Second place in the table goes to a 32-processor Digital AlphaServer (configured as four 8-way servers in a cluster). Considering that each of the processors in this system cycles at 350 MHz, we can conclude that this is probably the most inefficient design on the list. Fast processors that sit idle most of the time waiting for the memory or I/O are not of much value except to advertise high MHz ratings (this configuration clearly demonstrates why MHz alone is a pretty worthless rating).

The third-place system, which just edges out the AS/400, is a 28-way Silicon Graphics Origin2000 Server. As I mentioned in an earlier section, this system is notable because it is the basis for the ASCI Blue Mountain Project. Still, it takes 28 processors running at 195 MHz to just edge out an AS/400 with 12 processors running at only 125 MHz.

To achieve high performance on a commercial benchmark such as TPC-C requires a balanced system. We must consider the processor, memory, and I/O subsystem performance as a whole, with each component contributing to the overall performance of the system. The AS/400 gets its performance from incredibly efficient memory and I/O subsystems.

The major contributor to the AS/400’s increased system performance is the new memory subsystem. With this new memory structure, the AS/400 is now beginning to exploit its full 64-bit implementation. Variations of this memory subsystem will let the AS/400 incorporate even more processors and utilize much faster processors. The AS/400 approach continues to focus on a balanced implementation that delivers usable performance for business computing.

Conclusions

The only certainty in computers is change. Today’s technologies are becoming discardable at an astonishing rate, which is having an impact on today’s businesses. The future hardware and software technologies I have described in this chapter will only accelerate this trend. All these technologies will come and go; some will have value for customers, giving them a competitive advantage in their business. Once vendors recognize this value, they all will incorporate these technologies into their computer systems. Over time, even newer technologies will displace them.

Just as technologies come and go, so do computer systems and computer companies. History illustrates very clearly that this is especially true if a given computer system or company is tied to a technology that is either no longer of value to businesses or is commonplace in the market.

Imagine for a moment that it is the early 1980s. You have a business and are going to purchase a midrange commercial computer to run your business. Further, pretend that Rochester never got into the computer business back in 1969. Maybe IBM caught on to Rochester’s plan to build a new line of incompatible computers back in the 1960s and said No. Whatever the reason, assume for the purposes of our little saga that you have no Rochester systems to consider for your selection of a business computer. The hot technology of the day is office systems. Most vendors have separate data-processing and office systems. One of the first vendors to integrate these two systems is Wang Computers. In the early 1980s, Wang is the leader in office systems. This technology is obviously important to your business, so you purchase one of these office systems from Wang. Others in the computer industry also quickly recognize the value of office systems technologies, and soon this technology begins to appear in other midrange systems.

Now assume it is the late 1980s, and your business has decided to upgrade your technology with the purchase of a new midrange computer system. By this time, the minicomputer vendors, who started in the technical computing arena, have strong offerings in the commercial world. Led by systems such as Digital’s VAX, they now include most of the technologies businesses want. Digital, for example, has a software offering it calls All-in-One, which provides all the office functions most businesses want. You decide to replace the Wang system with a Digital VAX.

Pretend now it is the early 1990s. Businesses have become concerned that they are being tied into proprietary computer systems such as Digital’s VAX. They are seeing the disruptions caused in their businesses when they change systems every couple of years. Openness and standards are the hot topics. The most open system is considered to be a Unix system. Vendors such as HP who had moved away from their proprietary systems are touting the advantages and openness of Unix. Many businesses believe they can easily move between different systems and different vendors if they just adopt Unix. The appeal is strong, so you decide to go through the last system conversion you will ever have to do. You decide to replace your Digital VAX with an HP 9000.

It is now the mid-1990s. Almost every computer vendor has adopted open standards. The Unix systems have not delivered on their promise to move easily between systems, and they are more expensive to own than businesses originally thought. The hot topic is client/server computing. Some businesses have tried to replace their midrange systems with PC LANs, but that usually only works for very small businesses. Large, multiprocessor PC servers seem to be the answer. Also, the industry has discovered that proprietary operating systems, such as Microsoft’s Windows NT, can be every bit as open as a Unix system. As a result, Unix popularity in the commercial market is fading rapidly. You want the latest technology in your business, so one more time you bite the bullet and replace your HP 9000 with a Compaq ProLiant server running Windows NT.

Time marches on, and it is now the year 2001. The PC world has been far more expensive than anyone expected. Applications from various vendors don’t seem to work together any better than they did in the past. Even the anticipated cost reductions have not materialized because it is necessary to upgrade your hardware and software every six months, just to stay current. At last, the development of new technologies has allowed the creation of an entirely new concept in computer systems that finally solves all your business problems. This, for sure, will be the last time you need to convert to a new system. You throw out your Compaq server and install the new VR2000 computer from Virtual Reality Corporation. As the slogan for this new start-up company says, “If it seems too good to be real, it probably is Virtual.”

The saga ends here because no one knows whether the next hot system will be the VR2000 or something else. The only thing you can be certain about is that there will be one. You can also be certain that the latest system will not solve every business problem, any more than its predecessors have, and moving to it will again cause a major disruption in your business.