The idea is to let carriers create new services without having to reprogram the operations support system (OSS) software that runs behind the scenes. Amartus calls this software-defined service orchestration (SDS), a concept that it’s been discussing since at least April and that has now found its way into a product.
Amartus is describing it as a pragmatic step towards making SDN work for carriers. “SDN has a very simplified network architecture diagram. Service providers look at that and say, ‘I have to deliver services. How do I do that?'” CEO Michael Kearns says.
He sees potential business-model benefits here, too. If carriers could create services more quickly, they could start offering pay-per-use services, rather than the fixed-cost, fixed-term services that are offered today. (“Services” here refers to carrier services: creating things like MPLS connections or optical-transport links, or changing the bandwidths of those connections.)
SDN Orchestration Without Programming
Chameleon SDS is an option for Chameleon OS, Amartus’ network management platform for carriers. Chameleon OS itself communicates with both the network resources (switches and their element management systems, for instance) and the OSS layer, and it gets tailored for particular deployments by including (or not including) modules such as Chameleon SDS.
With Chameleon SDS, a carrier creates or modifies a service by describing it under well-understood industry models (they’re taken from the TeleManagement Forum). Chameleon SDS translates those models accordingly for the OSS and for the network elements.
In other words, instead of having to do any programming on the OSS, a carrier could now just create an abstract model of the service and communicate it to Chameleon SDS. This also means that the OSS and the EMS are divorced from each other; an operator creating a service doesn’t have to care what combination of OSS and network gear is affected.
Kearns points out some similarities with the Metro Ethernet Forum‘s Carrier Ethernet for Cloud (CE4Cloud) project, originally called Dynamic Responsive Ethernet. That initiative is about creating connections to the cloud — but those connections would be elastic, changeable on the fly, which could require automation similar to what Amartus is talking about. In general, automation for that OSS layer is going to be a bigger issue as SDN gains traction with service providers.