Fragmented CRYPTO frames in payload?

[Ceyhun Atacan]

I've been seeing a lot of fragmented CRYPTO frames in INIT msg payloads, peppered with PADDING and PINGs recently from Google clients. I wonder what the reason  is behind this. Is it interop testing or something? Is it possible to generate such packets with Quiche? TIA!

[David Schinazi]

Hi,

Yes, the packets come from QUICHE. Their purpose is to ensure that QUIC parsers correctly implement the protocol instead of looking for fields at fixed offsets. That'll ensure that we can add extensions later without breaking existing parsers.

https://github.com/google/quiche/blob/main/quic/core/quic_chaos_protector.h

David

On Fri, Feb 11, 2022 at 8:42 AM Ceyhun Atacan <oxyb...@gmail.com> wrote:

I've been seeing a lot of fragmented CRYPTO frames in INIT msg payloads, peppered with PADDING and PINGs recently from Google clients. I wonder what the reason  is behind this. Is it interop testing or something? Is it possible to generate such packets with Quiche? TIA!

--
You received this message because you are subscribed to the Google Groups "QUIC Prototype Protocol Discussion group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to proto-quic+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/proto-quic/e466b509-3f7b-468d-9f46-db08fbbe56c5n%40chromium.org.