Managing dynamic enterprise and urgent workloads on clouds using layered queuing and historical performance models

Abstract

The automatic allocation of enterprise workload to resources can be enhanced by being able to make what-if response time predictions whilst different allocations are being considered. We experimentally investigate an historical and a layered queuing performance model and show how they can provide a good level of support for a dynamic − urgent cloud environment. Using this we define, implement and experimentally investigate the effectiveness of a prediction-based cloud workload and resource management algorithm. Based on these experimental analyses we: i.) comparatively evaluate the layered queuing and historical techniques; ii.) evaluate the effectiveness of the management algorithm in different operating scenarios; and iii.) provide guidance on using prediction-based workload and resource management.

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

“…Bacigalupo et al [64] investigates a prediction-based cloud resource allocation and management algorithm. LQNs are used to predict the performance of an enterprise application deployed on the cloud with strict SLA requirements based on historical data.…”

Line: Evaluating Software Applications in Unreliable Environments

“…Theorem 1 follows from [14,Theorem 3.1] by showing that the sequence {X v (t)} v∈N+ verifies the conditions conditions (10), (11), and (12). As the QN model X(t) is closed, the entries X i,r (t) are bounded above by N , a condition that also holds for every X v (t)/v as well as for x(t).…”

“…In the introduction we mentioned that the lack of support to estimate response time percentiles has been marked as a limitation of current LQN solvers [12]. While simulation is a natural method to estimate responsetime distributions, it can be very time consuming, especially when estimating the distribution tail [16].…”

“…It has been developed in MATLAB, and its source code and binaries can be downloaded from [21]. LINE focuses on solving LQN models, which have become a popular abstraction to model software systems [5]- [8], [12], [22], [23]. LINE is particularly simple to use in combination with the Palladio Bench tool [6], a software-engineering tool based on the PCM paradigm.…”

“…This therefore limits the use of LQN models to tackle reliability-aware cloud resource provisioning, where, for instance, the virtualized environment can affect the application processing rates at very short time scales [11]. In addition, while existing methods and tools are effective in estimating mean performance metrics, the lack of support for the analytic computation of response time percentiles has been pointed out in the literature as a limitation of LQN models for SLO assessment [12]. As cloud applications typically serve different user classes, and support a large number of request types, metrics such as response time percentiles can therefore be needed at the user-type or requesttype level to effectively evaluate SLOs.…”