Deadline-aware Power Management in Data Centers

TL;DR

This paper addresses the challenge of optimizing power consumption in data centers by formulating a novel offline problem considering job deadlines and server activation costs, and proposes an online heuristic algorithm for real-time management.

Contribution

It introduces a new generalized assignment problem incorporating deadline constraints and server activation energy, along with a heuristic online solution for dynamic power management.

Findings

The offline problem is a new version of the generalized assignment problem.

The proposed online heuristic performs favorably compared to randomized routing.

The approach effectively balances power efficiency with deadline adherence.

Abstract

We study the dynamic power optimization problem in data centers. We formulate and solve the following offline problem: in which slot which server has to be assigned to which job; and in which slot which server has to be switched ON or OFF so that the total power is optimal for some time horizon. We show that the offline problem is a new version of generalized assignment problem including new constraints issuing from deadline characteristics of jobs and difference of activation energy of servers. We propose an online algorithm that solves the problem heuristically and compare it to randomized routing.