3 ms·

This seems to be a pretty ubiquitous view nowadays, but neglects the possibility that some projects might go better with an individual contributor just getting

• chillacy 10y ago
  An interesting analogy is the tradeoffs in going from one machine to a distributed system: a single SQL server will outperform a 3 node cluster for small loads, but the 3 node cluster will survive hardware failures and eventually out-scale the single-node setup as you add more nodes. It feels like larger companies are always reluctant to have a single point of failure, whether it's a machine in Utah going down or Greg getting hit by a bus.
  • dasmoth 10y ago
    It's a good analogy. Setting up the three-node cluster is an awful lot more work, and if you don't do it right there's a real chance that you'll end up with lower real-world reliability than the single node. That said, I can understand why large companies worry about single points of failure. What perplexes me more is that the you-always-need-a-team mindset is filtering down to smaller and smaller organisations.