Title :
A traffic driven power control algorithm for parallel Deep Packet Inspection
Author_Institution :
CEC Inf. Technol. Res. Inst., Beijing, China
Abstract :
The tremendous success of the Internet has made it a huge infrastructure that is connected with an enormous number of network devices to deliver information. With an equivalent increase of power consumption, the heat dissipation demands of Internet core routers are reaching the limits of cooling capability. As a balance of power and performance, this paper presents a traffic driven power control algorithm through the use of dynamic frequency scaling and a sleep mode for parallel Deep Packet Inspection (DPI) in Network Intrusion Detection System (NIDS). The simulation results show that it enables the network processing unit with the capability of adapting its frequency for both performance and energy efficiency. Besides, this paper presents the initial result on power aware parallel DPI system, while the follow-up work and its hardware implementation are the future work.
Keywords :
Internet; cooling; power control; security of data; telecommunication control; telecommunication power management; telecommunication traffic; Internet core routers; NIDS; dynamic frequency scaling; heat dissipation; network intrusion detection system; network processing unit; parallel deep packet inspection; power aware parallel DPI system; power consumption; sleep mode; traffic driven power control algorithm; Abstracts; Cooling; Engines; Tin; Dynamic frequency scaling; NIDS; Parallel DPI; Power efficiency;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2013 International Conference on
Conference_Location :
Tianjin
DOI :
10.1109/ICMLC.2013.6890853