Re:Streaming compression beats framed compression
by lambdaloop on Tuesday December 30, 2025 @12:11AM
(#46429754)
Does streaming compression work if some packets are lost or arrive in a different order? Seems like the compression context may end up different on the encoding/decoding side.. or is that handled somehow?
[ Parent
| on HN ]
Re:Streaming compression beats framed compression
by dgoldstein0 on Tuesday December 30, 2025 @12:17AM
(#46429795)
I think the underlying protocol would have to guarantee in order delivery - either via tcp (for http1, 2, or spdy), or in http3, within a single stream.
Re:Streaming compression beats framed compression
by duskwuff on Tuesday December 30, 2025 @12:14AM
(#46429779)
It sounds as though the data is being transferred over HTTP, so packet loss/reordering is all handled by TCP.
Re:Streaming compression beats framed compression
by dgoldstein0 on Tuesday December 30, 2025 @02:33AM
(#46430541)
Yes, or by http3's in order guarantees on the individual streams (as http3 is udp)
Re:Streaming compression beats framed compression
by gkbrk on Tuesday December 30, 2025 @12:59AM
(#46430037)
WebSockets [1] run over TCP, and the messages are ordered.
There is RFC 9220 [2] that makes WebSockets go over QUIC (which is UDP-based). But that's still expected to expose a stream of bytes to the WebSocket, which still keeps the ordering guarantee.