As someone who used to maintain a FOSS web server project for a few years, I've always lamented the fact that HTTP/2 did not displace HTTP/1.1 more thoroughly.
I've heard multiple variants of the argument that "HTTP/2's improvements don't warrant the extra complexity for internal networks", but the counterargument is inefficiencies like this one, caused by HTTP/1's connection and request/response models. (H1 pipelining is another problem.)
HTTP/2 is just a much more logical and efficient model. (Though it had its serious missteps too, like server push or the original priority mechanisms.)
As well as - last time I checked which to be fair was some years ago - being slower and heavier across (real life) implementations for short lived connections. HTTP/2 is not an "everything is better" solution, even if it is marketed like that. There are also a lot of biased benchmarks out there (comparing older and newer software versions or very far from real life scenarios, like the famous mosaic image). Hence I ended up benchmarking for my own use cases.
That said I really like HTTP/2 and HTTP/3 for their severe benefits in many areas. It just baffles me how many people have very wrong impressions of making things better by switching something to HTTP/2 without even confirming improvements. It's even worse when people end up switching to HTTP/2 but still establishing new TCP connections all the time. Have seen that way too many times.
[somewhat off topic from here on]
That's a general pet peeve though. People either not comparing new approaches to the old ones or doing it with huge obvious biases. Classic examples are "rewrite in new language", "containerize everything", "switch to Kubernetes", etc. claiming benefits that stem from just cleaning up code or infrastructure or simply upgrading to the latest release or redesigning based on past experience. I've been consulting for Kubernetes in fintech in the past, so I'm not that blind hater, actually benefited from the hype. But at the same time it's always worrying when people whom I consider smart become kind of religious about software.
kingmob | 15 hours ago
As someone who used to maintain a FOSS web server project for a few years, I've always lamented the fact that HTTP/2 did not displace HTTP/1.1 more thoroughly.
I've heard multiple variants of the argument that "HTTP/2's improvements don't warrant the extra complexity for internal networks", but the counterargument is inefficiencies like this one, caused by HTTP/1's connection and request/response models. (H1 pipelining is another problem.)
HTTP/2 is just a much more logical and efficient model. (Though it had its serious missteps too, like server push or the original priority mechanisms.)
reezer | 15 hours ago
As well as - last time I checked which to be fair was some years ago - being slower and heavier across (real life) implementations for short lived connections. HTTP/2 is not an "everything is better" solution, even if it is marketed like that. There are also a lot of biased benchmarks out there (comparing older and newer software versions or very far from real life scenarios, like the famous mosaic image). Hence I ended up benchmarking for my own use cases.
That said I really like HTTP/2 and HTTP/3 for their severe benefits in many areas. It just baffles me how many people have very wrong impressions of making things better by switching something to HTTP/2 without even confirming improvements. It's even worse when people end up switching to HTTP/2 but still establishing new TCP connections all the time. Have seen that way too many times.
[somewhat off topic from here on] That's a general pet peeve though. People either not comparing new approaches to the old ones or doing it with huge obvious biases. Classic examples are "rewrite in new language", "containerize everything", "switch to Kubernetes", etc. claiming benefits that stem from just cleaning up code or infrastructure or simply upgrading to the latest release or redesigning based on past experience. I've been consulting for Kubernetes in fintech in the past, so I'm not that blind hater, actually benefited from the hype. But at the same time it's always worrying when people whom I consider smart become kind of religious about software.
ur5us | an hour ago
Can you provide more details? I’m curious now.