A distributed evolutionary method to design scheduling policies for volunteer computing

Abstract

Volunteer Computing (VC) is a paradigm that takes advantage of idle cycles from computing resources donated by volunteers and connected through the Internet to compute large-scale, loosely coupled simulations. A big challenge in VC projects is the scheduling of work-units across heterogeneous, volatile, and error-prone computers. The design of efficient scheduling policies for VC projects involves subjective and time-demanding tuning that is driven by knowledge of the project designer. VC projects are in need of a faster and project-independent method to automate the scheduling design. To automatically generate a scheduling policy, we must explore the extremely large space of syntactically valid policies. Given the size of this search space, exhaustive search is not feasible. Thus in this paper we propose to solve the problem using an evolutionary method to automatically generate a set of scheduling policies that are project-independent, minimize errors, and maximize throughput in VC projects. Our method includes a genetic algorithm where the representation of individuals, the fitness function, and the genetic operators are specifically tailored to get effective policies in a short time. The effectiveness of our method is evaluated with SimBA, a Simulator of BOINC Applications. In contrast with manually designed scheduling policies that often perform well only for the specific project they were designed for and require months of tuning, our resulting scheduling policies provide better overall throughput across the different VC projects considered in this work and were generated by our method in a time window of one week.

Computing low latency batches with unreliable workers in volunteer computing environments

“…Other research related to VC scheduling includes using evolutionary algorithms to develop scheduling algorithms [25] and using P2P to perform load balancing in VC systems [26]. A possible future field of research is to use genetic algorithms in predicting worker availability for low latency batch computing, though other methods [13,17] appear to already be very successful and more applicable to low latency computing.…”

Task Scheduling in Desktop Grids: Open Problems

“…Applied to scheduling (e.g., [18,19,41,55]), they simulate the Darwinian evolution in a population of schedules to produce the best ones. Some authors use genetic algorithms to develop optimal scheduling policies.…”

A survey of task scheduling in Desktop Grid

“…Алгоритмы разделены на три группы: «наивные», не учитывающие историю работы; основанные на информации о вычислительных узлах (включая сбор статистики работы); и адаптивные -изменяющие параметры работы «на ле-ту». Несколько алгоритмов каждой группы обсуждаются детально, включая предложенный авторами генетический адаптивный алгоритм, схожий с представленным в работе [26] (более подробное описание этого алгоритма представлено в следующем разделе данной статьи). Также обзор дополнен детальным описанием систем имитационного моделирования BOINC SimBA, SimBOINC и EmBOINC.…”

“…В работе [26] предложен генетический алгоритм для нахождения оптимальной политики планирования заданий. Роль индивидуумов играют наборы правил назначения заданий.…”