Which of these best represents the near future of Enterprise Integration Technology?
I see 3 approaches to EI:
1. Piecemeal -- oldest / most widespread, point-to-point integration from scratch, customized vendor product, or with off-the-shelf.
+ No-long term investments, solutions small / targeted.
- Little reusability, extremely brittle to changes on either end, multiple systems difficult. High level of integration competence.
2. Middle-tier -- more recent next-most widespread, single product ( Java Application Server, .NET), all backend systems integrated indirectly via middle-tier apps.
+ Longer-term investment. Standardized approach to integration. most common back-end systems are easily fit in. Greater reusability, resistant to changes.
- Products are available but not equal. Many will be costly. High degree of middle-tier development and administration skills.
3. SOA -- newest, ESB product (AquaLogic, WebSphere ESB), back-end systems integrated indirectly via the bus. Business apps become processes which use services from connected systems.
+ Long-term investment in technology, planning, and business paradigm (for better or worse), which MUST remain viable for a good amount of time. Somewhat standardized approach some if not most back-ends fit with less effort. High reusability is strongly supported. Most agile to meet changing business conditions and goals.
- Widespread use of SOA will require changes across the enterprise as a major business initiative. High overhead for preliminary solutions. A very high level of skill in major areas of Business / IT planning and development. Licensing fees.
Good Answers (2)
Russell T.
PhD Student at George Mason University (Computational Social Science)
Best Answers in: Business Analytics (4), Risk Management (3), Change Management (3), Supply Chain Management (2), Enterprise Software (2), Economics (1), Staffing and Recruiting (1), Corporate Governance (1), Organizational Development (1), Distribution (1), Market Research and Definition (1), Interface Design (1), Product Design (1), Information Security (1)
I think in most large companies, and perhaps mid-sized firms, the gating item is corporate culture and capacity to manage change.
The ROI on SOA is not so clear and compelling that it motivates no-brainer investments (i.e. it's not going to deliver bottom-line cost savings in the current fiscal year). It's a whole new way of managing IT, and managing how the enterprise engages with IT.
So far, the IT vendors and consultants haven't done such a great job to make SOA-driven transformation more digestable.
Putting this all together, I'd bet that SOA will continue to lag the vision.
In contrast, Middle-tier approach mostly affects the IT budget and organization only. It will continue to be the prefered choice for budget conscious IT managers who have relatively little clout in the organization.
Christopher G.
Director, Service Portfolio Management
Best Answers in: Government Policy (4), Enterprise Software (3), Government Services (2), Organizational Development (2), Nonprofit Management (2), Career Management (2), Customer Service (1), Property Law (1), Product Design (1), Starting Up (1), Blogging (1), Computer Networking (1), Telecommunications (1), Software Development (1)
Dan,
I noted that you've used a typical SOA roadmap to articulate the EI Tiers which highlights more of a maturity model than distinct EI Tiers/Approaches. I thought it was important to highlight my interpretation of your question in order to articulate the crux of the question; What EI Tier best represents the future of Enterprise Integration Architecture?
I'd suggest that different Tiers and the maturity of Tiers will provide different TCO and economies depending on the organisations own size, processes and architectural requirements. Each EI Tier as you rightly pointed out has their advantages and disadvantages and different organisations will benefit from different Tiers and there isn't a one size fits all when it comes to EI and its best approached as a Maturity Roadmap for an organisation rather than an end state requirement.
Whilst I am a great believer in SOA, I understand the pain and cultural shifts in EA methodology and governance required to take an organisation to the current end point of SOA, where they will actually reap the benefits (SOA is also a maturing paradigm even though its a decade old). Knowing that, I know its not for everyone, though some of its foundations provide some good opportunities for organisations to employ some simple governance that may not lead to complete SOA, but better EA architecture and procurement strategies overall.
I think we will continue to see for quite some time different EI Tiers that will be adopted by different organisations based on their ICT maturity, skill accessibility and strategic significance to the organisation. Each EI Tier will have varied price points and length of maturity, and vendors will continue to push variations and vendor specific snapshots and applications based on their interpretation of their clients needs (though not sure whether this is mostly the case). I think COTS will still exist that have a mixture of your Middle Tier and SOA methodology. AJAX will make SOA more accessible but overall the EAI mix will not necessarily narrow. It makes people like Oracle who are focusing on multi vendor middleware Adaptors probably the most profitable.
Regards,
Chris
More Answers (1)
Sumana H. suggests this expert on this topic:
Rachel at the 451 Group analyzes these sorts of things for a living.
