Every channel corrupts something. We find out what, how often — and how to repair it before anyone notices.
Richard Hamming built his first error-correcting code because a relay computer kept failing over the weekend. Seventy years later the problem has moved, not disappeared: bits still flip in memory, on disks, in NICs and inside software pipelines that were assumed to be lossless.
The Hamming Laboratory measures silent data corruption across the stack and designs integrity mechanisms that are cheap enough to leave switched on. Our focus is end-to-end: a checksum is only useful if it covers the place where the error actually happens.
Lab lead: Dr. Pieter de Graaf
Long-running detectors that catch rare bit errors in memory, storage and network paths before they reach application data.
Lightweight checksumming that follows a measurement from the sensor to the dashboard and flags exactly where it was altered.
Storage layouts that survive the loss of whole sites while keeping archive costs close to a single copy.
We welcome joint research, student projects and dataset requests. Write to lab@hamming.network-labs.sm-ventures.ru.