The global amount of digital information is growing at a staggering pace of 50% p.a. and will exceed 60 Zettabytes in 2020. While storing and processing of such massive data will offer new business opportunities, it will also require new data centre and data centre networking architectures to provide the necessary scalability, resource sharing, and automation. Scalability is achieved by increasing the number of connected servers and storage devices as well as their interface and processing speeds. Warehouse-size data centres can easily host ten thousands of servers with their associated storage. Using multiple geographically dispersed data centres provides redundancy and further scalability. Server and storage virtualization improve the data centre utilization by resource sharing between multiple tenants or applications. Open source or vendor-specific software frameworks allow an automated control of compute and storage resources.
One of the most challenging issues when scaling data centre resources is the attached network infrastructure – inside the data centre, between multiple data centres and between the data centre and the user. A programmatic network control, a flexible allocation of networking functions, and appropriate switching and interface technologies are crucial to facilitate a seamless capacity adaptation and a coordinated orchestration of computing, storage and network resources.
This special symposium provides a forum for service & content providers, system integrators, equipment manufacturers, component suppliers and academia to discuss requirements, challenges and solutions for next-generation data centres. Key results from latest research as well as practical findings from commercial deployments will be presented.
Specifically, the following questions will be addressed:
- What will be the role of data centres in future service & content provider networks?
- What network architectures & connectivity is required (intra-DC, inter-DC, user to DC)?
- Which role will open source software frameworks play? Will there be open northbound interfaces?
- How can network appliances be virtualized and where will they be allocated?
- How will next generation data centres look like?
- How to best trade-off latency, bandwidth, connectivity, cost and power?
- Which role will optics play inside future datacentres (as well as inter-DC and user to DC)?
- Which optical interconnect technologies will be used (on-board, intra-rack, inter-rack)?
- Is there a place for optical switching, and if so is it circuits, bursts, packets?
- Will Silicon photonics bring a radical paradigm change for the data centre?
3 x 2 hour sessions of invited talks. The third session will include a 30 minute panel discussion.
View symposium 2 programme
Data Center Networking
Network Unifying Software Defined Transport and Datacenter Networking to deliver Carrier class Cloud Services
Network Technologies for
Next-generation Data Centers
Control Plane architectures for Photonic
Packet/Circuit Switching-based Large Scale
Photonic Interconnect Technologies for Data
Center and HPC in the EU Project PhoxTroT
- Juan Pedro Fernandez-Palacios (Telefónica)
- Harm J. S. Dorren (TU Eindhoven)
- Jörg-Peter Elbers (ADVA Optical Networking)
- Matthew Finnie, Interoute, United Kingdom
- Vijay Vusirikala, Google Inc, USA
- Andy Reid, British Telecom, United Kingdom
- Dominique Verchere, Alcatel-Lucent Bell Labs, France
- Chris Liou, Infinera, USA
- Salvatore Spadaro, Technical University of Catalonia, Spain
- John Dunne, Intune Networks, Ireland
- Bert Jan Offrein, IBM Zurich Labs, Switzerland
- Peter De Dobbelaere, Luxtera, USA
- Tolga Tekin, Fraunhofer Institute for Reliability and Micro Integration (IZM), German
- Kobi Hasharoni, Compass EOS, Israel