CloudTran reduces the "cost" side of the development equations: the risk of moving to a new architecture; time to market; and the cost of development.

• Reducing Risk - Out-of-the-box architecture

  Large applications achieve scalability by being distributed across many nodes, with related information being distributed across machines. Distributed development using data in the cloud is harder than previous Java development approaches. The architectural issues that arise include

  • how to distribute data across the grid, and making it easy to redistribute to reduce cross-machine hops.
  • dispersing and collecting it - and providing the application programmer with an 'undistributed' view of the data
  • making sure that what is read is consistent and updates are ACID
  • mapping between the types of data - the programmer view, the in-memory view and the storage view
  • how to efficiently create primary keys that are unique across the application
  • how to separate the deployment information from the application, so a reconfiguration doesn't cause a rewrite of the app
  CloudTran provides a proven architecture that is simple to configure and control.

• Time to market

  Whereever possible, CloudTran uses well-known tools and concepts, such as Java, services/SOA, object/relational modelling and messaging. This means that there is much less learning than going to a completely different set of concepts.

  What this means in practice is that CloudTran maps from these concepts into the concepts and configuration required in GigaSpaces XAP, which is the underlying platform.

  There are some new ideas: defining deployments alongside the application definition; collecting entities into groups; event-driven processing. These are necessary to build optimally efficient applications quickly. The effort here is in understanding the concepts and when to use them; all the mapping to GigaSpaces is done automatically.

• Cost of Development

  Automatic creation of boiler plate, programming patterns and configuration all reduces the cost of development. This is true both for initial development and subsequent enhancements.

  If your application has specific architectural requirements, these can be incorporated into the generation templates, further reducing costs.