Service Integration and Analytics Group - overview



    The service integration and analytics group focuses on exciting research problems at the intersection of Cloud Computing, Machine Learning, and Service Composition.    

    Our recent projects include cloud-native service composition using Docker and Netflix OSS that has been released as the IBM Workflow for Bluemix in IBM's new PaaS+services offering and Process Analytics to improve business outcomes of distributed end to end applications by providing operational analytics to the people making decisions on what to do next.  We have performed process analytics on Congestive Heart Failure patient data to understand treatments and outcomes and our work has become part of several IBM products  We work closely with several customers and visiting researchers. Previously, we played a key role in laying the foundations of SOA -Service Oriented Architecture - by providing core contributions for the Web services framework. For this, we received the IBM Extraordinary Technical Accomplishment and one of the papers has more than 1400 citations.

Current projects:

Process Analytics: Provide insight into distributed applications, at the business logic level, so that end users can make better choices and decisions. Using a combination of techniques from data and process mining, tailored to processes and task execution, we provide out of the box, easy to understand insight into the running applications.

Cloud-native service composition:  A JavaScript based language and corresponding runtime in IBM Bluemix for developers to quickly create new apps that compose interactions with people and Web applications.  In order to run such a stateful service efficeintly and reliably in the cloud, we are using Docker containers and Netflix OSS middleware components. You can try this out in the IBM Workflow for Bluemix beta on IBM Bluemix, with a DevOps model for the system itself and also for the workflow developer.



Business Process Insight and Analytics to enable the understanding, tracking, and continuous improvement through online analytics of people-centric semi-structured processes spread across multiple systems from potentially massive amounts of data. This enables an agile BPM lifecycle that blurs the distinction between modeling and execution and provides a nimble and evolutionary approach that is more closely tied to what is actually occurring. A first step in doing so is the ability to have a live representation of the underlying behavior, which we represent using Business Provenance. We are applying this to Case Management using a Case Health paradigm, in addition to applying it to fully ad-hoc pre-existing yet unmodeled processes in the Business Process Insight Platform.

Business Process Mashups: A lightweight approach for Business Process Mashups using REST-ful workflow scripting in a SaaS environment.

Service Oriented Architecture Our group was instrumental in defining the Web services approach to SOA. Our team members have been key contributors to the approach and its standards, being co-editors of several standards including WSDL and WS-BPEL, creating the first public SOAP engine and the first public BPEL implementation, and making several contributions in this space to the IBM WebSphere Application Server.