A New Method for Control-plane Congestion Control in Chip-scale Optical Networks
- DOI
- 10.2991/icmemtc-16.2016.254How to use a DOI?
- Keywords
- Many-core; Optical On-chip Network; Control Congestion
- Abstract
Recently, various optical networks have been proposed for chip-scale communication on large-scale many-core chips, including optical burst-switching (OBS). The overall throughput of OBS network is dually constrained by congestion and loss in both the data-plane (i.e., burst contention) and the control-plane (i.e., header packet losses). While previous studies mainly focus on maximizing data-plane throughput, little work has been done on the control-plane congestion problem. In chip-scale OBS, massive fine-grained data bursts, stringent transmission delay budget and moderate network operation frequency have actually constrained the electronic header processing capability, leading to serious network congestion. This paper proposes a new approach based on ingress regulation to overcome the control-plane congestion problem. In our scheme, before being injected, every concurrent packet flow is globally regulated and coordinated so that the aggregated flows do not exceed the header processing capacity of routing nodes, leading to the alleviation of control-plane congestion. We use real application traces to evaluate our method. Simulation results validate that the proposed approach can effectively minimize the control-plane congestion and improve system performance in terms of transmission delay and burst loss rate.
- Copyright
- © 2016, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Shi Xu AU - Renfa Li AU - Quanyou Feng AU - Wenhua Dou PY - 2016/04 DA - 2016/04 TI - A New Method for Control-plane Congestion Control in Chip-scale Optical Networks BT - Proceedings of the 2016 3rd International Conference on Materials Engineering, Manufacturing Technology and Control PB - Atlantis Press SP - 1290 EP - 1302 SN - 2352-5401 UR - https://doi.org/10.2991/icmemtc-16.2016.254 DO - 10.2991/icmemtc-16.2016.254 ID - Xu2016/04 ER -