Home > Articles

  • Print
  • + Share This
This chapter is from the book


When consumers request information, providers often refer to multiple local and remote sources to integrate several pieces of information in their responses. For example, if you use the Internet to make an airline reservation, it is common for multiple systems (some of which are not directly connected to the Internet) to be indirectly involved in the processing of your reservation. The "confirmation code" you receive when the reservation is made comes only after all steps of the transaction have been completed successfully.

Integration on the server side is common for most Web applications. To some extent, this is a medium-term problem. The Web is a young technology and most of its important processing still involves some legacy or intermediate proprietary systems. These systems have proved reliable and have seemed scalable. Certainly, they are still part of the loop because organizations believe in their ability to handle workloads, but the question is whether these systems are ready for Internet-level scale.

Consider an airline that migrates its ticketing to the Web. To do so, server-side processing is required to connect to a remote, proprietary invoice database for each request. In the past, hundreds of phone-based human agents had no trouble using such a system to do processing. But it may be the case that, for example, there are some hard limits to the number of concurrent connections to this database. When there were never more than a few hundred agents, these limits were never exposed. However, putting such a system in the server-side mix may turn out to be the bottleneck in a Web application.

  • + Share This
  • 🔖 Save To Your Account