I have been involved in the world of enterprise architecture for many years. As a practitioner, tech builder, sort leader, and expert.
Whether it’s a traditional enterprise, cloud, edge, or other special purpose system, what I like about the architecture is about 100 factorials to solve the same IT architecture problem. There is a way. Creating an optimized architecture for the least amount of money and risk actually relies on the experience of the architect rather than some routine methodologies and automated processes.
There are many things that are at stake. Under-optimized and costly architectures can certainly work, but while most people aren’t wise, they can cause businesses to lose millions a week. 30 technologies are used and 12 technologies work better. If you do not design for change, your business will be less agile.
But it works, doesn’t it? Now, let’s file for bankruptcy. That way, bad architects can go away and ruin other businesses.
The better you are at the architecture of new IT concepts such as cloud computing, the better you will understand the number of best practices. I’m often asked the question, “Can I automate the architecture (in this case, the cloud architecture)?” Is a good architecture reproducible? This means that you can define your architecture as a binary concept where you can repeatedly apply logic rules for optimized and consistent results.
About 15 years ago, I tried to break down SOA (Service Oriented Architecture), a component of cloud computing, into algorithms. I failed.
We were able to automate some basic decisions, such as choosing the size of the CPU and memory, but the idea couldn’t get going with more complex decisions. Databases, middleware, service design, security technologies-too many moving parts and too many dependencies to consider. Needless to say, we take advantage of the ever-changing current technology and the products and services of companies that may be closed or acquired.
Of course, thanks to cloud computing, affordable AI is available today. However, the same basic problem exists in that the variables are modified due to the excessive complexity that is part of these architectures. In addition, the training data used in this “AI cloud architecture” only defines past best practices and has little or no relevance to the way we work today.
What I mean is that cloud architecture is still the result of architect skills and knowledge. Tools such as checklists and predefined processes can help, but it depends on the talent of those who consider the past and future and probably rely on their instincts to make core technology decisions. .. We welcome it to be binary someday, but I’m sorry, it’s still analog for now.
Copyright © 2021 IDG Communications, Inc.
Is the cloud architecture binary or analog?
Source link Is the cloud architecture binary or analog?