Managing data returned from a data provider

By default, Decision Server Insights stores all of the enriched data that is returned by a data provider in a WebSphere® eXtreme Scale map. To manage the size of the stored data over time, you can evict data by using a time to live (TTL) algorithm or a least recently used (LRU) algorithm.

About this task

If you want to evict data from memory, you can use a TTL algorithm, an LRU algorithm, or both.

Important: Performance of the runtime environment might be impacted when eviction is frequent and the amount of data that is evicted is large. Nevertheless, systems that have a limited amount of memory can benefit from data eviction.

When you select an eviction policy, it applies to all of the maps that store the returned data.

  • Evicting data by using the timeToLive evictor
    The time to live evictor is a timer-based algorithm that evicts entries based on their last updated time. The time to live value must be greater than the maximum duration of the data providers cache.
  • Evicting data by using the LRU evictor
    The least recently used (LRU) evictor is triggered when the map reaches a specified size. To reduce the size of the map, the LRU evictor identifies the entries in the map that are least recently used and evicts them.
Parent topic: Administering Insight Server