Class TupleSerialization is an implementation of Hadoop's {@link Serialization} interface.
Typically developers will not use this implementation directly as it is automatically added to any relevant MapReduce jobs via the {@link org.apache.hadoop.conf.Configuration}.
By default, all primitive types are natively handled, and {@link org.apache.hadoop.io.BytesWritable}has a pre-configured serialization token since byte arrays are not handled natively by {@link Tuple}.
To add or manipulate Hadoop serializations or Cascading serializations tokens, see {@link TupleSerializationProps} for a fluent property builder class.
By default this Serialization interface registers the class {@link org.apache.hadoop.io.ByteWritable} astoken 127.