SOA: Making the Paradigm Shift Part 6 of N - The Current State of SOA and How to Make the Paradigm Shift
The Transition
| Old World | New World |
| Monoliths | Services |
| Tightly Integrated | Loosely Coupled |
| Process Centric | Workflow Enabled |
| Islands of Automation | Integrated Service Networks |
| Accountability | Results |
| Bending the enterprise to the will of technology solutions | Adapting technology to meet the needs of the enterprise |
The Current State of SOA
Most of what is labeled “SOA” today is really just Bottom-Up JBOWS. It’s not that these home-grown, spring-like-weeds services are not valuable. It’s just that there is no vision or alignment with the business drivers of the organization. Developers are kings and left to do what they want and there is no standards or governance. Reuse is impossible in this scenario but worse yet, each department is doing just what they need to get something up and thus producing another silo. It’s no great surprise when such “SOA” fails.
As I will show, in a later installment, the solution is not to take the opposite of JBOWS and take on Top-Down SOA. Such SOA is often referred to as “Boil the Ocean” SOA. An “Enterprise SOA Committee” is formed. A two-year effort begins to identify all the service candidates in the Enterprise. This kind of SOA is frequently pushed by IBM. In Microsoft’s approach, we see a “Middles-Out” Approach that is both Iterative and Practical. But more later.
- Lack of Compliance/Governance
- Don’t have every developer doing their own thing
- Security, Common Policy Repository, Common Schema, etc
- Deployment, Management, and Change are ad-hoc
- Centralized Management Experience
- Improve Execution Visibility
- Changes are risky and disruptive
- (Service Hell)
- Low Adoption Rate
- Nothing to encourage the construction and reuse of services
- Existing tools instigate retraining more than reuse of skills
Getting There
Be Practical
-
Avoid Common Pitfalls
-
Gain immediate results
Also see my colleague Brian’s tips here