Lead

CA Labs: Innovation drives business

CA is involved in product R&D for its customers to help them achieve better business results. By Vinita Gupta

CA realizes that innovation is necessary to stay ahead of competition. The company classifies innovation into three areas—process, product and technology. Any improvements in existing processes (requirements, development, testing and release) or inventing processes falls under process innovation. Enhancements to existing products like adding new features, more efficient implementation of existing features or improving the usability of products are all considered product innovation. Innovative ways of using existing technologies or inventing technologies come under technology innovation.

Dr. Sekhar Muddana, VP, CA Labs India, said, “Innovation is the key to drive business and solve people’s problems. We encourage our employees to publicize their ideas by writing white papers and journal papers, presenting them at conferences, and submitting for patents. At CA India Technology Center (ITC), we have devised several ways to encourage innovation. There is an innovation contest where we award reward points for any innovative activity. We also conduct speed rounds where individuals or teams can present their ideas to an innovation panel. We also established an idea management system where people can submit their ideas.”

About CA Labs

CA has established a research department, called CA Labs in 2005 to work on medium and long-term R&D projects. Through CA Labs, CA collaborates with both internal and external organizations to further innovation in EITM. CA Labs engage with academia, professional associations, industry standards bodies, customers and partners to explore novel products and emerging technologies.

CA Labs has research centers in the US, Canada, UK, Israel, Czech Republic, India and Australia. The centers have been working closely with each other and with universities, professional associations and government on various projects that relate to CA products, technologies and methodologies. The results of these projects vary from research publications, to best practices, to new directions for products.

“While we are always looking to the future, this isn’t technology for technology’s sake. Unlike other research organizations, we focus our research on technology that we can potentially apply to CA and its customers’ operations approximately three-to-five years down the road. In this regard, the researchers from CA Labs meet with various product groups within CA to address potential problems in research. They then identify the professors who are best suited to work on these projects, sign collaboration agreements and fund the research,” said Muddana.

Among the more than 25 research projects that are currently underway, here are three that give an insight into the kind of research that is being done. A successful project can result as a feature of the existing product or into a new product.

Taxonomy-Driven Semantic Search

Today, about 80% of all corporate data is unstructured. Keyword searches are inefficient because they are not context sensitive. This is especially challenging for knowledge-based teams such as those in technical support. The Taxonomy-Driven Semantic Search research project seeks to develop a computational framework, tools, techniques and methodologies for accessing the unstructured and semi-structured documents in an enterprise.

In this project, researchers from CA Labs and IIIT-Hyderabad will explore deeper Natural Language Processing (NLP) techniques and taxonomy-based models in extracting keywords and indexing enterprise content.

Muddana added, “We will be building models that will provide a better experience while searching for structured, semi-structured and unstructured documents in an enterprise. Using named entity (NE) extraction we will create taxonomy for a specific domain within CA. Using the keywords and taxonomy we can also classify and categorize the documents.”

By implementing a taxonomy-driven semantic model, efficient context-sensitive searches can be performed, thereby improving the search experience within an enterprise. CA is collaborating with professors and graduate students from IIIT-Hyderabad for this project. The collaboration, initiated in October 2008, is for three years. Every six months the results will be submitted to determine the status of the project.

Automated Business Process Alignment

With today’s tools, it is difficult to model business processes from both an IT perspective and an enterprise viewpoint. Being able to model these relationships, however, would establish compliance and variance in a customer’s IT environment. The goal of this research project is to link the expected behavior of business processes with the actual behavior of the systems.

Corporate service-oriented architecture (SOA) systems can be described as three layers of abstraction. The first layer deals with the source code and the infrastructure of the deployed system. At this layer are source code models, architectural models as well as the environment and platform-related models. These models are usually specified at a per service, per component or per server node basis.

The second abstraction layer deals with workflow models that are related to the runtime, process-centric, behavioral view of the system as a whole. At this layer are behavioral models such as activity diagrams and service level agreements.

The third layer of abstraction relates to business process models and other high-level enterprise-wide organizational models. At this layer, are business processes and organizational constraints; pre/post conditions and invariants; and formal or semi-formal models that describe the compliance and governance requirements of the system at the business process level.

Muddana revealed that given a collection of models at each level of abstraction (and a set of constraints, properties and invariants that these models must conform to), the challenge is to develop a framework that can be used to verify and validate whether instances of models at one level of abstraction violate the properties, constraints and invariants of models at the other levels of abstraction. In this respect, the problem also relates to projecting static and dynamic models and model features from one level of abstraction to models and model features of other levels of abstraction. “Features of static models can be extracted by static analyzers, while features of dynamic models can be extracted by logging and monitoring tools,” added Muddana.

By modeling the relationships among business processes, IT systems and enterprise views, customers can get a better understanding of how business and IT can support SOA. Results from this project will also help customers better establish the variance between planned IT support and actual business requirements. CA Labs is collaborating with researchers from the University of Athens in Greece and the University of Trento in Italy. This project began six months ago.

Internet Flow Visualizer

This project has recently received a research award of $815,000 from the US Department of Homeland Security’s (DHS) Science and Technology (S&T) Directorate to drive innovation in network traffic flow analysis. Together with research partner Dalhousie University of Halifax, Canada, CA Labs will explore new methods for visualizing massive quantities of network data, enabling the extraction of vital information from network traffic patterns, which can accelerate the identification of potential security threats. Under the award program, the US government fosters the creation of new security technology for the protection of the US cyber infrastructure. For CA Labs, the award will fund the delivery of a visualization framework that can easily present vast amounts of network information to facilitate security analysis across large, distributed networks. This project was initiated early this year.

Presently the R&D department is also working on incorporating technologies like SOA and virtualization into their products. The company is also exploring various options in cloud computing.

vinita.gupta@expressindia.com