Reliable Networks - impossible?
2 huge articles from Kyle about network partions and how it affects software vendors and online companies:
Jepsen: Testing the Partition Tolerance of PostgreSQL, Redis, MongoDB and Riak
Wow... a long list of outages due to network partitions from various companes like:
- Github
- Bonsai.io
- Amazon AWS
- CityCloud
- Fog Creek
- Freistil IT
- PagerDuty
- CloudFlare
.... and many more.
A very well researched article by Kyle Kingsbury, addressing the topic how hard it is to implement reliable distributed systems.