Modern distributed storage systems recently adopted erasure coding to reduce
storage overheads. We show how network coding can surprisingly make the
maintenance of such distributed erasure coded systems more efficient by orders
of magnitude compared to standard Reed-Solomon codes. We survey the developing
theory and practice of regenerating codes. Both lower bounds via information
theoretic cut-set arguments and constructive techniques will be developed in a
tutorial way. We will discuss recent implementations over Apache Hadoop
Mapreduce and open research directions.