By: Boro Petrovic, Chief Technology Officer, CTO, MICROS Systems Inc.
This is the first edition of MICROS’ monthly “CTO Corner”. At the end of every month one of our 3 CTO’s will be giving you their insights into technology, the industry, and how it affects your business.
I took a flying lesson recently. Prior to the flight, my instructor introduced me to the plane. While I was marveling the twin engine flying machine, he was going through the detail check list. Holding a letter sized plastic sheet with text written in bold red font he checked tires, wings, poured a bit of gasoline from the tank into a glass and inspected it visually, and so on and so on. After a great flying experience, while driving home I thought about the three legged stool analogy in this case. First, the plane was in a perfect shape, second, the Instructor was well trained and the third the standard operating procedures (check list) were in place and followed. All the three legs were perfect.
Weekend is over and I am back in the office. This time, turbulence in the world of IT makes me think again about the three legged stool analogy for a stable, reliable IT system. The hardware, system software and applications make one leg; the skilled personnel that manage it make another leg and the third leg is operating procedures used by the personnel. Are all of the three legs always in place? Are all IT systems reliable and stable? Unfortunately they are not.
“What do you mean it may not be working?! Those are plug and play components and all the vendors’ specifications state that this works!”, exclaims a young IT manager who is in the process of engineering a complex system (multi-tier cluster, SAN, virtualization, OS, database and application servers) using commodity components bought from different vendors: Servers from Vendor A, SAN from Vendor B, OS Vendor C, Database Vendor D, and application from Vendor E. And indeed, most often, upon connecting and configuring all components, everything appears to be working fine. I am trying to imagine my flying experience with this approach. Plane chassis, engines, instruments, etc. all arrived in separate shipments and are sitting on the hangar floor. I read the accompanying DIY manuals and after a few days of work putting it together, the plane is ready. Turn the key, the engine starts, propeller rotates, and all the indicators show two thumbs up. Engine is humming and I should take a flight…Would I? No way, of course. Most people wouldn’t. This is the job for test pilots.
So why is it that IT managers would not board a plane that is assembled in a garage, but would implement a mission critical IT system put together that way ? It’s because of the intuitively perceived cost of risk (COR). In many cases, the cost of adequate testing, or acquiring an enterprise class, preassembled system with a staff of skilled engineers is cost prohibitive, in which case IT managers have no choice but to take the risk! Or do they?
The alternative is to find a way to share the COR (Cost Of Risk). For example, let’s say the cost of testing/certifying a new configuration is $100,000 and let’s say that there is going to be 10 implementations of this identical configuration. Then the cost of $100,000 would be spread across of the 10 standard implementations, since testing/certification would be required only once. In that case, the testing/certification cost gets reduced from $100,000 per unit to $10,000 per unit. The same applies for the cost of developing standard operating procedures. A staff of skilled engineers required for 24×7 operations would have a lot of free time on their hands if it is only one configuration that they have to take care of. However, if there are many, same standard configurations, then the staff can be highly specialized allowing lesser cost per one configuration.
If the application software vendor is MICROS, then one alternative is to host the application in MICROS hosting facilities. That application instance would be one of the many, hosted in a standard configuration, allowing lesser cost per configuration.
That way, you can fly a plane that has been tested and operated by specialized pilots using tested procedures.
Sometimes, the need for all three legs has not been recognized. The perception was that only one leg is necessary, the SW and HW leg, and the other two are “overhead”. In those case we are often, well,…taken for a rough, unpleasant, and risky flight.