Real Time, RealisticallyYour definition of real time has deep ramifications for business processes and the technologies that support themReal time, the buzz term du jour, is quickly approaching overuse. But beyond the buzz, the term represents a significant ongoing trend: the acceleration of business processes. When a customer calls your company to change an account, the call-center application should immediately communicate this event via integration technology to sales and marketing applications that can use the information to business advantage. Current conventional wisdom states that IT systems that enable quick business reactions have value in terms of retaining customers, outperforming competitors, and keeping time-sensitive processes (such as manufacturing) on schedule. The hype surrounding real time makes speedy reactions sound good, yet ignores the thorny details of implementation. In particular, how close a business reaction must come to real time varies from one business process to another. Because IT automates the reaction, the speed required of an integration technology varies from project to project. Furthermore, business operations and IT systems have inherent limitations influencing the reality of real time. Therefore, when planning an IT project that will support a rapid business process, you must define real time realistically on a per-project basis. This is more than a mere planning exercise. Your definition of real time affects a project's scope, requirements, performance expectations, user satisfaction, complexity, and cost. Technical Definitions of Real TimeReal time can have multiple meanings. (See Figure 1.) For example, a few naive souls take the term literally, as if the laws of physics would allow instantaneous integration. Integration technologies such as message-oriented middleware (MOM) and database replication can be very fast, but not that fast. In many ways, the alternate term "near time," which takes a few seconds, offers a better description of actual system performance — a variety of integration technologies, including MOM, replication, and even frequent batch runs, can move information within 10 or fewer seconds — and a more realistic understanding of real time. With this definition in mind, the norm for fast-paced information integration today is still the 24-hour cycle. A common example is the operational data store (ODS) that supports operational reporting. In a nightly batch window, integration technology collects data about operations, loads it into an ODS, and then runs reports so that the next morning, executives can study the previous day's corporate performance. A few companies have sped up this cycle by executing batch runs and reports more frequently, so that operational reports are available more than once a day, sometimes hourly. Reporting aside, similar integration cycles support inventory control, fulfillment, logistics, and other processes of moderate time-sensitivity. Daily and hourly integration cycles are more latent than real time, but they're fast enough for the majority of business processes and realistic in terms of technical feasibility.
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
