Batch process layer

A batch process, from the perspective of data access, requires sequential access of data and the storage layer should be optimized for this operation. Hadoop as a technology reads and writes data as blocks and each of these blocks contains data in a sequence. Even the block level access in Hadoop is in a sequence to make sure that any batch to disk operation is fast enough even on spinning disks (commodity hardware).