As an enterprise architect, I support my customers with concepts concerning solutions, platforms, and dataflows and process visualization. That means I’m always working with an enormous degree of complexity. Over the years, I have worked with a lot with standard EA tools such as Orbus iServer, Enterprise Architect from SparxSystems, or Iteraplan. These solutions helped me to cope with high levels of complexity. The tools worked quite well for a long time, but due to the increasing demand for agility in today’s business and the IT environment, I realized that my EA software requirements had changed accordingly. I want to share the challenges I have faced and the conclusions I have drawn.
Away from large landscapes – Towards small incremental steps
The principle of agile thinking is to turn away from the waterfall model and turn to incremental steps. Of course, the Big Picture remained a necessary part of complex software landscapes for me. However, over time the number of smaller use cases began to increase. The challenge was then for me was to break down the complexity of the tools used to deal with this situation more efficiently.
Increased requirements for collaboration
Another component of agile work is regular communication and the ability to collect feedback as you go. The fragmentation of use cases naturally increased my communication needs accordingly. That’s when I started looking for a solution to manage collaboration (synchronous and asynchronous) in a more targeted way, to reduce the time invested.
As a software architect, you tend to work across different departments. With the move towards the cloud, the trend towards inter-department collaboration has increased. The need for documentation and reporting is therefore essential for the ever-increasing number of people. I came to realize that three further points became increasingly important to me: simplicity, availability, and relevancy.
Approaching the solution
I took a closer look at those issues and compared them with my standard situation. Documentation and reporting turned out to be my biggest challenge, both in terms of time and resources. The creation of the reports and diagrams were, until now, quite time-consuming. I had to export the diagrams, marry them to the essential information, and then put everything into a final Word document.
This process was an ineffective way of dealing with diagrams, which then gave me an idea. The company I worked for at the time already had Atlassian Confluence and Jira in use. Confluence was used extensively, but not for documenting IT solutions or processes. The only remaining challenge was to integrate the diagrams into Confluence. The embedding of the existing data via iFrame wasn’t ideal due to long loading times, authorization queries, and minimal user-friendliness. I then set out to find an embedding solution and its relevant plug-in, for Confluence. I was successful: enter draw.io.
draw.io for Confluence – advantages and trade-offs
Until now, I have only talked about the things that have bothered or challenged me in terms of classic EA tools. The decision to go for a lightweight solution, draw.io, of course, had a trade-off, which I don’t want to hide. Traditional enterprise architecture tools always have a repository connected to them. All components of the IT landscape can be easily provided with attributes and object relationships in the subsequent steps, and the processes can be handled holistically. However, I had to accept the limitation of not having this by default, and I am happy to explain this.
Simple, direct and fast
After editing the draw.io diagram in Confluence, a .png file of the diagram is created and displayed in the viewer. I can always view, zoom in, and discuss this status with my stakeholders. I had rarely seen such a simple, fast, and direct access before.
Numerous visualization libraries for Solution Architecture
The topic of the hour is the cloud. This means I’m now almost exclusively on the road for my customers providing them with cloud solutions. draw.io offers comprehensive libraries with shapes for AWS, GCP, or Azure, to name just a few. The libraries not only make the creation of my diagrams faster but also make them visually more appealing and easier to understand.
Access to the latest version at any time
The visualization is now integrated as a macro on the Confluence page, and this means that my stakeholders and I have access to the current status of the information. We can provide online comments and feedback directly on the page. There are no more media breaks – the data can be consumed comfortably in Confluence, and the integrated versioning still gives me the option to view older versions and restore them if necessary.
Transfer to EA Tool, if necessary
Of course, I don’t completely do without the already mentioned, classic EA tools. If I need to provide my solutions with architectural attributes, draw.io diagrams can be incorporated and further developed with minor tweaks.
The bottom line is that as an Enterprise and Solution Architect, I have never regretted adopting draw.io for Confluence and Jira, respectively. The advantages I gained from the change in agility and timeliness far outweigh the loss of the analysis functions. It doesn’t always have to be a significant and comprehensive EA solution; give it a try.