Is there a comprehensive comparison between Tyk vs Kong?

I have been developing microservices (Spring Cloud) for a while (~2 years) and heavily used Netflix Zuul. While it offers a lot of functionalities and great features, my developer mind wandered towards knowing about the alternatives and came to know about Tyk and Kong.

Reading from the individual documentation and blogs, I understood more or less both offer the similar features. I would like to know a comprehensive comparison between the two and any real-world examples where you have implemented will be a great help understand.

Solution

According to CI/CD both can comply with Infrastructure-as-Code approach, so i do not see difference in terms on Deployment Pipeline practices.

tyk API function-set is more compared to Kong, which may make sense if you rely your business on API(need to integrate with some Billing, ...) https://tyk.io/docs/tyk-rest-api/api-definition-objects/

On the other side, the API of Kong has limited functions and terminology IMHO is not understandable: https://galileo.gelato.io/docs/versions/2.0.0/

Kong uses Galileo reporting tool for DashBoard/UI, tyk uses its own DashBoard including not only Reporting functions, but also almost all Management Functions if you wanna go with the UI
If you need to transform your legacy APIs to external world, tyk has Transform function which can be used to transform XML<->JSON<->YAML<->Custom
On tyk you can code extension not only with Lua, but also with Go, Java. Python. .NET, Javascript ...
If you have DR needs, tyk has Multi-Datacenter option which is targeted for Enterprise level architecture including a Disaster Site
If you need performance tyk is written with Go. (We have benchmarked tyk to respond around 3000 req./sec. where Kong did around 2500 req./sec. on same VM with same APICall patterns)

So based on your needs, if any of your needs matches with one of the above, you can consider tyk, if not you can consider whichever you like more...