Vincent Chau

Low complexity scheduling algorithms minimizing the energy for tasks with agreeable deadlines

Power management aims in reducing the energy consumed by computer systems while maintaining a good level of performance. One of the mechanisms used to save energy is the shut-down mechanism which puts the system into a sleep state when it is idle. No energy is consumed in this state, but a fixed amount of energy is required for a transition from the sleep state to the active state which is equal to L times the energy required for the execution of a unit-time task. In this paper, we focus on the off-line version of this problem where a set of unit-time tasks with release dates and deadlines have to be scheduled in order to minimize the overall consumed energy during the idle periods of the schedule. Here we focus on the case where the tasks have agreeable deadlines. For the single processor case, an O(n^3) algorithm has been proposed in Gururaj et al. (2010) for unit-time tasks and arbitrary L. We improve this result by introducing a new O(n^2) polynomial-time algorithm for tasks with arbitrary processing times and arbitrary L. For the multiprocessor case we also improve the complexity from O(n^3m^2) Gururaj et al. (2010) to O(n^2m) in the case of unit-time tasks and unit L.

Energy Efficient Scheduling of MapReduce Jobs

MapReduce has emerged as a prominent programming model for data-intensive computation. In this work, we study power-aware MapReduce scheduling in the speed scaling setting first introduced by Yao et al. [FOCS 1995]. We focus on the minimization of the total weighted completion time of a set of MapReduce jobs under a given budget of energy. Using a linear programming relaxation of our problem, we derive a polynomial time constant-factor approximation algorithm. We also propose a convex programming formulation that we combine with standard list scheduling policies, and we evaluate their performance using simulations.

Throughput Maximization for Speed-Scaling with Agreeable Deadlines

We are given a set of n jobs and a single processor that can vary its speed dynamically. Each job J j is characterized by its processing requirement (work) p j , its release date r j and its deadline d j . We are also given a budget of energy E and we study the scheduling problem of maximizing the throughput (i.e. the number of jobs which are completed on time). We show that the problem can be solved by dynamic programming when all the jobs are released at the same time in O(n 4 logn logP), where P is the sum of the processing requirements of the jobs. For the more general case of agreeable deadlines, where the jobs can be ordered such that for every i < j, both r i ≤ r j and d i ≤ d j , we propose a dynamic programming algorithm solving the problem optimally in O(n 6 logn logP). In addition, we consider the weighted case where every job j is also associated with a weight w j and we are interested in maximizing the weighted throughput. For this case, we prove that the problem becomes \({\cal NP}\)-hard in the ordinary sense and we propose a pseudo-polynomial time algorithm.

Throughput maximization in multiprocessor speed-scaling

In the classical energy minimization problem, introduced in 24, we are given a set of n jobs each one characterized by its release date, its deadline, its processing volume and we aim to find a feasible schedule of the jobs on a single speed-scalable machine so that the total energy consumption is minimized. Here, we study the throughput maximization version of the problem where we are given a budget of energy E and where every job has also a value. Our goal is to determine a feasible schedule maximizing the (weighted) throughput of the jobs that are executed between their respective release dates and deadlines. We first consider the preemptive non-migratory multiprocessor case in a fully heterogeneous environment in which every job has a machine-dependent release date, deadline and processing volume and every machine obeys to a different speed-to-power function. We present a polynomial time greedy algorithm based on the primal-dual scheme that approximates the optimum solution within a factor depending on the energy functions (the factor is constant for typical energy functions of form P ( z ) = z α ). Then, we focus on the non-preemptive case for which we consider a fixed number of identical parallel machines and two important families of instances: (1) equal processing volume jobs; and (2) agreeable jobs. For both cases we present optimal pseudo-polynomial-time algorithms.

Throughput maximization for speed scaling with agreeable deadlines

We study the following energy-efficient scheduling problem. We are given a set of n jobs which have to be scheduled by a single processor whose speed can be varied dynamically. Each job

Energy Minimization via a Primal-Dual Algorithm for a Convex Program

J_j

Scheduling fully parallel jobs

Jj is characterized by a processing requirement (work)