I'd like to begin this thread by saying I have friends at @forrester (the analyst firm, not the folks who look after trees). They by-and-large do excellent work. I hope that I still have friends there after I finish this thread.
Today they released a Forrester Wave (motto: "Gartner Magic Quadrants that won't get us sued!") highlighting the landscape of Functions as a Service and oh doggy do I have thoughts on it. reprints.forrester.com/#/assets/2/374…
The analysis fails to grasp the actual decision points around which FaaS provider companies use. Effectively nobody is looking at the feature comparison highlighted in the analysis because that's not how adoption of FaaS happens here in the real world.
Instead, someone wants to do a thing with their cloud provider. They can either spin up a virtual machine / instance to do it, or they can try this newfangled FaaS thing.
As time goes on and adoption grows, what becomes essential to the Serverless adoption in an environment is the event-driven architectural model; this definitionally requires deep integration with their provider.
To wit, I use Lambda@Edge instead of @cloudflare functions not because it's a better product (HA!) but because Lamba@Edge integrates with AWS natively.
As a result the @forrester Magic Quadrant ("HOLY HELL DO NOT CALL IT THAT!!!") in question is really showing the relative adoption of various cloud providers.
To wit: "Forrester was unable to reach customer references for (IBM|Oracle) for this evaluation." YOU DON'T SAY. Their own analysis demonstrates this: Oracle Functions are technically excellent, but see little traction.
To distill this report's actual market analysis down to reality: "customers pick a FaaS provider because that's where their existing cloud infrastructure lives." The rest is just window dressing.
And lastly, the description of Oracle as a "non-participating vendor" is the most poignant diss in the history of industry analysis. Well done.
