CacheBox is incredibly flexible and if you would like to create your own eviction policy, you can! Below are a set of easy steps on how to do this:
Create a simple CFC that implements the following class cachebox.system.cache.policies.IEvictionPolicy
or use our convenience abstract class and inherit from cachebox.system.cache.policies.AbstractEvictionPolicy
Create your own execute()
method that will evict items (We recommend looking at existing policies to get an insight on how to do this)
Use the policy instantiation path in your cachebox
provider properties
The below code is the code used to evict objects from cache
That's it folks! Very easily you can create your own eviction policies and use our built in indexers to just sort the elements in whatever way you like. If not, you can always do it yourself :)
The CacheBox Provider offers several eviction policies that are used to evict objects from cache when certain situations occur within the cache environment. Again, these eviction policies are only for the CacheBox Cache provider. These eviction situations can include:
Maximum objects reached
JVM memory threshold reached
Manual eviction executions
Below is a refresher on what an eviction policy means:
Eviction Policy: The algorithm that decides what element(s) will be evicted from a cache when full or a certain criteria has been met in the cache. (Cache Algorithms)
Please note that when a cache element is evicted from the cache it usually is expired first so it can be gracefully collected when the CacheBox reaping procedures occur. Therefore, you might see sometimes that the cache actually goes above the maximum objects defined, this is normal.
Caution Please note that no eternal objects are ever evicted from the CacheBox provider. Eternal objects live as long as the CacheBox instance lives for (most likely application scope timeout). So please take that into consideration.
ColdBox ships with the following eviction policies:
Policy
Description
LRU (Least Recently Used)
With this eviction policy, the cache discards the least recently used items first. This is the default policy
LFU (Least Frequently Used)
This policy counts how often an item has been accessed and it will discard the items that have been used the least.
FIFO (First In First Out)
Just like it sounds, first one in, first one out. Great for implementing sized queues
FIFO (First In First Out)
Just like it sounds, first one in, first one out. Great for implementing sized queues
LIFO (Last In First Out)
Just like it sounds, last one in, first one out. Great for implementing sized stacks or timed stacks