Performance Prediction and Analysis of BOINC Projects: An Empirical Study with EmBOINC

Abstract

Middleware systems for volunteer computing convert a set of computers that is large and diverse (in terms of hardware, software, availability, reliability, and trustworthiness) into a unified computing resource. This involves a number of scheduling policies and parameters, which have a large impact on the throughput and other performance metrics. How can we study and refine these policies? Experimentation in the context of a working project is problematic, and it is difficult to accurately model complex middleware in a conventional simulator. Instead, we use an approach in which the policies being studied are "emulated", using parts of the actual middleware. In this paper we describe EmBOINC, an emulator based on the BOINC middleware system. EmBOINC simulates a population of volunteered clients (including heterogeneity, churn, availability, and reliability) and emulates the BOINC server components. After describing the design of EmBOINC and its validation, we present three case studies in which the impact of different scheduling policies are quantified in terms of throughput, latency, and starvation metrics.

A survey of task scheduling in Desktop Grid

“…Глава книги [3] частично посвящена обзору и описанию систем SimBA, SimBOINC и EmBOINC. Эмуляция Desktop Grid с помощью EmBOINC детально описана в работах [17,18], внимание также уделено средствам имитационного моделирования. Система имитационного моделирования BOINC SimBA описана в [19].…”

Reputation-Based Mechanisms for Evolutionary Master-Worker Computing

“…Of course, if all workers are malicious (an unlikely situation, as shown in [8,12,13]) there is no possibility of reaching eventual correctness.…”

“…The graphical representation of the data obtained captures the tradeoffs between reliability and cost and among all three reputation types, concepts not visible through the analysis. Here we present simulations for a variety of parameter combinations likely to occur in practice (extracted from [12,13]) and similar to our earlier work [10]. We have designed our own simulation setup by implementing our mechanism (the master's and the workers' algorithms, including the three types of reputation discussed above) using C++.…”

An Execution Environment for Robust Parallel Computing on Volunteer PC Grids

“…Multiple predictors of future availability of a host based on previous history are described and evaluated in [6,26]. Prediction of application execution time in a volunteer environment is studied in [14,29]. In contrast to research discussed above, this paper focuses on communicating parallel jobs rather than "bag-of-tasks" jobs, and on actual execution on a testbed rather than simulation.…”