Load Balancing of Databases on a Cloud

Cloud computing enables dynamic provisioning of resources for distributed computing applications. This allows for software and associated data to be hosted centrally in the cloud to provide Software-­‐as-­‐a-­‐ Service (SaaS). These services need to be highly available, with no single point of failure and are required to provide performance based on a Service Level Agreement (SLA). Relational databases are the most widely used form of databases systems and providing access to database as service is crucial for many applications hosted in the cloud. However, performance cost of database operations such as query, replication and index generation are difficult to determine. In most cases, these are monitored on an ongoing basis and databases and queries have to be reassigned to different computers within the cloud. In addition, the multi-­‐tenancy requires that more than one database be simultaneously hosted to achieve a better Return on Investment (ROI) from the cloud infrastructure, necessitating optimization for each compute node. In this project, we investigate approaches for modelling database operations and device load balancing techniques to maintain query level SLA in a dynamic cloud computing environment