Rohit M. Lotlikar

Policy Schedule Advisor for Performance Management

Policies are being increasingly used as a means of implementing autonomic computing features in IT systems. Since these policies consume system resources, the performance of a policy-based system may be hampered if resource intensive policies are scheduled at the same time when other policies and applications are being executed. However, with applications being accessed by users globally across time zones, and fast changing business needs, it is increasingly difficult to identify and maintain suitable schedules for these policies. In this paper we propose a framework to address the above aspect of the policy-based autonomic computing - the problem of determining when to schedule a given policy such that its impact on system performance is minimized, and then giving appropriate feedback. This feedback is aimed at assisting the policy maker in defining or redefining the policy schedule so that it can be executed more efficiently. In this framework, we make use of the underlying log data (i.e. resource utilization data) of a managed resource in order to determine an appropriate policy schedule. We demonstrate the efficacy of our approach using DB2 as a managed resource and policies for data management

RMI: A Framework for Modeling and Evaluating the Resiliency Maturity of IT Service Organizations

With increasing globalization and outsourcing, IT service providers need a way to assess the resilience of their operations (and of their vendors) to understand the impact that various failures can have on their output, and to guide their decisions on where to best invest for remediation. In this paper, we present the resiliency maturity index (RMI), a framework for characterizing and evaluating the resiliency of an IT services organization. The framework consists of (a) a model for capturing the hierarchical component structure of the organization and the relationships among (sub)components, and (b) a method to determine a quantitative score that indicates the overall resiliency maturity of the organization. We describe a case study of the application of RMI to a hypothetical IT service provider organization, based on our experience of applying it to actual real-world delivery centers. We also demonstrate the utility of the RMI framework through various realistic scenarios such as making investment decisions, .assessing the impact of organizational expansion, and outsourcing decisions.

Adaptive policies in information lifecycle management

Adaptive policies contain parameters and take into consideration performance feedback to modify values of these parameters adaptively. We propose a method (applicable to the domain of Information Lifecycle Management) to automatically adapt these parameters. Design issues such as selection of sensors, desired ranges for sensors and effects of parameter sensitivity such as over-correction and under-correction are discussed.

Analyzing policy dependencies using historical information

Autonomic computing is central to the success of IT infrastructure deployment as its complexity and pervasiveness grows. This paper addresses one aspect of policy-based autonomic computing - the issue of identifying dependencies between policies, knowledge of which is useful to the policymaker while defining or updating policies. These dependencies are determined via assesment of the impact of a policy on the sensors (measurable entities at runtime). Our approach uses a simple pragmatic model over the measured runtime information from the recent past. Both static and runtime information is combined to provide effective feedback.

Organizing for Delivery of IT Services Having Moderate Task Novelty

The grouping of IT Services by degrees of novelty and the appropriate organizational structures are explained. For IT Services having moderate levels of novelty, the cellular model of delivery may be appropriate, and the considerations involved in designing the appropriate organizational structure for cellular delivery is explained. A simulation framework is proposed for determining the appropriate organizational structure. This simulation framework includes a model for determining the optimal way to allocate tasks to cells, and also a procedure for determining a optimal design of cells.