EA TOOL SELECTION & SUPPORT

“To do good work, one must first have good tools”  (Chinese proverb)

One of the first major milestones in EA program maturity and productivity is typically the selection and implementation an EA toolset

Most architects know to reach a desired level of productivity they should automate and integrate their internal modeling, analytical and EA program processes

The majority of programs begin by using a combination of Microsoft Office suite tools (Visio, Excel, PowerPoint, Word, Access, etc), as well as open source and low cost modeling tools, and other enterprise data sources such as CMDBs and databases to support their architecture efforts. 

Some of these are great ‘point-solutions’ for program inception, but unfortunately they don’t offer a comprehensive and integrated approach.  Most EA programs using them soon reach a point of diminishing returns especially as their programs grow in size and increasing amounts of architectural content is created.  Common problems include an inability to leverage industry accepted frameworks, and issues with keeping EA content up-to-date, version controlling and team collaboration, etc. 

Not surprisingly selecting the right tool can be daunting task. There are a myriad of EA and modeling tools on the market each with comparative strengths and weaknesses many of which are pointed out in various analysts’ reports and product ratings.  Like many other genres of software these differences can range from very subtle to very significant making an ‘apples to apples’ comparison exceedingly difficult.  Such uncertainty causes many programs to procrastinate in taking the next step.

Many programs find themselves facing a common dilemma:  The possibility of a prolongation of diminished capabilities and lost opportunities vs. the prospect of making the wrong tool choice that could become a costly and time consuming diversion.   

Some key considerations include:

Product attributes:  There are only a few products designed specifically for the purpose of supporting the EA process.  Most were originally written for drawing or process modeling, system development or as information repositories.   Many tools today are compilations of these capabilities.  Their features and functions vary as do their overall effectiveness for supporting program processes and information needs. 

Cost:   There is a wide range of prices in the EA tool market.  They can run in multiples of $1,000s, $10,000s $100,000s and even $1Ms. Although it is seldom wise to make a decision on price alone, it is certainly a factor; especially in the current economic climate where ‘belt-tightening’ is the order of the day for many public sector organizations and corporations.

Ease of use: Most program managers want to be sure there is a minimum learning curve and that their teams will become productive and self reliant in the shortest amount of time as possible.

Transition/reuse:  Most programs invest significant time and money in developing team skills, establishing program workflow and output, which is often, archived in a variety of data formats, so a burning issue for most is how all this can can be leveraged when moving to a new tool. 

The disruption factor:  Any selected tool should complement and facilitate the EA effort and should not become its focus or an obstacle to it which could result in misdirected efforts or cause a team to lose sight of program goals and objectives.  

Key tool attributes should include: 

Meta-model selection/availability:  should provide access to all leading industry frameworks

Flexibility:  should enable users to extend and customize the underlying meta-model

Output:  can produce standard views, reports, as well as other ‘business friendly’ artifacts such as graphs, charts, lists, matrices, dashboards and has facilities to publish web enabled content

Openness:  imports data from and can interface with a variety of outside sources (SQL, XML, CMDBs, etc); exports EA data to other tools.

Usability:  is easy to learn, deploy and use; enables team collaboration and role based access and security

Integration:  changes made to architecture components or attributes are propagated across models

Manageability:  facilitates application and repository management; supports information fidelity through version controlling, check-in/check-out, conflict resolution, etc

Facilitates architecture based decision support:  component level properties and metrics for cost and performance analytics

Enables the concept of As-is and To-be architectures:  can perform gap, impact and option analyses, etc.

Support:  vendor has financial stability, organizational depth and maturity and product development capabilities which stay abreast of the latest industry trends and developments; as well as the ability to provide training, product support and consulting resources either directly or through partner networks.