitless OSPF Restart that might offer a similar capability, said Mark Seery, an analyst at RHK. Seery has seen ACEIS demonstrated and believes it will work, but says scalability is key.
"No one has demonstrated this type of capability for the number of sessions they're talking about," Seery said. "This technology is only useful if you can demonstrate that it scales to the number of typical BGP peering sessions that most people have," he added.
Widespread adoption of any of these technologies would take a long time, but in any segment in which it is deployed - such as on an individual Internet service provider's network - it should improve network performance, Seery said. From that point on, more rollouts should make the Internet more useful.
"If it is widely deployed, it will make the whole Internet more robust and more reliable and will make people feel better about using it," he said.
More reliable IP networks will draw enterprises away from legacy services such as frame relay toward IP-based services such as virtual private networks, which can take advantage of the broad reach of the Internet, Seery said. That shift will make it easier for companies to bring customers, new partners and mobile employees on to their networks.