Comprehensive testing methodology for the operational National Grid Computing infrastructure - GARUDA

Grid computing is a collaborative e-infrastructure to solve large multidisciplinary problems having interrelated demands of data, compute, scientific instruments and increased importance of collaboration. Grid is a distributed system which enables sharing, selection, and aggregation of heterogeneous geographically distributed autonomous resources dynamically depending on their availability, capability, performance, and users QoS requirements. Validating such a vast system has additional complexities of heterogeneity, multiple administrative policies, and issues of multi-layer software interactions. GARUDA is the Indian National Grid Computing Initiative connecting over 65 partnering institutes across 30 cities of India and aggregating over 70TF-15TB Compute-Storage power. The authors present the software engineering challenges and validation methodologies for GARUDA, to make it a production level grid. Integrated validation of GARUDA through Application testing, Health Monitoring, Middleware-Software testing, Performance benchmarking and Stress testing are described.