Misunderstood EA seeks acceptance

Enterprise architecture (EA) is not new. It's not hip, it's not being hyped (much) and it's probably not funky either. What it is, apparently, is a means for an organisation to align business and IT; that oft mentioned but seldom achieved Holy Grail.

EA is also, ITWeb is told, a means to reduce risk, improve efficiencies, rectify errant processes and save the planet; oh, no, wait, that was virtualisation, wasn't it?

Jokes aside, EA promises a lot that many other practices, frameworks and methodologies have promised over the years. And a CIO would probably be forgiven, given current circumstances, for wondering why he or she should add another long-term, expensive and resource-intensive practice to the list of practices (like SOA) currently being investigated or implemented by the organisation.

According to EA vendors and fans, CIOs should care because EA can deliver on its promises, provided, as usual, that it's done right.

“The first point is that EA is not just about or within IT. In my opinion, and a growing cohort of others, enterprise architecture goes beyond IT issues such as applications, data and technology, to embrace business architecture. This includes elements of direct interest to the business, including markets, stakeholders, products, services, processes, channels, culture, contractual arrangements,” says Promis Solutions chief architect Graham McLeod.

The key, apparently, is in the definition of the two words. Says Dariel Solutions CEO Malcolm Rabson: “Surely it is better for any organisation to have an architecture that is strategically aligned within the organisation, than to just have an architecture that merely exists?”

Further, as Sybase SA's Practice Manager Rudi Leibbrandt notes: “Enterprise architecture (EA) addresses the fact that IT should be driven by business and not vice versa.”

Part of the challenge, comments EOH Consulting divisional director Hubert Wentzel, is that EA is misunderstood in the industry. “It is thought to be part of integration,” he says. “And yes, it plays a role, but only against the bigger plan. There are three key pillars: information architecture, technology infrastructure and processes. These are three very diverse areas, and only one of them is hardcore IT. The rest is information management, how you disseminate information, what is your single source of record, data warehousing, BI, all play to that. It's not only an IT view, business has to take ownership too.”

EA goes beyond IT issues.

Graham MacLeod, chief architect, Promis Solutions

Says Leibbrandt: “EA describes how a business is really made up - from IT assets to how people do things to roles and responsibilities, organisational set-up. It's more than just the IT piece. Most CIOs these days are hard pressed to start another initiative to manage just those aspects of the business. The value is pretty low.”

And that's the key: value. Says RealIRM head Stuart MacGregor: “EA is a process not a project. You need to understand what makes it business-appropriate and sustainable. And you need to run your EA business with a clearly defined value proposition and service offering.”

The other value that frequently raises questions around EA is cost. As Wentzel notes: “EA has traditionally been perceived as really only doable for large corporates with big pockets that can afford an architect and team to look after these systems. It's got a stigma as not being affordable for the rest. The take we have is that, especially in these times, CIOs can't afford not to. You don't need to go out and get a one million foot view or try to model the world. The intent of EA is really around being able to define a set of rules, principles and disciplines that govern the way an organisation does things.”

Leibbrandt concurs: “Although CIOs may be reluctant... to invest in EA, they should see this as a strategic initiative driven from a business perspective and not as just another IT project. No one has time for another IT project now, especially not the ones that won't impact the bottom line,” he says.

“It is also important to remember that while we are in the middle of an economic recession, the economic climate will turn and when it does, businesses should have the structures they need to go full steam ahead,” comments Rabson.

According to Mike Rogers, senior manager, Technology and Delivery at Accenture SA: “If there is one thing an economic downturn is good for, it is its action as a compelling force to encourage businesses to improve their operational efficiency to maintain profitability. While South Africa has been riding high on a prolonged consumer boom, the inevitable upwards cycle is coming to a close as interest rates rise and inflation bites down. As companies look to maintain performance in potentially tougher times, the enterprise architect comes to the fore as a key role player who can identify powerful new means to enable cost avoidances, while also contributing to earnings enhancement.”

Another issue is the old standards-based versus proprietary conundrum. Says MacGregor: “CIOs need to adopt open standards, like The Open Group Architecture Framework (TOGAF), so that they are not tied into a proprietary methodology.”

States MacLeod: “There are probably more than a dozen approaches and frameworks. The Zachman Framework and TOGAF are the dominant ones. Zachman sets a good agenda in terms of what companies should be concerned with in EA; it doesn't always tell you how to go about it, though. TOGAF gives very good guidance in terms of the process of architecture and the methods and techniques but, today, is still a little bit unclear on some of the deliverables and the meta-models for organising and integrating them. Although that has improved somewhat with TOGAF 9, which has just been released. Some commercial frameworks, such as Integrated Architecture Framework (IAF) from CapGemini, and the Inspired Enterprise Architecture Framework (Promis Solutions) - also provide integrated meta models to assist in the integration of all aspects of the architecture.”

Another point to note, of course, is not to get stuck in analysis paralysis, trying to determine which methodology is best. Rather pick one and go with it, Rogers says.

MacGregor also notes that if the EA initiative is not close to the board or the strategic planning manager, it risks being tied to solution delivery and not being true EA.

Rogers agrees: “[It is important to] obtain sponsorship and support for the exercise. It is an exercise that requires the really close involvement of business, applications people, data architects, infrastructure and technology teams, and if they aren't onboard and willing engage and invest, the exercise will almost certainly fail through that lack of involvement.”

You should invest in EA, says MacLeod, “if you'd like to influence the future versus having it happen to you. EA lets you see how the context is changing, how the organisation needs to adapt to that future and then to plan the changes that you need to make in all dimensions to get to that desired future. It also helps organisations to respond to things in the environment rapidly and sensibly. For example, when you have a fire alarm, it's too late to start drawing the floor plan.”

Having a floor plan in place in advance of the economic turnaround, which, as Rabson notes, will come, will see the organisation concerned sitting pretty. EA may just be a good route to that point. Or not.

SDT embarked on an EA project, and was surprised by the results.

SDT develops life assurance administrator software. It embarked on an EA project around 18 months ago, with a view to formalising and codifying its products, creating some standard for software in the sector, and in a bid to bring its clients in line with the global Insurance Business Architecture standard. What it got, on the other hand...

EA is a process, not a project.

Stuart MacGregor, head, RealIRM

SDT executive director Freda du Toit says the company found it was having problems satisfying customers. “It was a misalignment in terms of what clients wanted and what we delivered. We embarked on the EA project because we thought we had all the business knowledge and could put something together as a model for the industry. What we discovered was that we would learn more about our own organisation in the process.”

SDT realised that the processes it used to engage customers were wrong in that it didn't guide customers through the process their business follows when they evaluate SDT's software, Du Toit says. “So we started an EA initiative, changed all of our internal processes, evaluated, implemented. Two major benefits have come out, one of which we never envisaged.

“On an SDT level we are suddenly able to be more efficient in our process in aligning what the customer gets and what we do. Product quality is better, customer satisfaction is higher and the cost of delivering it is lower because we're aligned from the beginning. Also, because we follow a life cycle approach, customers can understand very quickly what they are buying.

“We've created a framework in which IT can talk to the business so we both understand the cost of a decision and the return of a decision, if made. In the past, total cost was never seen, and we didn't understand the implications of small changes.

“This started as a pet project of mine because I thought it was interesting and related to business benefit, but actually we've created so much insight into how we operate that the level of benefit has far exceeded what we expected,” she concludes.