Chapter 4. Real-Time Computing

"There is never enough time, unless you're serving it."

Malcolm Forbes

In a way, real-time computing is something that the computing industry has been moving toward ever since it first began. One of the key value propositions in applying computers to solve business problems is that they can perform both routine and complex calculations faster than we can by ourselves and can therefore save us time. If the task is highly repetitive or complex, we can gain a good return on our investment in the technology. By putting a machine on the job, we can get our work completed faster and go home earlier. In the 1980s, many of us were taught how computers would provide us with a four-day work week and would enable us to enjoy more free time. It seemed quite logical at the time. Twenty years later, perhaps the opposite has been true in hindsight: With ubiquitous information and application access, our data and our work follow us everywhere we go, whether at home or at work. In a way, we have become slaves to the very machines that were supposed to liberate us. As we work with ever-increasing volumes of information in our day-to-day activities, a similar story seems true for the myth of the paperless office. So along with the increase in computing capabilities has come a steady increase in workload, driving us to create machines that are ever more powerful in an attempt to increase productivity and to stay ahead of the curve.

But along with this information overload has come a lot of business benefits as well. As the enterprise becomes virtualized, computers are climbing up the stack in terms of value. They are performing more and more business tasks and are capable of performing complex analytical calculations and routing event-based information and transactions around the world in seconds. The business world has traditionally worked in a batch environment. There were large time lags between the time of data collection, such as store sales at the cash register, and the time of data analysis, such as monthly sales reports. The time between the capture of busines events, their interpretation, and, finally, the reaction to those business events, was considerable. Business reports which historically have been produced weekly or monthly as part of large, time-intensive batch processing jobs within IT departments, can now be produced daily, hourly, or on an ad hoc basis very easily by anyone connected to the network. The computing power exists to generate the analysis and reports in minutes or seconds rather than hours, and the information is accessible at any time. In addition to raw computing power for turning batch analysis and reporting into real-time visibility, the ability to route event-based information and transactions around the world in seconds is also a powerful capability. It has particular merit within the supply chain where real-time supply chain visibility and supply chain event management can provide critical information on the location and status of goods and assets as they move through the global supply chain. This can enable participants in the supply chain to better manage fluctuations in demand, minimizing the bull-whip effect, thereby reducing inventory and costs. Supply chains can also be more proactively managed so that exceptions are detected early and can be quickly resolved before they become more critical issues.

With the advances in computing power and the connectivity of the Internet, business as a whole can now operate on its own schedule and at its own pace and even faster. The former shackles of technical limitations such as limited processing power, proprietary standards, and distribution of paper printouts due to limited connectivity, which bottlenecked information flows in the past, have been released a long time ago. The real-time enterprise is clearly within sight.

The term real-time computing was traditionally applied to real-time operating systems and other areas that required systems that reacted to input immediately and were capable of receiving a steady stream of input data such as embedded computer systems in the automotive, consumer electronics, medical, aerospace, and defense industries. Most operating systems such as UNIX or Windows are not real time because they have a slower response time. Today's definition of real-time computing has expanded to include operating systems, middleware, applications, and even entire business processes. The concept has been scaled up to meet the needs of the business in improving productivity and overall performance. This broader definition of real-time computing is the focus of this chapter. We'll look at how the concept can be applied for business benefit across a number of industry scenarios.