Bound to legacy systems

Without giving away too many details about my regular place of employ, some recent work there has reminded me of the joys of dealing with IT (Information Technology) users that box themselves into corners by letting themselves be bound to legacy computer systems.

What am I talking about? There are many systems around my work place, or in other areas within the larger organization, that can't be replaced and/or retired and wind up pretty much frozen in time because some amount of work was done that involved those systems and was done using the specific operating system, with the specific patches, and certain specific applications and development tools on them. All of those specifics can't be changed and/or won't be changed without an absolutely unbelievable amount of consternation and complaint and concern by the users or people that work with the users.

For example, there may be custom software or in-house developed software that was written using a certain development language. Because that software was written and compiled using older tools, those tools can't be replaced, nor can that software. Over time, the problem snowballs upon itself because the tools, the older operating systems, the older hardware, everything ages to the point that the vendors that made it won't support it or if they do, they'll charge outrageous amounts of money to continue to provide support for equipment that they really don't want to support (as it would cost them a ton of money to dig up replacement equipment and/or software if they had to find it themselves). The reason that those vendors do provide support is because they are pretty much forced to do so otherwise they might not be able to sell new equipment to the customer.

All of this grates on my nerves a bit because I know what it is costing on many different levels and feel that it a big example (in many ways) of what some might term waste, fraud and abuse. It's not fraud though, and it's not abusive, or at least not intentionally so. It is however somewhat wasteful and is part of the huge amount of hidden costs that wind up getting tacked onto projects that take on a life of their own and seemingly will never die because somebody, somewhere, is using them.

continued in the comments area. Please continue reading there!

Locked Post

Advanced Search

Watch this post

Do not email me updates for this post Email me updates for this post

Successfully updated karma reason!

Thanks for continuing here. Article continues...

I'll give a few examples that are not directly related to my work, but are good examples of the types of things that wind up happening over time.

Example #1 - the Mars Rovers that JPL (Jet Propulsion Labs) and NASA sent off to Mars. Designed and built for a short mission live, but still going strong, or at least still going. They are amazing devices, but they are also amazing devices that have continued to suck up research money and government money as they keep going, and going, and going. And they'll continue to suck up money that most people wouldn't expect because once they were built, the equipment that supports them was locked in time and won't be changed for fear that any changes that are done could result in a catastrophic failure of the systems or the rovers.

Example #2 - the satellites that have hung around and lived well past their initial life spans, such as the ones that have helped track hurricanes. They serve very useful missions and provide great scientific value, but they cost a lot of money to keep going because in order to support them there are systems that were built back at the same time, or well before the satellites were built and launched, that are still be used to receive the data from those satellites or to command and control the satellites with.

The same things happen to other government areas, and pretty much anyone that lets themselves fall victim to assuming a dependency on a specific operating system or specific tool or application. You get to a point where you can't upgrade or patch systems because if you do, there is no telling what the effect on the systems and application that run on them will be.

No money has been budgetted for the continued testing, and continued development, so instead money has to be budgetted for continued support of those systems. You create your own monsters by having to keep around Systems Administrators that will be able to support the old equipment and systems, and those people wind up having skills that deteorate over time from lack of use.

Good IT managers typically refuse to let themselves fall victim to these type of problems as they demand the use of industry standard systems and require that users continue to keep up with patches on their systems. Still, there are untold hours of testing that have to be done to make sure that the systems will function as intended and that the results of the testing match the expected results, or the results of every previous combination of runs that have been done. Scripting work has to be done, and/or test protocols have to be created and documented so that procedures will exist the direct how to confirm that the upgraded or newer systems will function as proper replacements for existing systems.

Things get more interesting over time when you find that some systems used hardware that is no longer found on new systems. For example some devices may have communicated over serial ports. 5 - 7 years ago those were plentiful on systems, but more and more those ports have disappeared because USB is able to handle those types of devices. Except that the legacy software may have been written in a way such that it tried to communicate directly with COM1: or COM2: while USB provides some totally different names and addresses for those ports.

Anyway, it is an interesting problem and continues to haunt my days, and the days of many other systems admins out there. Users that never knew to demand that software they were having developed be developed using highly portable code find later that they can't replace that software or get it moved to other newer systems because they can't get it recompiled for those systems. Or perhaps they can, but they have to pay a small fortune to have it done, and then tested to a level of satisfaction, etc.

Hopefully over time more users or managers of users will demand that the users of IT equipment and systems in their areas will not let themselves fall victim to the trap of being bound to legacy systems. If they succeed, they could save untold millions of dollars in long term support costs.

Successfully updated karma reason!