The storage architecture for HDFS system in Hadoop


The storage architecture for HDFS system in Hadoop

Tradition in HDFS system

Hadoop is always takes as large scale business with trendy processing scale and also with the branching applications. Sequential throughout plans are shown latency but in a small manner and also the already existing models of HDFS are attached with some disk of special kind that are spinning disks.

Hadoop data architect give enquiry in more interactive way and this is the reason that either application benefits or happened perhaps with lower and storage media of solid states disks and latency.

Performance metrics which must be classified for the storage events and the three classes according to which they are classified are as:


Data is secured only in the conditions when there is great durability provided in your storage device and also it is not only enough but the successful writing of data is one of the beneficial tasks for the medium.

Disk failure is not a new problem it happens eventually but driving the new disk and collect data show that how durable is the device is.

Certain statistical calculations annually rate for the disks between 1 – 2 years and there is a risk only for about 2 – 8 percent. This is studied with the study of failure disk.

Performance of Hadoop data architect

Storage performance of device depends upon the two key measures of device of Hadoop data architect and these are:

  • operations of IO in every single second

Affected work load is supported by some readable IO operations and they must be connected with the rotational latency with size comparable to IO and also with work load. Limit in the performance shows with the interactive enquiry applications.

  • throughout

Maximum raw writable and readable data is on the rate of technology that their storage can be supported with some typical measuring quantity that is in Mega Bytes per second or (MBpS).

The last but most considerable factor is to measure the cost for every single mega byte.