NMRG/EMANICS Workshop Position Statement Claudio Bartolini HP Laboratories I'm Claudio Bartolini (http://www.hpl.hp.com/personal/Claudio_Bartolini/index.html), a senior researcher with HP Labs Palo Alto. (However, given the European spin of the meeting, I'd like to let known that my team is based in Bristol, UK). My background is on architecture and design of software systems and frameworks. My current research is centered on business-driven IT management (BDIM), and this is the angle that I would like to bring into the discussion at the meeting. The goal of BDIM is to to enable an enterprise to manage its IT services in accordance with its business objectives. BDIM focuses on the impact of IT on business processes and business results and vice versa; besides the conventional IT metrics such as availability and response time, it looks at key performance indicators (KPIs), that is metrics that have significance from the point of view of the business supported by the IT. The BDIM approach aims at rethinking IT management from a business perspective, whether this be in an operational, tactical or strategic context. Below is a description of my medium term research plan, which I have begun to execute on together with my colleagues at HP labs and external collaborators at various universities in Europe and South America. Within an enterprise, the Information Technology (IT) function must deliver value by aligning itself with the objectives of the business that it supports. IT service management frameworks such as COBIT [1] and ITIL [2] define best practices and processes that support the IT function in the smooth running of their operations. However, neither these frameworks, nor any of the commercially available software tools today can help an IT manager plan and execute their courses of action by taking into account the risk of effecting the actions and their impact on the business. The objective of my research plan is to provide an approach to business-driven IT service management based on modeling the impact on the business of actions taken on IT systems and processes. Building on the modeling activity, I aim to apply existing and/or novel decision support and automation techniques in order to suggest and help execute the actions that minimize the impact on the business. Research plan * Review the state of the art of the academic and industrial research in technologies for business-IT alignment * Produce of a conceptual framework for business-driven IT management * Design and implement a software framework for business-driven IT management complete of modeling tools, decision support tools, and automation tools * Apply the tools and technologies for decision support and automation to at least one real-life scenario of IT service management Research questions The main research questions that I'm expecting to address in the course of my research will have to do with modeling, decision support and automation. Modeling Both the production of the conceptual framework and the design of the software artifacts will require formal models of the description of IT service management processes, and other related IT objects such as people technology and processes involved. I'm expecting that in the course of the modeling research activity I will tackle research questions similar to the following: * What are the key IT metrics that are of concern to IT managers? * What are the key performance indicators (KPIs) that measure the performance of IT service delivery and operations with respect to the value they contribute to the business supported by the IT? * To what degree the concerns of the business are addressed by measuring the level of alignment of observed metric to short term (e.g. service level agreements or SLAs) and mid/long-term (e.g. balanced scorecard metrics) objectives? * How to model the dependencies between IT metrics and business objectives, business rules and policies, using a variety of modeling techniques (logic-based, probability-based)? * How to link these metrics and KPIs to models of organization, objectives, IT facilities, people skills, application systems, technology, and data? * How to estimate risk and impact on the business of activities related to IT management processes? * What information models exist today that can provide a basis to start answering the questions listed above? (Possible starting points are COBIT and ITIL stated above, but also DMTF's CIM [3] - common information model, and possibly advanced modeling concepts from the semantic web, e.g. W3C's OWL [4], the web ontology language.) Decision support Providing decision support in order to maximize the alignment with business objectives requires determining the tradeoffs that exist between conflicting business objectives and analyze them to choose between alternative options available to an IT manager (for instance alternative service and process configurations) such that the optimum alignment with the objectives is achieved with respect to current business and system/resource conditions. The main research question to be addressed here is * What fundamental techniques for decision support (such as forecasting and prediction methods, mathematical optimization, simulation, etc.) can be applied to the most important decision to be made within the IT service management processes? Automation To close the IT management loop and act on the decisions made by the IT managers, it will be necessary to find answers to questions such as: * What fundamental techniques for automation (such as planning, scheduling, automated capacity planning and resource allocation), can be applied to increase efficiency in the execution of the IT service management processes? References [1] COBIT - www.isaca.org/cobit [2] ITIL - www.itil.co.uk [3] DMTF CIM - www.dmtf.org/standards/cim/ [4] W3C OWL - www.w3.org/TR/owl-features/