How does the trivial method of distributing data work
- Assume that there are n cache servers
- Calculate the
serverIndex using the formula: hash(key) % N
Problems faced by the trivial method of distributing data
- This method works fine when the server pool is fixed, and the data distribution is even.
- Causes problems when servers are added/removed. The result of the modular operation will change, giving a different serverIndex
- Most keys have to be redistributed when a server is added or removed, not just the affected/added server.
- This can cause a huge number cache misses.
Sources