Managing the Performance Impact of Administrative Utilities

Abstract

Abstract. Administrative utilities (e.g., filesystem and database backups, garbage collection in the Java Virtual Machines) are an essential part of the operation of production systems. Since production work can be severely degraded by the execution of such utilities, it is desirable to have policies of the form "There should be no more than an x% degradation of production work due to utility execution." Two challenges arise in providing such policies: (1) providing an effective mechanism for throttling the resource consumption of utilities and (2) continuously translating from policy expressions of "degradation units" into the appropriate settings for the throttling mechanism. We address (1) by using self-imposed sleep, a technique that forces utilities to slow down their processing by a configurable amount. We address (2) by employing an online estimation scheme in combination with a feedback loop. This throttling system is autonomous and adaptive and allows the system to self-manage its utilities to limit their performance impact, with only high-level policy input from the administrator. We demonstrate the effectiveness of these approaches in a prototype system that incorporates these capabilities into IBM's DB2 Universal Database server.

Monitoring scheduling for home gateways

“…However, application management must be performed carefully, so as not to impair the system functions with management-related performance overhead. This question has been studied in the context of Web Services [4], and for EJBs [5], or with a focus on Operating System management [6]. The performance question for OSGi management will be discussed in detail in this paper.…”

Control of large scale computing systems

“…MISO and MIMO systems are practical representations when there are multiple actuators being manipulated in a coordinated way and with an interrelated impact on system performance (e.g., utilities throttling [1]). In addition to the techniques and examples of such systems shown in Table I, general studies found in the control literature include an emphasis on computational aspects such as model reduction, sparse matrices, and parallel algorithms [3] [22].…”

“…In the last several years, there have been many examples of applying control theory to computing systems, including impact on commercial products [1] [2]. A common thread in this work has been to demonstrate the value of a control theoretic approach on a relatively small scale.…”

A control theory foundation for self-managing computing systems

“…The results of this work have been incorporated into an IBM database management product. More details can be found in [12]. Fig.…”

“…Control theory has also been applied to middleware to provide service differentiation and regulation of resource utilizations, as well as optimization of service level objectives. Examples of service differentiation include enforcing relative delays [10], preferential caching of data [11], and limiting the impact of administrative utilities on production work [12]. Examples of regulating resource utilizations include a mixture of queueing and control theory used to regulate Apache HTTP Server [13], regulation of the IBM Lotus Domino Server [14], and multiple-input-multiple-output (MIMO) control of Apache HTTP Server (e.g., simultaneous regulation of CPU and memory resources) [15].…”