IBM’s grid computing project:

 

 

Grid computing:

 

 

Reliable, scalable, robust

 

 

IBM Grid and Grow – starter set of services , growing portfolio of targeted solutions for business needs

 

 

Customers: insurance industry, requires sophisticaed modeling techniques, grid and grow for actuarial solutions that give companies the dynamic infrastructure they need. Can help companies lead to shorter product cycles leading to innovation etc.

 

 

Companies who want to seamlessly integrate their data. Discover and consolidate data across the enterprise. Single, unified data integration infrastructure

 

 

The Healthcare Industry- quick, secure access to images. IBM Grid Medical Archive Solution. Open source operating software.

 

 

Financial Industry: complex analytics and computations near real time, structured fixed income products, risk maangement, heding and portfolio balancing. Optimized analytic infrastructure. IBM Parallel File System. Optimized Analytic Infrastructure.

 

 

GRID COMPUTING: Issues to Address

1.      Grid Computing

2.      Market

3.      Which industry are we considering

4.      Competitors

5.      Customers

6.      Compelling needs:

 

 

 

 

 

 

Note:

Grid computing can only work if you can break your data down.

 

 

Our model:

Adding a business model to the existing system.

 

 

 

 

Question:

Is the processing power that we are promising them reliable with multiple users/servers as compared to the processing power from a single, large server system?

 

 

Question: How do we get enough users to employ our system so that we are viable?

Are there incremental steps we can take so that we are useful from the start?

 

 

 

 

Oracle: dealing with bigger customers

 

 

Goal: make system composed of 90-100% end users.

 

 

Charities: We’re going to make the biggest grid in the world.

            Computer stock market

 

 

Motivation: Super computers are expensive $14,000 (1 terminal/ 1 rack)

 

 

 

 

1^st Phase:

 

 

Initial investment:

(10-20 k per server) 

 

 

Volunteer computing- strike a deal with nonusers and encourage users to donate to them.

Biotech, ecommerce, webservices, finance companies,  start-ups (almost all tech start ups will need this CPU power, especially research-based startups)

Selling enterprise software

Flat fee subscription based à sales between companies (commissions and flat fee)

 

 

 

 

Major Technical Constraints:

-

 

 

Low quality computing power: its scattered all over the palce

Vs.

High quality computing power

 

 

à We can segment users

 

 

 

 

-->Final Phase