Using a New Event-Based Simulation Framework for Investigating Resource Provisioning in Clouds

Abstract

Today, Cloud computing proposes an attractive alternative to building large-scale distributed computing environments by which resources are no longer hosted by the scientists' computational facilities, but leased from specialised data centres only when and for how long they are needed. This new class of Cloud resources raises new interesting research questions in the fields of resource management, scheduling, fault tolerance, or quality of service, requiring hundreds to thousands of experiments for finding valid solutions. To enable such research, a scalable simulation framework is typically required for early prototyping, extensive testing and validation of results before the real deployment is performed. The scope of this paper is twofold. In the first part we present GroudSim, a Grid and Cloud simulation toolkit for scientific computing based on a scalable simulation-independent discrete-event engine. GroudSim provides a comprehensive set of features for complex simulation scenarios from simple job executions on leased computing resources to file transfers, calculation of costs and background load on resources. Simulations can be parameterised and are easily extendable by probability distribution packages for failures which normally occur in complex distributed environments. Experimental results demonstrate the improved scalability of GroudSim compared to a related process-based simulation approach. In the second part, we show the use of the GroudSim simulator to analyse the problem of dynamic provisioning of Cloud resources to scientific workflows that do not benefit from sufficient Grid resources as required by their computational demands. We propose and study four strategies for provisioning and releasing Cloud resources that take into account the general leasing model encountered in today's commercial Cloud environments based on resource bulks, fuzzy descriptions and hourly payment intervals. We study the impact of our techniques to the overall execution time, overall cost and cost per unit of saved time with respect to various instance types offered by the Amazon EC2.

The Failure Trace Archive: Enabling the comparison of failure measurements and models of distributed systems

“…Among the simulators that use FTA data [53,54], SimGrid was extended to support volunteer computing platforms [53]. The main challenge encountered by the designers of SimGrid is the scale of the simulated environments; they showed through experiments using FTA traces that their proposed scalability mechanisms perform well in practice.…”

An improved model for live migration in data centre simulators

“…Our work is based on the GroudSim [6] simulation back-end of the ASKALON system [9] that, due to its integration with the DISSECT-CF [5] Cloud infrastructure simulator, provides models for energy consumption of data centre components, as well as for VM migration and networking.…”

“…This section describes the design and implementation of our energy model in the DISSECT-CF infrastructure simulator [5], integrated in the user-oriented GroudSim simulator [6].…”

“…Our new approach extends on our previous generic VM migration model [4] implemented in a Cloud infrastructure simulator called DISSECT-CF [5], integrated as a back-end of the user-side GroudSim simulator [6]. Compared to other simulators, our new model increases the accuracy of the simulation of VM migration and similar major activities involved in the workload consolidation process.…”

Fostering Energy-Awareness in Simulations behind Scientific Workflow Management Systems

“…Although several cloud simulators exist today [6], [7], [8], they cannot support the requirements of current workflow management systems. They are frequently focused on the user side, therefore mostly considering clouds as a black box.…”