Toward sustainable networking: Storage area networks with network coding

Abstract

Abstract-This manuscript provides a model to characterize the energy savings of network coded storage (NCS) in storage area networks (SANs). We consider blocking probability of drives as our measure of performance. A mapping technique to analyze SANs as independent M/G/K/K queues is presented, and blocking probabilities for uncoded storage schemes and NCS are derived and compared. We show that coding operates differently than the amalgamation of file chunks and energy savings are shown to scale well with striping number. We illustrate that for enterprise-level SANs energy savings of 20-50% can be realized.

Building a network coded storage testbed for data center energy reduction

“…We give a brief overview of the theoretical background presented in [1]. Consider a DC storing one or more copies of file f composed of T chunks f = {f 1 , .…”

“…Reference [1] showed that blocking probability is a function of the incoming user tra c. Using curl loader we measure P e b as a function of the number of users present in the system n. For each n, 50 trials were run and confidence intervals are calculated across all data points.…”

“…One aspect of maintaining a quality user experience is keeping blocking probability small, or the probability that content is unavailable when requested. This paper builds upon the analytical work presented in [1], which showed that reductions in blocking probability, using algorithms such as NCS, can significantly reduce DC energy consumption. We are implementing a testbed to refine the DC network models, and to get early estimates of potential energy savings.…”

“…P ROJECTIONS indicate that the worldwide DC industry will require a quadrupling of capacity by the year 2020 [1]. DCs are designed with the goal of serving large numbers of content consumers concurrently, while maintaining an acceptable user experience.…”

“…This paper builds upon the analytical work presented in [1], which showed that reductions in blocking probability, using algorithms such as network coded storage (NCS), can significantly reduce DC energy consumption. We demonstrate that NCS can be deployed in a DC management software with a commercial grade feature set, without interference.…”