CARPO: Correlation-aware power optimization in data center networks

Abstract

Power optimization has become a key challenge in the design of large-scale enterprise data centers. Existing research efforts focus mainly on computer servers to lower their energy consumption, while only few studies have tried to address the energy consumption of data center networks (DCNs), which can account for 20% of the total energy consumption of a data center. In this paper, we propose CARPO, a correlation-aware power optimization algorithm that dynamically consolidates traffic flows onto a small set of links and switches in a DCN and then shuts down unused network devices for energy savings. In sharp contrast to existing work, CARPO is designed based on a key observation from the analysis of real DCN traces that the bandwidth demands of different flows do not peak at exactly the same time. As a result, if the correlations among flows are considered in consolidation, more energy savings can be achieved. In addition, CARPO integrates traffic consolidation with link rate adaptation for maximized energy savings. We implement CARPO on a hardware testbed composed of 10 virtual switches configured with a production 48-port OpenFlow switch and 8 servers. Our empirical results with Wikipedia traces demonstrate that CARPO can save up to 46% of network energy for a DCN, while having only negligible delay increases. CARPO also outperforms two state-of-the-art baselines by 19.6% and 95% on energy savings, respectively. Our simulation results with 61 flows also show the superior energy efficiency of CARPO over the baselines.

Optimizing rule placement in software-defined networks for energy-aware routing

“…For instance, the authors in [20,1] use mixed integer programming to optimize router power in a wide area network. Furthermore, other works on saving energy for data centers have also been presented [13,15]. In general, these works show that up to 50% of network energy can be saved while maintaining the ability to handle traffic surges and guaranteeing QoS.…”

“…The idea is to use OpenFlow to control traffic flows so that it minimizes the number of used network elements to save energy. Similarly, the authors in [15] have set up a small testbed using OpenFlow switches to evaluate energy savings for their model. OpenFlow switches have also been mentioned in many existing works as an example of the traffic engineering method to implement the EAR idea [27,14].…”

“…However, as we can see, the testbed setups with real OpenFlow switches are quite small. For instance, it is with 45 virtual switches onto two 144-port 5406 chassis switches, running OpenFlow v0.8.9 firmware provided by HP Labs [13] or 10 virtual switches on a 48-port Pronto 3240 OpenFlow-enabled switch [15]. However, we argue that when deploying EAR with OpenFlow in real network topologies, much more real OpenFlow switches should be used and they will handle a large amount of flows.…”

“…As prior works in literature, OpenFlow has been used to evaluate energy savings in data center network [13,15]. However, as the limited rule space is ignored, the evaluation may not be as expected, especially in large scale data center networks.…”

“…As shown in literature, many existing works have used OpenFlow as a traffic engineering approach to deploy EAR in a network. For instance, the authors in [13,14,15] have used real testbed with OpenFlow switches to evaluate their energy savings proposals. In these works, the flow table of each switch is assumed to hold an infinite number of rules.…”