WRENCH: A Framework for Simulating Workflow Management Systems

Abstract

Scientific workflows are used routinely in numerous scientific domains, and Workflow Management Systems (WMSs) have been developed to orchestrate and optimize workflow executions on distributed platforms. WMSs are complex software systems that interact with complex software infrastructures. Most WMS research and development activities rely on empirical experiments conducted with full-fledged software stacks on actual hardware platforms. Such experiments, however, are limited to hardware and software infrastructures at hand and can be labor-and/or time-intensive. As a result, relying solely on realworld experiments impedes WMS research and development. An alternative is to conduct experiments in simulation. In this work we present WRENCH, a WMS simulation framework, whose objectives are (i) accurate and scalable simulations; and (ii) easy simulation software development. WRENCH achieves its first objective by building on the SimGrid framework. While SimGrid is recognized for the accuracy and scalability of its simulation models, it only provides low-level simulation abstractions and thus large software development efforts are required when implementing simulators of complex systems. WRENCH thus achieves its second objective by providing highlevel and directly re-usable simulation abstractions on top of SimGrid. After describing and giving rationales for WRENCH's software architecture and APIs, we present a case study in which we apply WRENCH to simulate the Pegasus production WMS. We report on ease of implementation, simulation accuracy, and simulation scalability so as to determine to which extent WRENCH achieves its two above objectives. We also draw both qualitative and quantitative comparisons with a previously proposed workflow simulator.

Using simple PID-inspired controllers for online resilient resource management of distributed scientific workflows

“…The analysis of the influence of using multiple concurrent controllers at different levels is challenging because of (1) the large number of controllers, which can be deployed per workflow task and related performance metrics, or per execution node or core; (2) the "chain effect" of distinguished, and possibly conflicted, actions resulted from the actuator output of the controllers; and (3) the dynamic behavior inherent to distributed systems. Future work include the design and implementation of a solution for analyzing the challenges aforementioned using an accurate, scalable simulation framework [47], which will allow us to design a realistic system for usage in production systems. The bottom of the figure shows the step response function of a memory controller attached to a standard cluster (32 cores, 64GB RAM), which has more potential to arise memory overflows.…”

Accurately Simulating Energy Consumption of I/O-Intensive Scientific Workflows

“…To evaluate the accuracy of our model, we extended a simulator [29] of the state-of-the-art Pegasus [5] workflow management system (WMS), which is the WMS we used to perform the experiments described in Section 2. This simulator is built using the WRENCH simulator framework [28], which can be used to build simulators of WMSs that are accurate, can run scalably on a single computer, and can be implemented with minimal software development effort [4]. We extended the simulator by replacing its simulation model for power consumption (the traditional model in Eq.…”

Bridging Concepts and Practice in eScience via Simulation-Driven Engineering

“…and analyze simulation outcomes. For an exhaustive description of the WRENCH architecture, its functionalities, and APIs, we refer the reader to the latest WRENCH research article [2] and the project's online documentation [3].…”

“…Although WRENCH is a young project and its first stable release was made available just over one year ago, the framework has already impacted CI research, development, and education. In the last year, we have delivered four WRENCH's stable releases, which were downloaded over 300 times, either directly from our GitHub repository 1 or via our Docker Hub repository 2 .…”

“…In this paper, we discuss our experience using simulationdriven engineering (see Figure 1) for bridging theoretical and practical aspects of CyberInfrastratructure (CI) development. To this end, we introduce the WRENCH project [2], [3], a CI simulation framework that provides high-level simulation abstractions for building accurate and scalable full-fledged simulators of CI scenarios with minimal software development efforts. We then briefly describe two use cases in which WRENCH was used for implementing simulators of production application execution frameworks.…”