MapReduce offers a robust and conceptually simple framework for implementing
index construction in a distributed environment. By providing a
semiautomatic method for splitting index construction into smaller tasks, it
can scale to almost arbitrarily large collections, given computer clusters ofsufficient size.