Utility-based placement of dynamic Web applications with fairness goals

Abstract

Abstract-We study the problem of dynamic resource allocation to clustered Web applications. We extend application server middleware with the ability to automatically decide the size of application clusters and their placement on physical machines. Unlike existing solutions, which focus on maximizing resource utilization and may unfairly treat some applications, the approach introduced in this paper considers the satisfaction of each application with a particular resource allocation and attempts to at least equally satisfy all applications. We model satisfaction using utility functions, mapping CPU resource allocation to the performance of an application relative to its objective. The demonstrated online placement technique aims at equalizing the utility value across all applications while also satisfying operational constraints, preventing the over-allocation of memory, and minimizing the number of placement changes. We have implemented our technique in a leading commercial middleware product. Using this real-life testbed and a simulation we demonstrate the benefit of the utility-driven technique as compared to other state-of-the-art techniques.

Enabling Resource Sharing between Transactional and Batch Workloads Using Dynamic Application Placement

“…We formally describe the technique, and then demonstrate that it not only performs well in presence of heterogeneous workloads but it also shows consistent performance in presence only of long-running jobs compared to other wellknown scheduling algorithms. We perform our experiments with a simulator already used and validated against a system prototype in [25,18]. While here we mainly focus on the description and evaluation of the management of long-running jobs, transactional workloads were extensively covered in [18].…”

“…This paper focuses on batch workload, as the benefits of our approach in managing transactional workload have been shown previously [18]. This permits us to study the algorithm's behavior with a reduced number of variables, while also providing an opportunity to compare our techniques to existing approaches.…”

“…In this section we present three experiments performed using a simulator previously used and validated in [25] and [18].…”

“…In [17], the authors evaluate a similar problem to that addressed in our work (restricted to transactional applications), and use a simulated annealing optimization algorithm. Their strategy aims to maximize the overall system utility while we focus on first maximizing the performance of the least performing application in the system, which increases fairness and prevents starvation, as was shown in [18]. In [19], a fuzzy logic controller is implemented to make dynamic resource management decisions.…”

“…The application placement problem is known to be NP-hard and heuristics must be used to solve it. In this paper, we leverage an algorithm proposed in [18].…”