Multi-agent systems are well suited to benefiting from cloud computing facilities. They provide decision support to the user in challenging environments. They are often applied to integrate heterogeneous information. The agents can help us in managing complex data, but scalability is a big challenge because many agents are commonly needed. Each of them may need a significant amount of computing resources. In tasks such as data mining and integration, these shortcomings become more important. JADE is an existing platform that is designed to handle scalability in multi-agent systems. Many preconfigured JADE platforms are created and linked together in different servers to achieve a measure of scalability. Amazon EC2 provides us with virtually unlimited computing resources via an IAAS platform using cloud computing. There are many techniques in Amazon EC2 that can be used for increasing scalability. The main purpose of this project is to provide a flexible platform that makes decisions in order to choose the best possible scalability technique dynamically. This is according to the changing environment in which system is deployed and nature of different applications running within it.