This is a sample analysis/recommendation from the 4Q99 time frame...
WHY LOOK AT THIS?
Non-traditional telecom competitors are using commercial vendor products and computer industry standards as major components in their systems. This may be a
competitive advantage for them. If it is, the traditional telecom vendors should consider doing the same thing.
Telecom products contain technologies that were developed by each
equipment vendor specifically for a particular product line. Some of these technologies
were introduced in the 1970's and 80's time frame and each vendor
has continued to evolve these technologies as components of its systems. It takes a considerable
amount of effort to maintain this infrastructure and to keep it up to date. Non-telecom competitors may not have to expend this effort since they are
using off-the-shelf components (computers, operating systems, etc.) without any custom enhancements. This is an advantage to them. Also, for any
given component in their system, the industry for that component is moving the technology along and
they has a choice of
manufacturers for that component. Each telecom equipment vendor, on the other hand,
is the only manufacturer for its components and is alone in trying to advance its
technology. A limited choice in manufacturers is a disadvantage for them. Also,
they are not likely to keep up with the industry in moving technology
forward.
BUT WHAT ABOUT THESE PROBLEMS?.....
We understand that today's performance (capacity & reliability) of such a system may not be what we are used to, but we assume that the industry will move forward...especially on the industrial specs, military specs, and carrier-grade telephony specs versions of these components.
What if the industry doesn't have exactly what we want? (which is always the case). They should buy today's version and later buy the updated version after it is introduced in 6-18 months. Every time I have seen a look at an emerging technology that could be obtained from a supplier (e.g., IC chips, DSP Speech Algorithms, Computer Systems), by the time it could be developed in-house (6-18 months), the industry would be ready with its next generation....and the industry would have absorbed the development cost. Therefore, I suggest waiting for the suppliers. This is done now with Intel Pentium chips whereas in the past, custom chips were made. The telecom equipment vendors can't do everything (vertical integration) involved in delivering a system anymore. They have to chose a slice (e.g., service software only,...or network integration & testing) and let other vendors do the other slices (e.g., computers, operating systems, databases).
What about vendor management issues? For example, what if computer vendor "A" is not satisfactory. I see two actions that can minimize this issue. First, do not ask the vendors for custom products and alterations. This raises their costs and they simply pass it back. Most vendors now have products that are industrial specs, military specs, or carrier-grade telephony specs. The telecom equipment vendors (and their customers) should just live with today's products. Yes, the suppliers should be told of the problems so that they can incorporate them into the next release (6-18 months out)...but the next release is an industry release and not a custom job. Second, with standards, unreasonable suppliers can be easily replaced. For example, in a Windows environment, one can easily move back and forth among many computer vendors. If vendor "A" is not satisfactory, one can easily move to vendor "B" or another vendor without the huge porting costs of moving among UNIX machines.
Suggested Suppliers:
Why do it?
Cost of Windows solutions tends to be lower. Windows RFP responses tend to be 25% the cost of UNIX. New hardware peripherals are introduced on Windows Server before UNIX. Therefore, it is faster...and may be easier to add hardware in a Windows environment. Customers want a Microsoft look and feel. I add that it may be easier to find new hires that know Windows than know UNIX. I also think that with Windows, future hardware ports within the Windows world should be easier. Also, the operating systems world is going Microsoft. The feeling is that no one can beat them in the long run. They are taking over from UNIX servers. They have a telecom division headed by a former telecom exec. They have SS7. The top hardware vendors (HP & Compaq) have announced moves to NT.Windows Server isolates apps from the Control Computer...making the computer hardware question almost irrelevant. The control computer is tied to the operating system it runs on. Let's look at UNIX. UNIX is not just one operating system. Each computer hardware manufacturer has their own flavor of UNIX. Therefore, if your computer is running UNIX, your platform is tied to a particular manufacturer's computer hardware. This is why, in the telecom world, we always hear about "what hardware platform it runs on" or "what is the next hardware platform"...because the app must be customized to run on each flavor of UNIX.
Microsoft does not have this problem with Windows Server. The hardware vendors that use Windows have agreed to not change Windows and use the one version that comes from Microsoft. Therefore, Microsoft does not have to customize Windows for each hardware vendor... The effect is.. if you use Windows, it doesn't matter what hardware vendor you choose, it works.
By the way, UNIX tried to get a standard version back in 1988. The effort failed because competing hardware vendors wanted to say that their version of UNIX is better than the next vendor's UNIX...hence keeping multiple versions of UNIX
[The Road Ahead pgs 60-61]
In conclusion, I suggest using Windows Server to fix the hardware restriction problem.
Note: Linux may also have some of the same advantages as Windows.
Anthony Clark
