Cloud computing offers IT organizations the ability to create geo-distributed, and highly scalable applications while providing attractive cost-saving advantages. Yet, architecting, configuring, and adapting cloud applications (latency-sensitive web applications and bulk data processing applications) to meet their stringent performance requirements is a challenge given the rich set of configuration options, shared multi-tenant nature of cloud platforms, and dynamics resulting from activities such as planned maintenance.
This project is developing novel methodologies, algorithms, and systems that can enable application architects to (1) judiciously architect applications across multiple cloud data-centers while considering application performance requirements, cost saving objectives, and cloud pricing schemes guided by performance and cost models of cloud components; (2) automatically learn effective application configurations and configuration-to-performance prediction models through statistical machine learning techniques; and (3) create applications that can adapt to ongoing dynamics in cloud environments through transaction reassignment over shorter time-scales, and application migration over longer time-scales.
The impact of this research is multi-fold: (1) Enable IT organizations to significantly reduce costs by optimally moving their operations to the cloud; (2) create benchmarks based on operationally deployed applications and collecting workload traces which will be made available to the research community; (3) make developed algorithms and systems widely available as open source software; (4) inform the design of a nation-wide health-care cloud in Thailand; (4) introduce cloud computing related topics in the undergraduate and graduate curriculum; and (6) train multiple Ph.D., M.S., and undergraduate students, with explicit effort to recruit and train students from under-represented minority groups.
