Technologically, I believe this move to a "Buy-and-Integrate" mentality (vs. a "Build-Everything" mentality) will open the door to a new class of products to assist with services integration. And, if you agree that the importance of leveraging external services will be elevated for the CIO, then I believe a significant enabling technology will be a rebirth of the need for a robust "service integration bus".
Why? As I mention in my blog, enterprises integrating external services require:
- Identity and access management for each provider
- Data compliance, legal and regulatory audit access across each provider
- Security compliance systems
- Provisioning, including capacity forecasting
- Performance (e.g. SLA) monitoring
- Cost and budget tracking (i.e. for billing, showback and/or chargeback)
- Disaster Recovery / Redundant service sources where needed
Some would call the above integration functions "Glue Logic." Indeed in the past, many of these functions were hand-integrated across the few external services that were leveraged, and custom-engineered into each internally engineered stack. But time is changing the model. With more turnkey services sourced from cloud (IaaS/PaaS/SaaS) providers, the need for a more efficient integration function will escalate. Integration will need to be standardized and replicable, scalable and responsive to the business' needs.You may recall one component integration approach has been the Enterprise Service Bus, primarily associated with SOA leveraging SOAP protocols. This Integration Bus was originally to orchestrate access and workflow between component services within the enterprise. (By the way, Microsoft offers a great overview of ESBs - albeit BizTalk centric - here on MSDN).
I believe that the "2.0" Integration Bus will be one which brokers higher-level services generated from external, public cloud providers - not just internal component services. And it will use more generalized interactions than SOAP, since the providers and their environments will be less standardized.
To this end, there are some great current/upcoming thoughts suggested by Mike Ponta of the notion of a "Cloud ESB", and can't wait to hear more. A quick survey of the market also yielded what looks to be potentially promising integration technology coming out of Mulesoft called their CloudHub.
Conceptually, the "mashed-up" service was the 1.0 version of this integration concept. But as enterprise IT begins to regularly tap and integrate multiple external services, the 2.x integration busses will need a more structured, standardized and rapid approach to integration.
I can't wait to see what else the market will generate. Stay tuned.
- The CIO as a Business Service Conductor Blog by Andi Mann
- Is today’s CIO really just another supply chain manager? Blog by Charlie Betz