| Grid computing can provide many corporations with a much-needed technology boost and introduce another paradigm shift in IT resources, by introducing a new service without a large expenditure or increase in budgets. Corporate IT management should evaluate grid computing for internal use within their IT operations. Grids are suitable for data intensive/computational intensive applications, such as business intelligence, data mining, and statistical analysis. Grids will give outsourcing providers a much need shot in the arm by providing an infrastructure that can meet even the most demanding of clients. Outsourcing organizations are poised to take advantage of the new utility approach and the on demand operating environment. Large hardware companies who also have outsourcing subsidiaries are also prime candidates for this technology. It is likely that grid computing will evolve into specific industry types or application areas and provide services to clients in those markets. Three models are developing: -
Compute grids provide shared access to high performance computing facilities and resources, where the emphasis is on high-end computing capacity. -
Data grids provide shared access to data/transaction-intensive applications with large databases, file systems, or mass storage. -
Application grids provide shared access to specific application sectors, such as financial, manufacturing, and health care. This analysis can be taken further by proposing the types of grids that corporations can implement. Today, many corporations are moving ahead with their grid projects, as the business case and the need are established, and grids can compliment their existing IT infrastructure. Four types of grids are evolving: -
Enterprise/corporate— A grid is established as a secure, internal, and private grid drawing on existing resources within a single enterprise. -
Partner/client/supplier— A designated set of partners, clients, or suppliers share the grid for the business benefit of all. This has particular cost benefits if large number of participants take part. -
Service/utility— A general all-purpose grid is created that can be made available to all that need it. Clients use the resources and pay for what they use. This is a good example of grid on demand. -
Educational— Universities and colleges collaborate in a shared pool of resources. Universities have established networks and have been using some of their resources for some years. IBM prefers to use the following terminology to describe the above classifications. -
Infra-Grids -
Intra-Grids -
Share resources/networks with external partners -
Designed for specific applications -
Utilizes excess IT capacity -
Extra-Grids -
External collaboration with clients/partners/suppliers -
Resource and optimization utilization -
Multivendor system integration -
Inter-Grids -
Sharing of data storage across the public Internet -
Monitored access and security -
Resource sharing and aggregation -
Multivendor hardware/software/network protocols Figure 11.2 illustrates the relationship of these types of grids Figure 11.2. The relationship between the four types of grids. 
Early adopter companies are already experimenting with elements of on demand computing, sometimes called eUtility computing, selectively outsourcing a portion of their business processes and computing applications, based on the types of grids specified above in the diagram. Several vendors already offer utility-based pricing for server and storage capacity. Web services and grid protocols are gaining credibility, even at a time when technology spending is mostly confined to essentials. The rapid development of grid computing in the next few years will drive spending to new levels. Estimates range from spending in 2004 of $350 million to growth in 2009 to over $5 billion, with corporate grids being the leading type of grid implementation. | 
