Live Layer Protocol: Streaming Signed, Verified WebAssembly Layers into Running Applications

Shipping a code change today means running it through a build-test-release-deploy pipeline and then restarting the program that runs it. The pipeline is where safety checks live, so it cannot simply be removed. We describe the Live Layer Protocol (LLP), which moves those checks to the moment a change arrives at a running application. Code is delivered as a stream of small, signed, content-addressed WebAssembly modules (layers). A Live Host owns all application state, verifies each signed manifest, rehearses the new layers against a copy-on-write shadow of the real state, commits all of them atomically or none, and reverts a layer that traps or exceeds its time or memory limit. The same manifest stream drives client devices and server or edge nodes, and a fleet layer adds channels, percentage rollouts with health gates, convergence tracking and key rotation. We report a reference host in JavaScript, an independent second host in Go written from the specification, and a conformance suite of 161 language-neutral checks that both pass. On one 4-core machine, a warm swap takes 2.3 ms (p50), a change reaches 1,000 simulated hosts 0.58 s after the file is saved, and a layer that crashes after going live is reverted with state restored in 2.3 ms. A red-team review produced 39 findings, of which 29 are fixed (several only partly or opt-in) and 10 remain open. We replay the failure mechanisms of four public outages in simulation and state where the protocol contains the failure and where it does not. LLP v0 is experimental: it catches crashes and hangs, not wrong answers, and several results come from single-machine simulation.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24349381
Primary Topic
Distributed systems and fault tolerance
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Live Layer Protocol: Streaming Signed, Verified WebAssembly Layers into Running Applications

Nikhil Ajay Kadalge
Zenodo (CERN European Organization for Nuclear Research)
Distributed systems and fault tolerance
preprint

Live Layer Protocol: Streaming Signed, Verified WebAssembly Layers into Running Applications

Nikhil Ajay Kadalge
preprint en

Abstract

Shipping a code change today means running it through a build-test-release-deploy pipeline and then restarting the program that runs it. The pipeline is where safety checks live, so it cannot simply be removed. We describe the Live Layer Protocol (LLP), which moves those checks to the moment a change arrives at a running application. Code is delivered as a stream of small, signed, content-addressed WebAssembly modules (layers). A Live Host owns all application state, verifies each signed manifest, rehearses the new layers against a copy-on-write shadow of the real state, commits all of them atomically or none, and reverts a layer that traps or exceeds its time or memory limit. The same manifest stream drives client devices and server or edge nodes, and a fleet layer adds channels, percentage rollouts with health gates, convergence tracking and key rotation. We report a reference host in JavaScript, an independent second host in Go written from the specification, and a conformance suite of 161 language-neutral checks that both pass. On one 4-core machine, a warm swap takes 2.3 ms (p50), a change reaches 1,000 simulated hosts 0.58 s after the file is saved, and a layer that crashes after going live is reverted with state restored in 2.3 ms. A red-team review produced 39 findings, of which 29 are fixed (several only partly or opt-in) and 10 remain open. We replay the failure mechanisms of four public outages in simulation and state where the protocol contains the failure and where it does not. LLP v0 is experimental: it catches crashes and hangs, not wrong answers, and several results come from single-machine simulation.

Zenodo (CERN European Organization for Nuclear Research)
Distributed systems and fault tolerance
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon