Andrew Lord, British Telecom, UK
; Harald Bock, Coriant, USA
To meet high capacity and low latency requirements driven by IoT and 5G mobile together with the delivery of video content, it is essential to combine optical transport, packet switching, NFV and cloud capabilities into a seamless and efficient end-to-end network. Various emerging industry consortia are seeking to drive aspects of the open interoperability required in such an environment, e.g. TIP, openROADM, MEC and fog computing.
A new generation of optical nodes tailored for this cloud environment is now emerging, extending the paradigm of packet-optical nodes to ‘cloud-optical boxes’ which may include packet and/or optical functions.
How is this going to change the way today’s networks are designed and operated? On the one hand, cloud data centers strongly rely on disaggregation of HW and SW functions, which requires open traffic interfaces and open control APIs, whilst on the other hand, the split between functions needs to be defined and may differ depending on use-cases. This is resulting in different market requirements: e.g. while a large part of the industry uses open compute project switches and DWDM DC-I equipment separately, the TIP initiative defines an integrated switch with DWDM interfaces.
This workshop will bring together experts from optical equipment vendors, cloud providers and network operators to express their opinions on ‘cloud-optical boxes’ and the appropriate function split required.
Why use cloud-optical whiteboxes?
How can you implement a network based on those cloud-optical whiteboxes?
What are characteristics of those white boxes?
What are the functions included into such boxes in a network?
What is the targeted level of disaggregation?
How do you control such networks?
What other key elements of openness are needed to enable fast innovation?