Grid Computing: Center of Excellence
Overview | Grid IPs
Grid computing enables service-oriented, flexible and seamless sharing of heterogeneous resources for compute- and data- intensive tasks. It has captured the high performance computing space by better exploiting existing computing resources while reducing administrative overheads related to managing distributed information systems.
Grid computing enables enterprises to:
- Reduce the load on backend systems
- Reduce maintenance cost
- Better manage operations by eliminating the clutter of vendor products
- Make quick decisions by leveraging data from multiple and disparate sources
The Grid Computing Center of Excellence (COE) at Infosys Labs pursues projects in multiple areas. The COE's vision is to create an adaptive services Grid through service-oriented architectural constructs and platform.
Grid Computing is a conglomeration of different technologies, as shown in Fig. 1. Work on distributed and high performance computing in the past has contributed significantly to the development of distributed resource sharing. Newer technologies such as P2P and distributed shared memory model have paved the way for a resource-sharing platform over heterogeneous resources. The combination of web services and virtualization results in a robust, distributed-computing and resource-sharing platform over heterogeneous resources.
Research directions
The Grid Computing COE pursues research in the following areas:
Performance and scalability
With the exponential growth of transactions, back-end systems face performance and scalability issues. Research is focused on improving performance and scalability of legacy systems.
Management and flexibility
The IT landscape is getting increasingly complex with different products and open- source components. The COE develops solutions to flexibly manage heterogeneous IT assets in the enterprise.
Infrastructure utilization and consolidation
Maintenance costs continue to rise as servers and products get hosted in data centers to cater to diverse business needs and increasing business volume. Inorganic growth adds to the complexity of data centers. Companies must consolidate IT resources and improve infrastructure utilization to reduce cost. Our research focuses on infrastructure management solutions to improve resource utilization and consolidate workload on fewer enterprise servers.
Information virtualization
As data is disparate and heterogeneous, virtualizing information is critical to enable access to real-time information across an enterprise. The focus is on developing innovative Enterprise Information Integration solutions by harnessing the power of data grids to achieve scalable data integration.
Next: Grid IPs