A Smart-NIC-Based Power-Proxy Solution for Reduced Power Consumption during Instant Messaging

The major challenge to achieving efficient power management for local computers that serve as edge devices of the Internet is to achieve long sleeping times while at the same time maintaining network connectivity. By modifying the network interface card (NIC) with a low-weight proxy module that can handle a certain amount of network protocol semantics and hence still keep the local computer alive on the Internet, a longer host sleeping time can be achieved, while keeping many of the components, including CPU, memory, motherboard, hard drive, etc, remain in a very-low-energy mode. In this paper, we pursue a power-proxy solution to the widely-used Instant Messenger (IM) applications. Yahoo! Messenger® is selected for a case study to investigate its communication protocol YMSG. Furthermore, the decision on inactivity time-out values is a crucial factor for energy savings, but the random nature of idle periods makes the decision another challenge. We formulate the decision problem into a stochastic program and propose an adaptive algorithm to set up appropriate inactivity time-out values. Simulations show that the proposed power-proxy solution combined with the inactivity time-out setup algorithm can efficiently reduce the power consumption of an edge computer.

[1]  Alan D. George,et al.  Smart-NICs: Power Proxying for Reduced Power Consumption in Network Edge Devices , 2008, 2008 IEEE Computer Society Annual Symposium on VLSI.

[2]  Ann Gordon-Ross,et al.  A resource efficient content inspection system for next generation Smart NICs , 2008, 2008 IEEE International Conference on Computer Design.

[3]  Bruce Nordman,et al.  Electricity used by office equipment and network equipment in the US , 2002 .

[4]  P. C. Gunaratne,et al.  Design and evaluation of new power management methods to reduce direct and induced energy use of the internet , 2007 .

[5]  Alan D. George,et al.  The next frontier for communications networks: power management , 2004, Comput. Commun..

[6]  Alan D. George,et al.  Real-time performance analysis of Adaptive Link Rate , 2008, 2008 33rd IEEE Conference on Local Computer Networks (LCN).

[7]  Allen C.-H. Wu,et al.  A predictive system shutdown method for energy saving of event-driven computation , 1997, 1997 Proceedings of IEEE International Conference on Computer Aided Design (ICCAD).

[8]  Alan D. George,et al.  Power-Proxying on the NIC: A Case Study with the Gnutella File-Sharing Protocol , 2006, Proceedings. 2006 31st IEEE Conference on Local Computer Networks.