Current thinking tends to focus on making the access network completely passive, he says, pushing all possible smart functions deep into the heart of the core network inside massive data centres. But Dunne believes that this approach simply won’t scale, economically, performance-wise or energy-wise.
The alternative, being explored in the PIANO+ project, is to keep intelligence in the access layer, by creating a fully programmable optical mesh network based on Layer 2 connectivity. The network contains distributed data centre functionality – i.e. servers, storage, and processing power – to support local services with high quality of service and low latency. Best-effort internet traffic can be mixed with the services this network will deliver, without any compromise on the quality of the local services.
The project will model the traffic created by low-latency, interactive real-time services and show how this architecture can support a new business model for access networks. “We are trying to demonstrate that the access and second mile network could play a key role in generating revenues by being actively involved in delivering those services,” said Dunne.
There are a number of other partners in the project. The Polish Supercomputing Research Group will build a network demonstrator in Poland. The University of Essex, UK, will be researching the control plane issues of merging the network with the distributed IT infrastructure. And last but not least, components firm the Centre for Integrated Photonics (CIP), will be investigating how optical integration of components could reduce the cost of this solution.
To view the article please click on the below link: