Distributed cache is a facility provided by MapReduce Framework. It is provided to cache files (text, archives etc.) at the time of execution of the job. The Framework copies the necessary files to the slave node before the execution of any task at that node.

In Hadoop, the Distributed Cache is a mechanism that allows you to cache files (like jars, zips, and other files) needed by the MapReduce job across all nodes in the cluster. The purpose of the Distributed Cache is to make these files available to all the nodes in a Hadoop cluster during the execution of a MapReduce job.

When you submit a MapReduce job, you may have additional files (such as libraries, configuration files, or even small datasets) that are required by the job. Instead of manually copying these files to each node in the cluster, the Distributed Cache automatically distributes these files to all the nodes before the MapReduce job starts.

This mechanism helps in improving the performance of the job by reducing the data transfer time, as the required files are already available on each node. It is especially useful for distributing read-only data or files that are used by the mappers or reducers.

In a Hadoop MapReduce program, you can use the Distributed Cache by adding files to it before the job starts. These files are then shipped to the task nodes and can be accessed by your map and reduce tasks during their execution. The API provides methods for adding files to the Distributed Cache and accessing them in your MapReduce code.