In conversation with a client of mine recently I was reminded of the need to always look deeper before embracing technology change. My client's IT architecture consists of a number of packaged applications, with one core ERP style application at its heart. Interfacing between these applications is currently achieved using a number of ETL style interfaces. Not necessarily pretty, but it works.
The vendor of this core application has recently announced that it is effectively rebuilding the application and in a SOA style. Sounds good on the face of it. My client and I were having a conversation as to what this would mean to him and his business users. We talked about all the nice things SOA would bring around loose coupling, agility etc. and agreed that these were good things. However, when we asked ourselves why this would be a good thing for the business users we came up short - there were very few reasons we could come up with regarding why they would (as opposed to should) care. It was a classic example of why it is so difficult to make a business case for SOA. We instinctively knew it made sense but couldn't assign any monetary value to any of it. Sure it might bring benefits in the future, but not now.
My client will migrate towards a SOA, at least in the context of this core application, but it will be driven by a need to stay in support as opposed to any great desire to go through the large migration project and all the headaches that the upgrade will inevitably bring.
We are both big advocates of SOA, but in this instance and in the current budgetary climate we were left saying SOA - So What?
Subscribe to:
Post Comments (Atom)
Simon,
ReplyDeletewhile agreeing with the sentiment of your blog, sometimes IT needs to make an infrastructure call, which may not meet an immediate business need, but sets the business up well for the future. Whether the current "holy grail" of getting wider business sign-off should always be obtained for this type of work is another debate.
Ian.
Hi Ian, thanks for the comment. Absolutely agree with your thoughts. The point I guess is that this is a classic exampe of how hard it is to get the business to sign off on these type of investments. Earlier in the year I visited a large bank in Europe who had made a very significant investment in their SOA platform over a number of years but getting an ROI from it was proving impossible especially now that getting funding for the 'consumer' projects of the infrastructure was now nigh on impossible for them - yet another nice twist. Hopefully some of that will resolve itself as projects get funding again in the coming years (we all hope!).
ReplyDeleteHi Simon,
ReplyDeleteAbout the rebuilt ERP platform. I think it’s worth remembering that when software vendors have to re-build their core platforms, it is not for their current clients’ benefit! It is to get the next wave of clients. They then have to 'spin' to existing clients that is good news. Stay sceptical and believe the benefits when you see them!
The client may well need new infrastructure, which does not have an immeadiate business need visible from outside the IT department. But this may or may not be it. The new platform might include features which could potentially be useful. But, until a benefit has been identified, it's just a cost. And by ‘identified’, I mean identified in your particular client’s environment! Not in the labs or some case study.
To think that the new stuff is automatically beneficial, is basically Gartner’s ‘Mistake No. 1: Irrational SOA exuberance’.
http://www.gartner.com/it/content/754400/754413/twelve_common_soa_mistakes.pdf
The new platform may well offer some new possibilities, but it definitely won’t do all the work by itself (if it does, I want one for Christmas). For now, I think any upgrade job should be looked at a cost of ownership.
see ya,
John Marmelstein