Cloud vs Grid Computing

Cloud computing is on demand provisioning of virtual resources and services around them.

• If virtual service is storage, then API will be provided to access and manipulate data storage.
• If virtual service is instances(execution), then API will be provided to create image and deploy them. 
• If virtual service is database, then API will be provided to do all SQL operations and store data

Grid computing is ability to join multiple computers to solve large computational problems.

• Auto-node discovery
• Dynamic deployment 
• Redeployment of a problem
• Topology and collision resolutions
• Load balancing
• Monitoring 
• Elasticity of Computation by adding nodes 
  

Grid Computing Process: 

1. Take a problem
2. Determine the optimal node topology for it
3. Dynamically deploy the problem onto the chosen topology
4. Provide means to divide the problem into multiple parts for parallel execution
5. Distribute the problem within topology
6. Gather results
7. Return them to user

Cloud bring the user applications to run on nodes, and Grid makes these applications into clusters for execution.
Future will be Cloud application becoming Grid aware directly without Cloud platform intervention and Grids can collaborate with other cloud grids to utilize virtual resource more efficiently.

Formula:

• Cloud = virtual hardware resources + APIs to access them
• Grid =  virtual hardware resources + APIs to join them

Reference: 

• http://gridgain.blogspot.com/2009/05/grids-vs-clouds.html