IT Nirvana

Over the last few posts I have raised the unfulfilled concept that was prevalent in the early days of Open Systems - namely the Corporate or Enterprise Database. This concept was where corporate data would be stored and accessed centrally from a central managed place. Of course the central database never happened and there are lots of reasons for this, but I like to think of the Corporate Database as a kind of IT Nirvana. Twenty years ago if we could show IT managers what their systems landscapes would look like today I think this idea may have taken off and just possibly we could have a achieved that blissful IT state.

Instead we face the following issues:

- Systems complexity - get your local IT architect to print out your current systems landscape. Even better get him to walk through all your interfaces.
- Data Warehouses - Often the holy grail of the data warehouse is the search for a single view of the truth. This is because your multiple systems will all have different versions. If the data was kept in one place then we wouldn't need to build expensive and complex data warehouses.
- Master Data Management - The Corporate Database is the MD solution
- Enterprise Data Bus - Why would we need a bus if all the data sits in the same location?

So are we likely to find spiritual enlightenment in IT soon? Well not so long as we can't see past the next quarter, or FY budget for that matter.

Congestion Blues

My average work commute speed is about 8mph*, which is just better than the average London rushour car journey at 7mph - worse than the horse and cart achieved at the start of the 20th Century.

Cars are great so what's went wrong? Well the problem is the tarmac and the fact that we all want to use the same bit at the same time. So if that's the problem then what are the possible answers?

- Build more roads - yet studies have shown that new roads encourage people off public transport and into cars with no net benefit to congestion.
- Build great public transport - yet most governments are too bankrupt to fund the sort of public works programs required to make major infrastructure improvements here.
- Work from home - the best commute is the one you don't have to take. With boadband and mobile technology we have the infrastructure and tools that enables us to really embrace this solution, but so far it's only a minority solution.

So what's this got to do with IT? The answer is below:

- Build more roads - is the equivalent of building more interfaces between our systems. Except that in our case each road is made from different materials, using different construction techniques, different traffic lights and different road signs.
- Better public transport - is the equivalent of building an enterise bus architecture. A great potential solution but still very costly and complex.
- Work from home - back in my earlier post "Redundancy Blues" I lamented the lack of the single corporate database. Imagine if all the data our company held was in one database. We would ensure that it ran on the best hardware, backup and recovery would be easy, we wouldn't need to build data warehouses to try to discover a single version of the truth. In short,if we had realised the potential benefits offered to us by Open Systems we'd be in a far better place than we are today.

Oh, well, must go and get my pick axe. I'm building a new road today.

*Please note that I consider myself very lucky that my commute is only half an hour. Across the world many, many people spend about 2+ hours per day commuting to work and back - that's about 25% of the time that they are actually there!

Where did my Open Systems go?

Early in my career I bought into the concept of 'Open Systems' in a big way. I'd been working on proprietary mainframe systems for a couple of years and figured that 'Open' had to be the way to go.

For those unitiated in the world of Open Sytsems it effectively was a euphamism for UNIX and RDBMS technology. The 'Open' basically mean that systems could be developed on one platform (say Sun) and ported to another (i.e HP or IBM) without much effort. Portability was the key to keeping your vendors sweet and it was also ensure that cross skilling your staff was easier. The same was theoretically possible with SQL based databases.

So did this happen? Not really. I used to work in a software porting team for a UNIX/RDBMS based product and because we applied strict standards to keep our code open we managed the transition without too much effort. This was not typical though. Most IT guys out there will tell you that porting from one server or database to another is a major undertaking.

That has resulted in the vast majority of IT shops out there often selecting a particular flavour of UNIX or RDBMS as its system of choice and that kind of defeats the whole argument of Open Systems if you think about it.