Robust Workload Estimation in Queueing Network Performance Models

Abstract

Traditional approaches for capacity planning are based on queueing network models. However, modeling with queueing networks requires the knowledge of the service demands of each class of workloads at each device described in the model. In real systems, such service demands can be very difficult to measure.In this paper, we present an optimization-based technique to address the problem. The technique is formulated as a robust linear parameter estimation that can be used with both closed and open queueing network models. We consider the case where aggregate measurements (throughput and utilization) are available. Such measurements are typically much easier to obtain than the service demands. We present experimental results which prove the effectiveness of the constrained and robust linear estimation.

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

“…Variants of these regression methods have been developed to cope with problems such as outliers [34], data multi-collinearity [35], online estimation [36], data aging [37], handling of multiple system configurations [38], and automatic definition of request types [39,40].…”

“…Casale et al [34] presents an optimization-based inference technique that is formulated as a robust linear regression problem that can be used with both closed and open queueing network performance models. It uses aggregate measurements (i.e., system throughput and utilization of the servers), commonly retrieved from log files, in order to estimate service times.…”

Transaction Profile Estimation of Queueing Network Models for IT Systems Using a Search-Based Technique

“…To solve these problems, researchers proposed to focus on measuring the end-to-end parameters, particularly the TRTs and RUs, and to infer the service demands from them [2] [3] [4]. This typically involved running the tests with a number of different load values (time consuming) and simplifying the system to a single node approximation (rarely applicable).…”

“…Casale et al [2] proposed to do a linear regression between the RUs and the service demands. While Kraft et al [3] proposed a similar approach but to use TRTs instead of RUs.…”

“…Additionally, some functionality, such as caching, can be only triggered when multiple users access the system, which leads capacity engineers to infer these service demands from multiuser measurements [2] [3] [4]. Such approaches have problems (we will discuss some of them in Section 2) which lead to various approximations and lengthy procedures.…”