by Allan Borra, MSCS
Sitting as Project Manager AND Enterprise Architect
In a series of blogs, I’m sharing my experiences, challenges and eureka moments related to my dual role as a project manager and an architect to a solution development project driven by Enterprise Architecture. I’ve managed software solutions development projects before and have varying levels of applying project management disciplines and I’m usually using the software development lifecycle (SDLC) method. So it was a fun and learning experience for me to alter my usual “success” recipe by using Enterprise Architecture (EA) discipline in the mix. This methodology is similar to what is called Solution Architecture as defined by Gartner1.
It is observable that locally there are varying degree of awareness, compliance and maturity of Enterprise Architecture work in organizations. I happen to be working for a pioneering and leading EA consulting company who engaged a solutions development project for a government agency. The government agency has no enterprise architecture or capability in place and the agency management has no (with some misconceptions even) to little knowledge about Enterprise Architecture.
IT groups in government agencies, especially the one we are currently engaged in, are very comfortable with traditional SDLC methodology. It fits nicely with the current government procurement law2. Nevertheless, we had a unique opportunity here to really push for an Enterprise Architecture-driven software development to complement the SDLC methodology (or any software development methodology for that matter), due to the fact that the terms of reference (TOR) explicitly calls out Business/Data/Application/Technology (BDAT) architecture requirements and standard notations such as Archimate.
Challenge 1: Changing Mindsets
Herein lies the first challenge: managing client expectations on project activities and deliverables. I’ve heard of an anecdote that goes “Change Management is easiest without people”. It was a really a concern for me back then that the client wanted the solution as soon as possible and that, even as the TOR calls out deliverables on BDAT architectures, these are not “primary” requirements for them. The client is accustomed to the SDLC or traditional methodologies that allow for only a certain level of business and functional requirements elicitation that is just enough for the software design and implementation. This meant that a semblance of the software solution could be out in a month or even as soon as codes are translated from the functional requirements.
|Figure 1. TOGAF Architecture Development Cycle|
Enterprise Architecture, using The Open Group Architecture Framework Architecture Development Methodology (TOGAF ADM)3, entails going through different phases as shown in Figure 1. Critical to the requirements of the Terms of Reference for the project are Phases B, C and D. The method allowed us to view all their process documentations, if available, or elicit and model processes that are undocumented. These process models and artefacts take part in the Business Architecture. We then looked into their data both from a high or conceptual level and from a low or physical level. It took great lengths of mapping these conceptual data models to the business processes, moreso getting the complete business processes, interactions and information flows documented themselves. Then we looked into their existing applications, modelled and related it to the processes to which these applications serve. We looked at their current infrastructure as well and modelled a target architecture by which the infrastructure can fulfill the requirements of the applications that services the processes that fulfills the intended performance of the solution that complies with the TOR. Needless to say, we considered legacy, third-party data and applications, out of scope processes in the mix of the overall architecture. Ultimately, all of these need time without outputting a single line of code yet. Three months spent on EA work without a line of code yet for only a year-long project. For sure the client was very impatient.
It was a painstaking “selling an idea, selling yourself” stint for me as I ventured in winning over the client and have them be on our side for this methodology: EA-driven solution development. What partly worked was communicating the value proposition of doing the EA work. For sure, there’s tons of references out there that list these: 1. Alignment of technology to business strategies; 2. Aids in achieving business strategies; 3. Managing complexity of the enterprise; 4. Faster design and development of solutions; etc.. At the start of every meeting, I give a few minutes to orient key client stakeholders on the EA framework and the value this provide. The orientations ran for more than a month until we had some considerable artefacts and architectures to show for. This worked in a way that when there are differences of expectations, it’s always a heavy, difficult conversation with clients, while, if there’s a level of agreement on expectations, conversations are smooth and affable. And we are having these kinds of conversations, already. But I did mention that communication only partly worked, because at the end of the day, there’s still no solution or line of code to show. I just can’t take all of that away from them, for now. And then, eureka! The architecture surfaced the complexity! When one can visually breakdown a complex network of interacting components and dimensions of the organization, or in this case, the solution and its interactions with the organization, managing all of that becomes simple. In my next blog, I’ll discuss how EA + tool shows the complexity and how it aids the communication: bringing the client to our side. I’ll share as well, how visualizing the complexity helped me as PM with decision-making based on these architecture artefacts. 1Gartner IT Glossary: Solution Architecture. Accessed June 2017. Retrieved from
Talk to Us.
Unit 903, Antel Global Corporate Center, Julia Vargas Ave, Ortigas Center, Pasig City, Philippines, 1605