A Telecommunications Architecture for Cloud Computing Performance

This session will highlight some of the issues in telecommunications for cloud computing and present a framework architecture to discuss where these issues arise and how to minimize their impact. Cloud computing is often associated with servers, storage and virtualization. Although one of the key enablers of cloud computing is wide area communications, it is under represented in discussions of cloud performance. In this paper and session we propose a taxonomy to discuss the various points within the cloud where telecommunications factors can improve or degrade cloud performance. By defining specific communications Reference Points (RP) within the cloud computing model, software designers and network architects can better understand the challenges of the cloud—each reference point presents its own unique requirements and challenges for creating, maintaining and managing connectivity. Creating proper communication paths and configuring them correctly at these reference points or interfaces will insure that cloud performance is maximized. Presentation Outline & Content (draft): During this session, MRV OCS will discuss the four primary communication paths and how to manage a proper traffic flow among these paths: 1) Client Device to Application Server 2) Application Server to Application Server 3) Application Server to Middleware Server 4) Middleware Server to File Server The communication reference points will be defined individually. For each reference point we will: 1) Analyze the type and characteristics of the data that is communicated. 2) Describe the expected traffic patterns at the reference point. 3) Discuss the reliability requirements, the need or lack thereof, for real-time traffic delivery (QoS), and other characteristics of communication, such as security. Performance requirements will differ between these four paths. The data communications network must be aware of the requirements of the data flows based on the needs of each category. This suggests that a new framework is required to discuss: 1) Traffic management in cloud networks 2) Communicating devices on cloud networks 3) Locations for traffic enforcement points within cloud networks If the communication channel adds significant delay variation then application performance may be affected. Likewise if the communication network encounters congestion it could have significant negative impacts on cloud service users. Many cloud applications require significant bandwidth; if bandwidth is inconsistent or constrained, cloud applications may be impaired. In severe situations, time sensitive transactions may be impacted by bandwidth constraints. This session will highlight these issues, propose a reference architecture to describe them and highlight work being done within the IEEE, ITU-T and MEF that address them.

- by Mannix OConnor

Author`s Bio:
Mannix O’Connor has held Director or VP positions at Hitachi, Corrigent Systems, Lantern Communications, Nortel Networks, Bay Networks and Synoptics. Mr. O’Connor was the founding Secretary of the IEEE 802.17 Working Group and was and a founding member of the Resilient Packet Ring Alliance. He was co-chair of the MEF Access committee and his work on Ethernet protocols is extensive starting as a member of the Gigabit Ethernet Alliance. Mr. O’Connor is a coauthor of the recent publication, “Delivering Carrier Ethernet” published by McGraw-Hill. At the Ethernet Academy he published a peer-reviewed paper on management of DOCSIS PONs. He has spoken on switching and optical networking at Supercomm, ComNet, Networld & Interop, Convergence India, Comdex Argentina, and the Congreso Internacional de Telefonia IP Mexico. He holds an MBA from George Washington University and lives in San Jose, CA.