Jarir K. Chaar  Jarir K. Chaar photo       

contact information

Director, Analytics Hosting and Runtimes
Thomas J. Watson Research Center, Yorktown Heights, NY USA
  +1dash914dash945dash3154

links

Professional Associations

Professional Associations:  IEEE   |  IEEE Computer Society

profile


During his 21 years at IBM, Dr. Jarir Chaar has built a record of excellence in a number of areas in computing. He was a major contributor to Orthogonal Defect Classification, a methodology for measuring software quality, and received an IBM Technical Innovation Award for his work in May 1994. Next, he led pioneering work on heterogeneous, peer-to-peer, and disconnected workflow management systems. His reference implementation helped establish the Object Management Group workflow management facility and received a U.S. patent in September 1999. He also was a co-author of the first IBM Global Technology Outlook -- The IBM Research yearly perspective on technological trends that have the potential to influence the future of the IT industry. Between 2001 and 2004, Dr. Chaar built a joint technical agenda and implemented a rich portfolio of projects between Research and IBM Global Services (IGS). In 2005, he became an IGS executive with the responsibility to transform its delivery capabilities. He adapted factory-floor assembly models and automation principles to the delivery of application and business services by geographically distributed services teams. His approach, branded as Application Assembly Optimization (AAO) by IBM Global Business Services (GBS), is a technology-based approach to optimizing the end-to-end application development and maintenance processes for GBS. Dr. Chaar led the definition and execution of the business strategy and technical plans of the AAO initiative which was recognized by industry analysts as a unique, differentiating capability for GBS services in the marketplace. He built its integrated delivery platform and led a large scale effort to adopt the approach and deploy the platform to GBS global delivery centers (GDCs). He has been recognized by senior management as “a true asset to this initiative - the father of invention for most of the work and continues to be a key part of its future success”. The initiative is on plan to deliver $300 million in total savings by 2015, with the platform coordinating the activities of 60,000 developers and testers across 12 GDCs located in 20 cities around the world. From a technical perspective, Dr. Chaar has been the overall Chief Technology Officer for AAO, led its development team, and authored 37 patent filings, with 7 patents granted so far. He was also awarded the prestigious IBM Corporate Technical Award in May 2011. His most distinctive contribution includes defining the work packet construct to allow for scheduling techniques that produce optimal cyclic schedules by using a reservation table per work packet to define the flow of a single work order and a collision vector to prohibit illegal schedules. His work is a major step in industrializing software delivery by geographically distributed teams. AAO enables a structured way of distributing work to remote teams. The work packets represent a standard envelope by which every work order is delivered. Work packets include workflow, instruction (normative guidance), metrics collection, and risk management / exception handling mechanisms. Each work packet constitutes a subset of activities that are bound to a larger project plan Work Breakdown Structure managed through a traditional project management tool. Dr. Chaar turned work packets into a standard model for AAO services that specifies a service as a well-formed set of activities together with their sequencing and durations; complexity of service is used to compute the duration of its activities. The model was used in creating a catalog of 100+ AAO services covering the development and test of Custom Applications, Enterprise Applications (SAP, Oracle) and Analytics. For each AAO service in the catalog, its reservation table and collision vector are used to determine its theoretical maximum initiation rate (rate at which AAO services are assigned to teams) and define the set of all its legal initiation sequences. Higher initiation rates are achieved by including cycles that involve multiple initiations of an AAO service at unequal intervals. The technique also handles scheduling multiple AAO services from the catalog. For application maintenance, Dr. Chaar developed techniques, based on ticket data analytics, for continually reducing the costs and managing the risks associated with an application portfolio. He built a highly automated ticket management system to enable the optimal sharing of skilled resources across accounts. Problem tickets are automatically routed to the appropriate resource pool and within the pool to the appropriate resource. Routing is based on ticket severity and associated SLA, and, resource skills, availability and knowledge of the target application. He also implemented algorithms for handling the backlog of tickets across accounts with the objective of eliminating SLA penalties and maximizing the utilization of technical teams.