Semi‐asynchronous approximate parallel DEVS simulation of web search engines

Abstract

Summary Discrete Event System Specification (DEVS) is a formalism for the modeling and analysis of discrete event systems. Parallel DEVS (PDEVS) is an extension of DEVS for supporting Parallel and Discrete Event Simulation, which is a powerful tool for evaluating the performance of large scale systems. In this work, we propose an optimistic approximate and semi‐asynchronous parallel strategy. The level of optimism is efficiently managed throughout the simulation execution, and it is automatically adjusted based on the simulation evolution. Load balance and model partitioning is automatically made by means of an algorithm that takes advantage of the communication pattern of the simulated model. Our proposal is designed for Web search engines, which are complex and highly optimized systems devised to operate on large clusters of processors and dealing with dynamic and unpredictable user query bursts. The results show that our proposal is able to reduce both execution times and memory usage of standard optimistic simulations of Web search engine models, at the expense of small errors.

Approximated Rollbacks

“…Other proposals have studied the tradeoff between relaxing strict causality of the events and its effects on performance in (speculative) PDES [6,17]. Essentially, these solutions skip running some state rollbacks if the effects of processing events out of strict timestamp order is considered acceptable-in terms of the final statistics computed by running the simulation.…”

Why High-Performance Modelling and Simulation for Big Data Applications Matters

“…The combination of efficient and reliable simulation software and purpose-built hardware optimized for simulation workloads is crucial to fully exploit the value of simulation and big data. Synchronous and asynchronous distributed simulation have been one of the options that could improve the scalability of a simulator both in term of application size and execution speed, enabling large scale systems to be simulated in real time [43,44]. ScalaTion [45] provides comprehensive support for discrete-event simulation and big data analytics.…”

Infrastructure and Energy Conservation in Big Data Computing: A Survey

“…The simulation power can be increased by deploying both HPC infrastructures and computing models that enable fast job execution and deliver the highest possible computing performance for the simulation workloads. Synchronous and asynchronous distributed simulation is one of the options that may improve the scalability of a simulator, both in terms of application size and execution speed, enabling large-scale systems to be simulated in real time [36], [37]. Investigations conducted in the field the development of modern simulation technologies have led to the design of general-purpose and problem-oriented software tools for Big Data systems simulations.…”