A few weeks ago I was at the ICSE 2007 and I envisioned an entire enterprise of applications build on top of the Salesforce.com’s APEX platform. I just got out of a TechEd 2007 presentation that demonstrated how to really build diverse applications using the Dynamics CRM as an application platform. I got really jazzed because they would handle the backend upgrades and we could handle the configuration elements we were most concerned with.
Instead of having two choices of build or buy for our organization, perhaps we have a third choice. I’m not sure what a good name is yet for this third option, but it involves leveraging the Dynamics CRMplatform to build out almost custom applications.
I get more excited about this than I did with the APEX platform because I have an existing Microsoft developer base than can really leverage their experience and bring it together inside of unified application architecture. I really like the idea of a unified application infrastructure, maybe because it feels more familiar than a message based SOA infrastructure. With SOA I can isolate key processes and upgrade them individually. With the common infrastructure approach I can quickly see us paying a lot in custom development every few years to reintegrate our new infrastructure.
Perhaps a SaaS application platform is really the way to go, and until we get there we can leverage the existing infrastructure using a SOA as a backend.
Subscribe to:
Post Comments (Atom)
3 comments:
The good news I think is that softare vendors are starting to get it too and really the pace of change in integration interfaces is starting to slow down a bit. It's still a very new industry though and folks are still figuring out the real cost of changing things on both the vendor and consumer sides.
yes, change is the hard part. I love the implied efficiency of purchasing a configurable package over custom development. Here are some questions I struggle with:
So when they come out with the new release, and we upgrade, do our custom configurations break?
If so, how much work is it to detect breaks and repair them?
As one who's test infected, can we write automated tests to confirm our configurations are working? Then at least we have a suite of tests by which to efficiently measure impact of an upgrade and progress in upgrading.
So I went over to the Dynamics CRM booth here at TechEd and posed my question on this topic. He told me that their group heard the upgrade and migration story that Salesforce has been out in the market with. He couldn't say anything official, but the upgrade from 3.0 to 4.0 will not be painless, but won't be as painful as a traditional upgrade. We talked about the importance of developing to the APIs. He also mentioned that it's really important to have a vendor that knows what they are doing and not someone who can do the work.
Post a Comment