Nihita Goel
Nihita Goel, Ph.D.
Scientific Officer(F) and Head, Information Systems Development Group
Tata Instititute of Fundamental Research, Mumbai, India


My research interests include workflow systems, web service orchestration, BPM and runtime monitoring. Some of work done is described below:

Runtime Monitoring of business process workflows: The business process workflows of todays organizations are a web of interconnected web-services provided by their business partners and other stakeholders. Since the process execution involves task executions that are outside their control domain, one of the biggest challenge in realizing such workflows is to ensure that their behavior remains consistent with the requirements on a real-time basis.
The developed online monitoring system named WF_SLA_MON works by observing all external interactions to keep a watch on a set of given properties. On violation of properties, either ``alerts" are set out for the workflow engine or a reactive action is taken. The system runs along with the workflow engine using properties, formally specified using temporal logic (SL) formula or MSC specification, along with pre-defined functions for computing non-functional properties. From the given properties, we derive the observers as automata and integrate them with the underlying workflow engine.

SLA monitoring: The aim here is to explore how from a given SLA of a BPMN system, monitoring properties are arrived at and how our above system enables the service provider to runtime monitor these properties. We also wish to analyse this for a cloud based workflow enviornment.

An executional framework for BPMN: BPMN is widely used in Model Drive Architectures (MDA) for enterprise-scale solutions. We shall realize an executional platform for MDA framework using BPMN. We transform BPMN into an executional framework using Orc where Orc is a web orchestration language that provides uniform access to computational services, including distributed communication and data manipulation. The interesting features of Orc are its capability to specify patterns like multi-merge, discriminator, arbitrary cycles and several multiple instances etc. cleanly. It supports the realization of the map-reduce paradigm for distributed computing and thus, provides a powerful MDA approach for business analysts to express their solutions. The framework enables creation/simulation of mock scenarios and the use of verification/validation/debugging in an integrated way. We use a graph based approach where a Business Process Diagram (BPD) diagram is validated and then converted to a set of Orc computation structures.