Architecting an organization's web strategy. What is the preferred method here?

Question

I'm working on a solution for an organization that needs to last them long term (as subjective as that is).

They currently have a Joomla installation whereby the CMS talks directly to the database, from there I have started "bolting on" the additions that they need for their environment (mobile app, kiosk, and lobby display). I've also started writing a RESTful service to talk to Joomla (not the joomla database) and the apps.

Here's a diagram of how it's currently going.

^{note: they already have the "public & admin" as well as the central db, google cal, email, and vimeo}

Since I'm just getting started with this, I'm thinking that I should probably push for an appropriate design from the ground up. I'm suggesting this so that they have a good foundation moving forward.

Here's my suggestion to them

proposed architecture http://www.gliffy.com/pubdoc/3221292/L.png

Being that they already have some pieces in place, is it worth pushing for a new architecture? Also, am I on the right track with regards to the new architecture?

Lastly, it's worth noting that they're not totally satisfied with Joomla and are looking for a possible rebuild for that portion of their project anyways.

Answer 1

First of all, there isn't really much of a comparison. The current architecture obviously grew as bits where added, and your approach is a re-design, the obvious advantage being you know exactly what the layers are and how individual components interact. It's obviously the better approach, you are introducing a middle service layer that essentially makes the top layer "data storage" agnostic.

There isn't much to discuss really, what you are proposing is fairly close to a typical three-tier architecture. The only possible disadvantage is that there is no obvious fall back for the middle tier, the CRUD web service. If that fails for whatever reason, everything fails. Depending on the company and the availability needs for your services, that might or might not be important, but always worth considering.

Since this is a proposal, I feel obliged to warn you that it's a fairly radical change. If it was presented to me on one of the mature projects I'm working on, I'd need a fairly good estimate of the cost before discussing it. Typically radical architecture changes, especially from the ground up re-designs, happen gradually in quite a few iterations. Don't be discouraged if your team rejects the approach, but if that happens make sure you evaluate any feedback and come back with an adjusted and a bit less ambitious proposal.