Kademlia: XOR Distance and the DHT That Stuck
Maymounkov and Mazières: k-buckets, XOR metric, and why BitTorrent and Ethereum-style discovery still look like this paper.
Tag
21 articles · 75 min of reading
Maymounkov and Mazières: k-buckets, XOR metric, and why BitTorrent and Ethereum-style discovery still look like this paper.
How Akamai's edge platform caches and accelerates origin fetches, and how SureRoute measures real paths so a download does not follow a congested BGP default.
How Zoom's Video Mesh and cloud media routing try to keep RTP local to an enterprise while still reaching the global meeting fabric.
Singh et al. SIGCOMM: hoverboards, flow programming, and why GCP's VPC can be software without looking like a software switch on your VM.
Cardwell et al. in ACM Queue: bottleneck bandwidth and RTT, why bufferbloat broke loss-based control, and what BBRv2 still argues about.
Langley et al. and the IETF QUIC RFCs: TLS 1.3 in the handshake, streams over UDP, and the middlebox tax HTTP/2 never escaped.
A practical comparison of WebSockets, Server-Sent Events, and WebTransport to help pick the right real-time transport for a given use case.
Why Spotify's early client-side streaming relied on peer-to-peer delivery, and why it later moved entirely to server and CDN-based delivery instead.
How Google's Jupiter network fabric used Clos topologies and centralized control to scale datacenter bandwidth by orders of magnitude.
How WhatsApp built low-latency, end-to-end encrypted voice and video calling on infrastructure originally designed for asynchronous store-and-forward messaging.
How Google's Maglev system moved network load balancing off specialized hardware and onto commodity servers without sacrificing speed or reliability.
A practical walkthrough of VPC networking concepts, from subnets and route tables to security groups, aimed at developers who deploy but don't design networks.
A practical comparison of gzip, Brotli, and zstd for HTTP response compression, covering ratio, CPU cost, and where each one actually wins.
Inside the public postmortem for Slack's January 2021 outage, where a networking degradation turned into a self-reinforcing overload across the stack.
A practical look at how HTTP/2 and HTTP/3 multiplexing affect application design, when it helps, and where old HTTP/1.1 habits still hurt you.
How Cloudflare uses anycast routing so a single IP address is announced from hundreds of data centers, with BGP steering each visitor to the nearest one.
CDNs now run edge compute, cache dynamic responses, and shield origin from thundering herds — far more than a cache for static files.
How Cloudflare replaced its NGINX-based proxy layer with Pingora, a Rust framework built for memory safety and lower resource use at edge scale.
Transport-layer and application-layer load balancers make routing decisions with very different visibility — here's what each one trades off.
Why Netflix stopped renting third-party CDN capacity and started shipping its own caching appliances directly into internet service providers.
Most backend engineers treat TLS as something a load balancer handles — until a service-to-service call or a certificate expiry proves otherwise.