CBDi 저널에 실린 SOA 관련 기사를 보면 지난 해까지 각 업체는 SOA 관련해서 제대로된 제품 라인업을 갖추지 못한 채 인프라 차원의 제품군만을 제공했었고 이로 인해 SOA를 채택한 대부분의 프로젝트가 쏟아부은 노력에 비해 별다른 성과를 얻지 못했었다고 적고있다.
기사를 대충 훑어본 바 올 해는 보다 실질적이고 구체적인 실행 지침과 방법론, 그리고 통합 플랫폼과 관리 툴등이 등장할 것으로 예상되며 SAP 의 ESA, Oracle 의 Fusion 등을 눈여겨 볼 필요가 있을 것같다. 아래는 기사 중 일부 인용한 것.
1. Practice guidance and methodology get hot
Why? Enough people have now realized that their first SOA attempts didn’t cut it. That just buying technology doesn’t deliver SOA. Rather each enterprise has unique requirements for SOA that address the complexities of existing environments and applications together with new service based strategies.
We expect simple concepts like SPP (Service Portfolio Planning) and RISS (Rich Service Specification) to get traction.
2. (Service) Modeling makes a comeback.
As enterprises mature their SOA thinking, they will look for repeatable, systematic approaches to SOA and they will turn to modeling, looking for trusted techniques that allow them to manage their business processes, services and projects. The vendors have mostly allowed this opportunity to pass them by, and there is big opportunity here for the brave.
3. EA Evolves into ESA (Enterprise Service Architecture)
For the same reasons that first SOA attempts are deemed to be sub-optimal, enterprises are set to place more responsibility and accountability with enterprise architecture teams. Enterprises want and need shared services, they understand the benefits and are going to realize that individual project teams need
to work within city planning like constraints. However EA teams are not universally accepted as providing practical solutions to business problems; they are often divorced from the action. We expect the demand for SOA implementation will drive the evolution of current EA practice to adopt real accountability with shared responsibility based on good working relationships with LOBs and projects that facilitates the development and evolution of practical SOA policy.
4. Integrated platforms and management tools get to first base.
One of the reasons that adoption of SOA and Web Services has been slow is that the platforms and tools have not provided integrated solutions. It’s easy to deliver services, it’s hard to manage the entire lifecycle of delivery and operation to agreed SLAs. The rash of standalone products and platforms are easy to acquire for individual projects, but not for enterprise adoption where service lifecycle and
operations management needs to integrate with the mass of existing technologies and tools already essential to day to day operations. The work of CA, IBM and HP in this area will start to pay off in 2006, and we expect consolidation around vertically integrated platforms to be a major battleground.
To survive standalone tools will work hard to consolidate an existing niche.
5. Enterprise Applications show the way.
Talking of vertically integrated stacks of course the enterprise application vendors are right in the middle of this marketplace and in 2006, led by SAP with Oracle following in 2007, they will be delivering SOA based application components. This will be a major catalyst in the marketplace and cause many enterprises that have been sitting on the sidelines to get their feet wet.
6. It’s Services AND Components.
We figured a long time ago that services were just the tip of the iceberg. While services provide lots of benefits the long term goal has to be the componentization of the application base in order to deliver real adaptability, and services are a practical strategy to achieve this goal. Watch closely to see the quality of componentization in the SAP ESA and Oracle Fusion (and of course other vertical application providers’
products). Those with good component based architecture will be nimble enough to lead the market in the long term.
7. The Next Big Thing after SOA – Supply and Assembly Paradigm!!
You can bet your bottom dollar every vendor is asking themselves right now – what is the next big thing? And I wouldn’t be surprised to see a focus on this during 2006. As SOA matures we will come to see business solution acquisition and delivery using manufacturing and assembly patterns, in a very similar
way to the auto and or high tech industry. Supply of core business services is a classic provisioning problem and requires rigorous contract agreements that manage the supply chain.
Core business services form the motherboards, subframes and platforms on which application specific solutions are delivered. The lead time for core business services will be shorter than silicon or auto engineering platforms, but not much. At the same time solution delivery projects will use increasingly
high abstraction programming models such as DSLs and BPM to integrate the service layer with the business process and knowledge management layers.
원문 : CBDi Journal 2006 / January, “PREDICTIONS FOR THE COMING YEAR”
P.S : 이 기사 출력해서 훑어보다가 ERP 관련해서 SAP 기술 영업하는 분과 만나러 갔었는데 그 자리에서 SAP의 ESA 개념을 짧게나마 들을 기회 있었다. 그 분 말 요약하면 SOA 라는 개념이 엔터프라이즈 차원의 Business Context 에 그냥 적용하기에는 너무 일반적인 개념이어서 ESA 라는 신조어를 들고 나왔다는 건데 한마디로 ESA = SOA +
Business Enterprise-scale business solutions.
Technorati Tags: SOA, ESA