AI vs IA

Artificial Intelligence has become a mainstream topic and an industry with a market size reported to be in the billions. This term was first coined back in the mid 50’s by John McCarthy who was also the inventor of the Lisp programming language. At about that same time, a similar concept was proposed by William Ross Ashby in his “Introduction to Cybernetics” book. While Intelligence Amplification has not received the same amount of media attention as Artificial Intelligence, its subtle influence is still very much quietly present in modern times.

While related, the two terms are not identical. Both AI and IA use Machine Learning algorithms but what they do with ML is quite different. With AI, the machine does all the “learning” and the humans that consume AI are just supposed to trust what it predicts. With IA, the computer assists the humans in thinking for themselves. That is interesting to me.

I decided to explore what IA could do to help software engineers, specifically with regards to understanding how microservice architectures behave. I set up an experiment where I captured the performance data of 6 different implementations of a feature identical microservice as they ran under load in both Amazon’s and Google’s managed Kubernetes environments. I then processed that data with ML algorithms and studied the resulting trained models to see if I could gain any insights into these microservices.

With the help of ML, I was able to learn two things that I did not know before. I discovered a performance affecting bug in one implementation of the microservice and I uncovered a performance related secret about AWS.

In the course of this investigation, I evaluated the following open source technologies; Jupyter, Tensorflow, R, scikit-learn, Weka, KNIME, Mahout, and Spark MLlib.

If this is of interest to you, then feel free to click on the links below to read about this learning adventure in more detail.

EKS vs GKE

Analyzing MSA Performance with ML

Evaluating ML Open Source