6 comments

[ 3.1 ms ] story [ 27.0 ms ] thread
I don't think this is a joke, but sentences like this make me wonder a bit:

    If Snowman Architecture sounds too informal to you,
    feel free to refer to it by it's formal name: Vertically
    Aligned Synergistically Partitioned (VASP) Architecture.
It probably isn't a joke. I see stuff like this all the time.
Sorry if this stings a bit...

We already sort of do this (without the bad bits) and have done since about 2001.

This is a little buzzword filled and doesn't deal with the system-level practicalities of enterprise architecture. It seems like an internal technical proposal written up to sell to the business rather than a practical discussion. Mind you, what can you expect from a CTO...

Regarding practicalities, it's a royal pain in the arse managing component and messaging contract versioning between "tall" architectures like this.

Also it doesn't work well with a SaaS model as the logical partitioning is more complicated with different isolation requirements for different clients. Complexity goes up O(log N).

Synergistic partitioning is pretty obvious and doesn't require an essay.

Keeping things within budget is about managing scope and coupling, that is all.

Nothing new here.

Thought the same thing after a quick scan. This is kinda how we do EA already...

However, EA literature often repeats the obvious, maybe we're not the right audience...

eckyptang: I think you may be mistaking synergistic placement of methods at the technical level (which I don't recommend) which synergistic placement of methods at the business level (which I do.) And you are right about the difficulty of managing messages. This is why the "simple" concept is critical. You might check out my Math of IT Simplification White paper for more on this. - Roger
"Wow, this just what we've been looking for! Where have you guys been all these years?"

The sad thing is it's not what we think that matters. If clients fall for the buzzspeak, it's irrelevant if that what they're paying for, believing it's "new" and "different", already exists.