So I have a ton of feedback from a number of people about my SOI blog. There are a couple of key points that I think I have not been clear about:

  • This blog is not the answer to life, the universe and everything in Infrastructure... yet :). Its a set of  top down ideas which will hopefully coalesce into something useful. Its a journey, not the destination.
  • This is about infrastructure, not the whole enterprise. I have specifically excluded anything that is business, project developer or process focused which of course includes ESB, SOA, workflow, Orchestration, Tools, SDLC. It is about the common computing elements used by an organization, not how those elements are used for business value or projects. It is not a model for computing in the enterprise, that is a topic that a lot of other people are working on and is rolled up in EA, ESB, Connected Systems, Enterprise Service Architecture or whatever else people want to call it.

Clearly a model for computing throughout the enterprise is critical and is what the Microsoft Connected Systems story is about. This blog is about the Infrastructure or platform part of that.

 

A number of people have pointed me to other articles and work going on in this area. For example there is a "Superplatform" report (snappy title) from the Burton Group. The super platform is SOI with some additional stuff that I have left out (specifically SOA and applications such as mail etc).

 

There is also an interesting report from MWD on "What drives service infrastructure deliverables" (thanks Neil)which looks at what are the motivations behind Service Infrastructure (without explicitly defining what it is).. The report focuses on quality and commercial drivers and in particular looks at Contracts. I went to an interesting talk given at the Microsoft internal technical training event (Tech Ready) entitled "Systemic Quality Impact Analysis" last week which also looked at SOI from a quality aspect and have just reviewed a draft whitepaper on the same topic by  Lewis Curtis, George Cerbone and Scott Anderson which I hope will be published soon. The interesting thing about SOI is that it as a poor business case (businesses are very "project" focused) but is critical to the health of the organization (most business critical apps wont run without the network for example). So whilst we think of the business value of SOA we have to think of the quality and contractual implications of SOI.

 

Finally I had a number of people asking me about infrastructure topologies, i.e. how we can support specific physical scenarios (branch, retail, cluster, geocluster).. I hope to get to these elements as I drill down through the levels. Architecture is a bit like archaeology!

 

Is anyone aware of any other papers or documents which are looking at or discussing the service infrastructure area? Please post or send me links to anything you have seen... I'm an incredibly fast reader and will read and review them.

 

Thanks for all the feedback!