Blogger: Richard Watson
It was standing room only yesterday at the Software as a Service (SaaS) sessions at our Catalyst conference in Prague. There’s a real thirst for insight into SaaS.
In the first session after lunch, Mike Rollings opened our eyes to architectural implications, and non-trivial data and process integration issues of adopting SaaS. Mike identifies the biggest impact of SaaS as “the elimination of IT boundaries”.
One of our concerns at Burton Group is improving the relationship between enterprise IT and the business they support. After years of disappointment, the relationship can sour. In the most dysfunctional cases, business regards IT as irrelevant, certainly not the solution provider of choice. One danger of SaaS, as Roman Stanek of Gooddata.com quoted is “the CIO is the last to know” [about a SaaS contract].
When discussing SaaS, I’d like to see a SaaS suitability rating for each class of applications, like an electrical appliance energy efficiency rating, or credit-worthiness. So maybe we’d give ‘AAA’ to the productivity suites that Guy Creese surveyed yesterday evening and maybe a single ‘A’ rating to CRM. Beyond that, right now, I’m a sceptic. Enterprise applications participating in a next generation architecture may not fit – and be graded no higher than junk.
My scepticism recognises the challenge teams will experience when eliminating application boundaries and creating a seamless cross-application user experience. Removing the skin and unpicking the tissue of assumptions that holds enterprise applications together is painstaking work. The data semantics, trust and infrastructure assumptions twisted into the tissues cannot easily be mapped into a multi-tenanted homogeneity.
Despite maintaining a healthy scepticism, we’re doing a lot at Burton Group to lay the groundwork necessary for taking those boundaries down:
- The Infrastructure Service Model (ISM) is our model of service-based infrastructure, created by Jamie Lewis and Anne Thomas Manes. One of the key features of the ISM is that it enables policy-driven management and control of infrastructure capabilities.
- We have Chris Howard to thank for relating the concepts in Stewart Brand’s book How Buildings Learn to layers of system architecture.
- Kevin Kampman’s work on interoperable identity services
- Bob Blakley’s research on relationship-based and federated identity
- Dan Blum’s work on Shifting Defenses, Zones and Dynamic Perimeters”
- Joe Bugajski’s overview of “Data Integration: Fantasies and Facts”
- Guy Creese’s evaluation of SaaS versions of the safer AAA, and AA rated applications
- Craig Roth’s overview on surveying more than 300 SaaS implementers
- Eric Maiwald does a great job of combing the fine print of SaaS contracts to highlight security and operational risks in “Considerations for Risk Management When Choosing Software as a Service”.
- and not forgetting Jack Santos’ “Software as a Service and GoogleApps: IT Strategy Implications”, Mike Rollings’ full perspective on “Architecting for SaaS” and his “Related Research Summary: Cross-Domain Context for Build, Buy, or Borrow Decisions” which provides a more comprehensive list than this of Burton Group research as a context for SaaS.
Everywhere you look, both here at Catalyst this week, and in our research, we’re taking on the challenges of SaaS so that our clients can reap the genuine benefits of off-premise, on-demand IT.
Are we there yet? No, but we’re getting there thoughtfully, so that our clients don’t create another generation of legacy technology silos.
No way out (“Huis Clos”) is a play by Jean Paul Sartre about dealing with the boundaries created by our minds. Garcin and the other characters looking for an exit came into my mind yesterday. The most famous line from No way out is “Hell is other people”. By the way, I don’t believe this is Sartre anticipating multi-tenanted databases!