IDs

These are unique values, primarily used as a way to distinguish different data (even if data in a row is the same)
This is used for primary keys, indexes etc
Issues crop up when creating Ids when
- a lot of unique ids per time interval ie 10000 per second
  - due to lots of writes to the systems
- Uniqueness constraints may not be possible through default database implementation
- Type of ids may be constrained due to disk space
  - ie can only have 64 bit, where as a common standard might be UUID (128bit)
- Ids may be constrained by ordering
  - ie may need to be ordered by date of creation
  - Issues with using time in distributed creation can be clock synchonisation
- Issues with distributed systems creating ids
  - synchonization problems, may end up with incorrect same id
  - availability of the system
- Size of id can limit total number of ids that can Generated
  - Thus need to do a reset or implement a new type of generation and thus migration of old ids

Creation

Last updated 2 years ago