Abstract

Interconnection networks within data centers are facing serious problems due to demanding and time-critical data services, growing importance of IT services and dramatically increasing traffic volume. Three main technical challenges are of key importance to be addressed: 1) providing ultrahigh capacity inside data centers, particularly for large-scale site infrastructure hosting millions of servers, 2) improving cost- and energy-efficiency throughout all parts of the interconnection network with the data center, and 3) offering low latency to support emerging time-critical services, such as connected vehicles, tactile Internet. In this tutorial, we summarize our attempts to address the aforementioned challenges. Our work is targeting both data plane and control plane of optical networks providing interconnects inside the data centers. The data plane forwards the data traffic in the physical infrastructure, while the control plane is in charge to ensure the optimal network operation. For the optimal design of intra-data center networks, a cross-layer approach is instrumental, where the data plane constraints are taken into account by the control plane in order to efficiently schedule the traffic flows while satisfying end-to-end quality of service requirements. Recently, we also have a specific focus on exploring spatial division multiplexing for high-performance optical data center networks.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call