Ben Northrop

The Current State of SOA

• Martin Fowler tries to untangle what people mean when they say "SOA". In his assessment, the term is completely semantic free, and is beyond salvaging. This was in 2005! David Ing piles on.
• Another great article by Martin Fowler. SOA, for many, is just "planned design" cloaked in a different robe. It entails that we have the time to design the "right" service APIs up front, and re-use them later. Even if this was the case, many businesses don't have the time. Fowler asks, can SOA be done in an evolutionary way?
• Anne Marie Manes proclaimed that SOA is dead, but services live on. Some excellent replies to this: Kurt Cagle, Gary Barnett.

SOA and the Database

• A nice, tangible scenario of how to how to shuttle data between services using a publish-subscribe model, so that services can keep data locally.
• Explanation of the Data Service pattern: "Data services abstract complex data structures and syntax. Data services federate disparate data into useful composites". Basically, it seems that Data Services are like a database view at the service tier. Not sure if that works in practice.
• Great article from Microsoft, laying the core design principles of SOA and then making the distinction between data inside and data outside.
• Good overview of transactions in the SOA world.
• How do you manage foreign key constraints within an SOA architecture. The answer: data synchronization! Another good link on this.

SOA and the User Interface

• Service Oriented Modeling and Architecture

SOA and Loose Coupling

• Good, honest viewpoint of what loose coupling means for SOA.

SOA and Security

• A great introduction to security in service oriented architectures, covering at a high level SOAP message monitoring, federated authentication, application proxy, contract management, certificates, keys, and encryption, and audit logging.

Canonical Data Model

• Cool animation of the benefits of SOA and a canonical data model.
• Solid argument for boiling your CDM into its least common denominator, rather than a superset of all possible fields associated with an entity.
• "Systems that pass data to each other share commonly understood semantics. Explicit data semantics is the key to success in an EDA". CDM's help create commonly understood semantics.

I'm currently living in Pittsburgh, PA, working as a Senior Technical Consultant for Summa, and just finished up my Masters degree in Philosophy. (thesis here!) My blog is mostly on understanding how to better analyze, make, and influence decisions in the world of enterprise software development. Subscribe here or write me at ben_northrop at yahoo dot com.