SOA Governance Monday (Okay Tuesday): Design time, and the missing pieces, part II

Core issue with design-time SOA governance technology is how deeply the technology goes toward serving the true notion of 'design'

I had a bunch of good comments and a bunch of e-mail around my last "SOA Governance Monday" post. Also a lot of self-serving shout outs from the vendors out there (I do check the IP addresses guys). So, I figured I would provide some clarity today, after everyone is rested from a nice three-day weekend.

The core issues with design-time SOA governance technology is how deeply the technology goes toward serving the true notion of "design." Most don't go that deep, and many SOA designers are left wanting more robust features and functions, including true modeling and simulation capabilities based upon SOA design and development best practices.

Thus, most SOA architects are bypassing this technology altogether, and for good reason. There is design going on around governance in the narrow and SOA holistically, I've seen it, but they are not using design-time SOA governance technology. This was further validated based on the amount of pushback I'm seeing out there in the field, and here on the blog. Of course, each vendor has its own approach, there is no consistent design pattern leveraged, which is another issue I brought up last week.

So, how do you fix the issues with the existing design-time SOA governance tools? I'll do the best I can in 100 words :

First, you work on a high-level and practical approach that makes sense for most SOA problem domains, and figure out how that approach binds into a larger SOA design and development methodology. There simply is no context with the existing tools today, thus no value.

Second, you're either a run-time tool or not, so figure out how deeply your tool needs to drive into that part of the architecture. My advice would be to do it well, or don't do it at all. There are many run-time SOA governance tools that are already doing it well these days, you're going to have to figure out how you orbit around them.

Third, learn how to define a clear ROI for this technology. Right now they are just being sold along with the hype, not sure if you're up on current events, but the hype is not driving SOA anymore, results are.

Finally, get together and create standards, technology, and approaches that are consistent from tool to tool. Right now they are all different, thus proprietary, and that's a hard sell these days when people look at durable standard technology that will be strategic over a long period of time.