Have you ever worked on a project that didn't turn out like you hoped? Maybe it took longer than planned, cost more money, or didn't have everything it was supposed to have. This happens a lot, and it's often because the architecture wasn't designed well.
Architecture is like the blueprint for a system. Just like you need a good blueprint to build a sturdy house, you need a good architecture to build a system that works well. But it can be really hard to tell if an architecture is any good and continues to get harder as systems become more complex.
Currently, experts just look at an architecture and give their opinion on whether they think it is good. But that's not very scientific or reliable. Luckily, a smart person came up with a new way to measure architectures called QSLS (Quantifying System Levels of Support).
Here's how QSLS works:
1. The architect looks at all the pieces of the architecture (called mechanisms) and figures out how they relate to each other. Sometimes they have standards (which can be viewed as a specific collection of mechanisms)
2. QSLS uses AI to analyze how well the mechanisms support the important characteristics, quality attributes, and business goals of the system.
3. It turns these relationships into numbers and does some math to calculate "support levels."
4. The support levels tell us how good the architecture is at meeting the system's needs. Low support levels mean the architecture needs to be improved in certain areas.
QSLS lets us measure architectures in a systematic way instead of just guessing. Because we have measurements, we can compare different architectures to see which one is best. We can also pinpoint specific problems in an architecture and fix them. Overall, QSLS helps us make smart choices when designing a system.
Along with measuring the architecture, QSLS also generates a measurement for RISK. RISK is key for program management to understand where problems need to be addressed and how to focus resources to address them. Before QSLS, programs identify risk based on engineers feel based on their history. As systems grow in complexity, actual computed risk numbers become key to being able to manage complex programs.
Recorded Demo
True Case Study
A retired DOE large project manager ($B programs) named Bob understands how difficult it is to develop actual measurements of levels of support and levels of risk in developing systems. He's seen firsthand how many projects fail because they rely on opinion instead of facts. He even tried using AI to analyze some pretend project data and got some interesting results. He thinks there's a big need for tools like QSLS.
“I am finding that PMs today provide qualitative status updates, not quantitative, and this
leads to poor performance on projects. Most projects are delivered late, over cost, or scope is reduced, or some combination of all the above is experienced by the owner/client.
I entered some quantitative 'dummy project data' into ChatGPT and received misleading replies as it relates to project status.”
Bob is now working with QSLS Engineering to use our measurement outputs to provided information needed by Program Managers.
Conclusion
In conclusion, QSLS is an exciting new approach that uses AI and math to measure how good an architecture is. It could help us build better systems that do what they're supposed to do, on time and on budget. As systems get more complex, we need smarter tools like QSLS to design them well. The future of architecture is looking bright!
Comentários