A Performance Interference Model for Managing Consolidated Workloads in QoS-Aware Clouds

Abstract

Cloud computing offers users the ability to access large pools of computational and storage resources on-demand without the burden of managing and maintaining their own IT assets. Today's cloud providers charge users based upon the amount of resources used or reserved, with only minimal guarantees of the quality-of-service (QoS) experienced by the users applications. As virtualization technologies proliferate among cloud providers, consolidating multiple user applications onto multi-core servers increases revenue and improves resource utilization. However, consolidation introduces performance interference between co-located workloads, which significantly impacts application QoS.A critical requirement for effective consolidation is to be able to predict the impact of application performance in the presence of interference from on-chip resources, e.g., CPU and last-levelcache (LLC)/memory bandwidth sharing, to storage devices and network bandwidth contention. In this work, we propose an interference model which predicts the application QoS metric.The key distinctive feature is the consideration of time-variant inter-dependency among different levels of resource interference. We use applications from a test suite and SPECWeb2005 to illustrate the effectiveness of our model and an average prediction error of less than 8% is achieved. Furthermore, we demonstrate using the proposed interference model to optimize the cloud provider's metric (here the number of successfully executed applications) to realize better workload placement decisions and thereby maintaining the user's application QoS.

Quality-of-service in cloud computing: modeling techniques and their applications

“…Zhu et al [23] builds a VM consolidation algorithm that makes use of an inference model that considers the effect of co-located VMs to predict QoS metrics. In this method, the workload is modeled by means of a Kalman filter, while the resource usage profile is estimated with a Hidden Markov Model.…”

“…The pattern and trend are first analyzed and then synthetic workloads are created to reflect future behaviors of the workload. Zhu and Tung [23] uses a Kalman filter to model the interference caused when deploying applications on virtualized resources. The model accounts for time variations in VM resource usage, and it is used as the basis of a VM consolidation algorithm.…”

CloudScope: Diagnosing and Managing Performance Interference in Multi-tenant Clouds

“…This work can be categorized into two groups: (1) machine learning-based approaches [1], [4], [5], [6], [7], [8], [9], [10] and (2) queuing model-based approaches [11], [12], [13]. The first group uses sophisticated micro-benchmarks and online training to predict the performance interference of different applications.…”

Virtual network embedding with collocation: Benefits and limitations of pre-clustering

“…For example, the meta-data may include information on the total amount of available resources in the corresponding subtree or cluster; alternatively, it may provide additional information, such as topological information (e.g., mincut) or the type of workloads running in the corresponding cluster. The latter information may be useful, e.g., to learn about potentially interfering workloads [28] (e.g., a disk-intensive VNet may not be embedded together with another disk-intensive VNet, or two services which should not be executed on the same machine); however, we will consider scenarios where resources can be isolated well from each other, and will concentrate on the available resources.…”