Grant: Discovery Grants Program - Individual

In today's world, performance and availability of software systems are two critical revenue impacting factors. Poor performance of software or an hour of its downtime can result in significant financial loss to businesses.Increasingly, software applications are being deployed in clouds because cloud computing offers several advantages. For example, it relieves the application service providers from buying and maintaining data centers thereby reducing the operational costs; it allows dynamic scaling of computational resources as required on a pay-per-use basis; and, it promotes easy deployment in multiple geographic locations at minimal cost. Moreover, in mobile domain, cloud computing can facilitate offloading of intensive computations from resource-constrained mobile devices to resource-rich cloud, thereby reducing battery power consumption.Although the aim of cloud computing has been to reduce computational cost, unfortunately the concept has been hit by performance and availability issues. Examples abound: Frequent outages of Microsoft's Azure is one. Rackspace's cloud service degradation is another. The onus is therefore on the software providers to ensure that their softwares meet the performance and availability requirements in presence of such issues. One way to achieve this could be through utilizing multiple clouds. However, managing applications in multi-cloud environment is a costly and complex task and hence calls for efficient strategies.The goal of the proposed research is to develop novel modeling techniques, tools and algorithms that would help in evaluating performance (response time or throughput) and availability of software systems utilizing multiple clouds. Such techniques and tools will assist the software providers to plan in advance and manage the computational resources in a cost-effective manner.The proposal includes (i) development and verification of novel models to evaluate performance and availability of applications in clouds; (ii) development of tools to aid in the modeling process; (iii) optimal deployment and provisioning of applications in multi-cloud environment that would minimize cost, and would meet their performance and availability requirements; (iv) computation offloading of mobile apps to improve their performance and reduce the battery power consumption of the mobile device. The outcome of this research effort, if successful, will help improve the quality of software systems and reduce business cost. Such developments in Canada can help strategically position the software industry here and help advance it globally.