There is talk about about re-layering the transport parts of HTTP: provide a common multiplexing, streaming protocol that works over TCP as well as QUIC.
Motivations are twofold:
- extending HTTP now requires double work or breaks at CDNs (they talk QUIC to clients, but not to backends)
- TCP will not go away and HTTP/2 is weak in regard to DoS handling. No one wants to put more effort into extending h2.
So, there might be a HTTP/5 coming on top of a tcp/quic multipleXing thingie (TUIX?).
Stefan Eissing /
npub179…nr96e
2024-11-22 08:43:49
Author Public Key
npub179eu5lkc7wklv8f44w6fz4qpc8gpkk2x8e32k7ha43pgvyj4jkqqqnr96ePublished at
2024-11-22 08:43:49Event JSON
{
"id": "d9571253607ab8f8943811f3f05306eb8ae544063b38cbdbe30d8f54ec6db8c3",
"pubkey": "f173ca7ed8f3adf61d35abb4915401c1d01b59463e62ab7afdac428612559580",
"created_at": 1732265029,
"kind": 1,
"tags": [
[
"proxy",
"https://chaos.social/users/icing/statuses/113525720998111865",
"activitypub"
]
],
"content": "There is talk about about re-layering the transport parts of HTTP: provide a common multiplexing, streaming protocol that works over TCP as well as QUIC.\n\nMotivations are twofold:\n- extending HTTP now requires double work or breaks at CDNs (they talk QUIC to clients, but not to backends)\n- TCP will not go away and HTTP/2 is weak in regard to DoS handling. No one wants to put more effort into extending h2.\n\nSo, there might be a HTTP/5 coming on top of a tcp/quic multipleXing thingie (TUIX?).",
"sig": "7a6860f50155a25de867f03b5d123f73301d7c23efe63dbce7ca287abab86bf03d0a1dc2ac83ce769d1d8d6701d61dbde48e913fd50dd95b4e875b1d195dc687"
}