Content centric networking (CCN) is a new network architecture aiming to solve many fundamental problems of existing IP networks. CCN is unique in that it widely deploys caches on each router to reduce redundant data transmission. Introducing caches into routers, however, causes the networks to consume extra energy. In spite of the great research effort that has been put for improving IP networks' energy efficiency, little research has been done for reducing CCN's energy consumption. We aim to fill this void by making CCN more energy efficient. To our end, the general idea is to make traffic flow on a small portion of edges in the network and shut down underutilized edges. Shutting down redundant edges is a commonly adopted technique for greening IP networks, where the core problem is to decide which edges should be closed. Most existing green mechanisms for IP networks predict the prospective traffic matrix of the network and calculate the edges that should be shut down in an offline manner. In CCN, however, estimating traffic matrix is difficult as the wide spreading in-router caches make the traffic pattern much more complicated to read. Facing this challenge, we design an online traffic algorithm called Multiple Tree based Traffic Engineering (MTTE). Data are delivered on multiple tree-topology networks dynamically created based on the physical network and the current network traffic pattern. The trees are carefully constructed so that minimal edges are contained in the trees and congestion is mitigated. Simulations using real ISP network topologies show that by using MTTE, up to 50 of the edges can be shut down. Moreover, to the best of our knowledge, MTTE is the first online green mechanism for CCN.
展开▼
