Session

We are going to walk Linux Networking Humans through how to build Compassionate Networks that work for AI use cases, using Human Networks as examples to draw parallels. Our observation is that Linux Networking in general, with switches in between, is not very compassionate: it floods the network with packets, without any network resource reservations. Based on our experiments we believe there are fundamental flaws and semantic errors in the way we think about communication. Network packet drops, it turns out, are the mother of TAR (Timeout And Retry) and cause timeout storms, reconstruction storms, and metastable datacenters. So we go back to the fundamentals and ask: are packet drops necessary? As we scale the datacenter to millions of nodes, it turns out they are not – they cause congestion, transaction loss, and coherence problems in distributed systems and AI/ML infrastructures. Compassionate networks are the opposite of imposition networks, and most of our networks are imposition networks as of now. The opposite of an imposition network is a Promise Network.

The technical substance: Modern congestion control treats the network as a substrate onto which packets are imposed, and from which loss, latency, and Explicit Congestion Notification (ECN) marks must be reactively interpreted. We argue for an alternative approach that was visible in the 1976 Ethernet paper but never pushed from the transport layer down to the link layer. Metcalfe and Boggs’s end-dally in the EFTP protocol is the bilateral closure of a transmission. Open Æthernet (OAE) generalises it into a link-layer admission primitive in which every frame is gated on a peer-issued token, so a sender without one cannot transmit. The consequence is structural rather than statistical: uninvited frames never enter the fabric. Congestion control then becomes an admission discipline within the network, rather than a feedback loop above it. We frame this as a Promise Theory problem and sketch the Linux kernel surface it would require. This is a moonshot proposal, presented for community critique.