Tuesday, August 18, 2026
Ryan Scherbarth (nvidia) joined the channel
David Ozog joined the channel
Sayan Ghosh joined the channel
Jiaqi Lou joined the channel
bsahu joined the channel
Fabian joined the channel
Francois Labonte joined the channel
Daud Arslan joined the channel
Shahbozbek Hakimov joined the channel
Tong Xu joined the channel
myoung-gyun.suh joined the channel
xi.chen joined the channel
Mahfuz joined the channel
Varun Chotalia joined the channel
pierre-louis.benard joined the channel
Matthew Fricke joined the channel
Advm joined the channel
ahmadrazarehman033 joined the channel
Howard Wang joined the channel
rashid-ahmed.kukkady joined the channel
Wednesday, August 19, 2026
A
Really interesting that ESUN is meant to keep the fabric common across transports. The MRC paper also describes a mix of mandatory, optional and implementation-specific mechanisms. Where is the actual compatibility contract between the transport, NIC and switch? Is there a machine-checkable multi-vendor negative-test suite for simultaneous congestion, packet loss and port failure, or is qualification still mainly end-to-end workload testing?
3 replies
M
ESUN defines the frame format, LLR, CBFC and optimized headers for transport protocols to follow. Once thr transport does that it can be easily forwarded via ESUN switch. Things like packet loss, port failure etc are handled at a layer above the ESUN layer, most likely in transport layer
A
Thanks Mohan, that helps draw the line. Is there a common multi-vendor test for the seam between ESUN and the transport, for example, when link-level retry can’t recover or a port drops while CBFC backpressure is active, or does each operator qualify that combination themselves?
Thursday, August 20, 2026
M
I think each operator have to qualify themselves as the transport protocol is specific to them.