Cache Invalidation and Eviction

r- Data stored in cache is not there forever, it can become stale (not correct)

• The data needs to invalidated, evicted and replaced

• Cache invalidation, is the process where the data is updated or removed from the cache

TTL - Time To Live

• This is time value, which stored along with the cached data

• When the TTL expires it will either remove the data or update it

• Having a very low TTL, means that the cache is updated frequently, and you lose the performance benefit of a cache and possibly overload the system

• Have a large TTL, leads to stale data being used

• Finding the TTL, can be very hard, requires investigation and metrics

On update

• This updates the cache at the same time data source is updated

least recently used (LRU)

• invalidate the data which was least used

least frequently used (LFU)

• keeps the most frequently used data

First in first out

• treat as queue

Freshness caching verification

• where the application executes a lightweight procedure to determine if the data is still valid every time the data is retrieved. The downside of this alternative is that it produces some execution overhead

Active application invalidation

• where the application actively invalidates the data in the cache, normally when some state change is identified.