Questionnaires For Web Server Stats
The two compelling reasons to extend the leading edge are to improve response time and to provide more meaningful content and track visitors. Fewer network hops and shorter (network) distance between the customer and Web server site means quicker response time. Second, targeted content can serve users better by including important cultural and language differences that are simply lost in uniform content and a single language. Even within a single country, organizations may want to separately target subsets of users - industrial clients and consumers, for example - with different content. At the trailing edge are the servers that need to stay closer to the information sources within your enterprise. Two compelling trailing-edge motivators are fine-grained demand for some narrow target audience(s) and the need for internal, real-time processing of information during a user's session. An example of the first motivator can be seen at Massachusetts Institute of Technology, where some 18,000 computers (desktops, workstations, servers) are potential Web servers on a multitude of specialized topics. MIT takes the trailing edge very deep into the enterprise; its access into the Internet has no firewall because it would be impossible to filter on the legitimate combination of IP addresses outside MIT that need to talk to the 18,000 endpoints within MIT. The combination is just too large to apply traditional firewall techniques. The second motivator shapes FedEx's Web tracking system. When a user submits an air bill number, inquiries to four servers create a unique composite answer describing your package's path and status. To keep response time short, FedEx performs all processing within its intranet. The user gets data from four servers, some of which may be deep inside the FedEx intranet. Leading and Trailing Edges The leading edge is primarily driven by performance. Can you serve your customers better by buying a big pipe into the Internet or by spreading out the load over several circuits? Will there be a tradeoff between a long access line to a single, good place on the Internet against many shorter lines into regional locations? In general, the closer users are to the content, the better the performance, but as the number of sites increases this general rule hits the point of diminishing returns. Furthermore, short, low-speed lines might be inexpensive for the first few sites, but become expensive when compared with configurations favoring larger economies of scale. At the trailing edge, as more servers and sites are made accessible to more users, the value to those users of the Internet-intranet access increases. Theoretically, this value could increase constantly, if the added content were always useful and accessible. Many organizations are just realizing the cost and effort required to make a lot of data truly useful. Search engines and tutorials on the subject matter must be developed, and the costs of coordinating the content climbs as sites are added.
| 
