ALT-26 · Paper · 23 Oct 2026

Mirai: the botnet made of default passwords

Cameras and routers with factory passwords became a botnet that knocked major websites offline in 2016. A study of who built it, how it spread, and why its source code outlived it.

Timeline
  1. Aug 2016Mirai appears and grows quickly.
  2. Sep 2016Attacks on the KrebsOnSecurity site and the host OVH; the source code is published on 30 September.
  3. 21 Oct 2016Attacks on Dyn disrupt access to large websites.
  4. 2017Three young men plead guilty in the US to creating Mirai.
Peak infected devices
600k+Antonakakis et al., USENIX Security 2017

How it spread

No exploits were needed: a short list of default credentials was enough. Infected devices scanned for more victims and reported them to a loader, which installed the bot. Rebooting removed it — until the next scan reinfected the device minutes later.

Why it mattered

Once the source code was public, variants multiplied. Mirai turned weak defaults into a public safety issue and fed directly into laws that now ban universal default passwords on consumer devices.

Our reading

Analysis — Altumptic's interpretation of the sources below.

  • Defaults are policy: whatever ships is what most people run.
  • A device that cannot be updated is a liability with a long tail.
  • DNS providers are concentration points; depending on one is a design choice.

What it means for you

  1. Ship devices with unique per-device credentials and no open Telnet.
  2. Keep IoT on its own network segment, without inbound access from the internet.
  3. Use more than one DNS provider for anything that must stay reachable.

Sources

  1. Antonakakis et al. — Understanding the Mirai Botnet, USENIX Security 2017
  2. CISA — Heightened DDoS threat posed by Mirai and other botnets (TA16-288A)

Altumptic was not involved in this research. Figures are as published by the sources above.

Cite this entry

Altumptic Lab (2026). ALT-26 — Mirai: the botnet made of default passwords. 23 Oct 2026. altumptic.com/research/mirai-2016