Files
zeekay abf99d05fc papers: reorganize into per-paper dirs, add Antje Worring as co-author
Restructured from flat .tex/.pdf files into per-paper subdirectories.
Added Antje Worring as co-author on all papers.
Synced PDFs to pdfs/ for download site.
2026-04-18 21:50:07 -07:00

16 lines
1.6 KiB
TeX

\section{Related Work}
%% ============================================================================
\textbf{Protobuf and gRPC} \cite{grpc}: Google's RPC framework provides excellent schema evolution, cross-language support, and HTTP/2 multiplexing. The allocation overhead is well-understood \cite{protobuf} and acceptable for most applications. For consensus transport, the p99.9 tail latency is the limiting factor.
\textbf{Cap'n Proto} \cite{capnp}: Kenton Varda's zero-copy serialization system pioneered the approach of reading data directly from wire buffers. ZAP simplifies the memory model (no segments, no far pointers) at the cost of schema evolution.
\textbf{FlatBuffers} \cite{flatbuf}: Google's zero-copy alternative to protobuf. The vtable design enables schema evolution while maintaining zero-copy reads. ZAP trades the vtable for fixed-offset access, gaining 10\% encode and 10\% decode throughput.
\textbf{Protobuf-mux} \cite{Protobuf-mux}: The IPFS/Ethereum P2P networking stack. Protobuf-mux provides peer discovery, NAT traversal, protocol multiplexing, and pubsub---features ZAP does not implement at the transport level. ZAP is a lower-level primitive; Protobuf-mux operates at a higher abstraction.
\textbf{QUIC} \cite{quic}: Google's UDP-based transport protocol (RFC 9000). QUIC's 0-RTT handshake and multiplexed streams are advantageous for web traffic. For consensus transport between pre-authenticated nodes, the TLS handshake overhead is amortized over long-lived connections, making TCP with ZAP competitive.
%% ============================================================================