I assumed you'd essentially be testing Netty against Netty, and you wouldn't be able to find much of a difference, despite the limitation of the OS TCP stack. But it seems even the TechEmpwer Round 22 results support your gut feeling.
Using Jetty vs Netty fell down to just where I copied the setup code from. Had I managed to get error-less results I planned to fix that, but as it’s the server that accepts the connections it turns out to have been handy that I used both, because I think it implicates the OS, not the server or library?
Love the Big Al’s shirt!
Thank you. I have a big bike ride planned this summer, so there will be more cycle attire!
@@PairingWithDuncan enjoy! And thank you for these videos. They are a great way to see another developer’s workflow and thought process.
Could share the GitRepo Link for this benchmarking snippets
Certainly - its github.com/dmcg/ktor-documentation
I assumed you'd essentially be testing Netty against Netty, and you wouldn't be able to find much of a difference, despite the limitation of the OS TCP stack. But it seems even the TechEmpwer Round 22 results support your gut feeling.
Using Jetty vs Netty fell down to just where I copied the setup code from. Had I managed to get error-less results I planned to fix that, but as it’s the server that accepts the connections it turns out to have been handy that I used both, because I think it implicates the OS, not the server or library?
Would be interesting to consider memory and CPU usage in addition?
Ooh, yes. So many variables!
There is even more, when you consider different servers and clients 😮