0

I am newbie in Hadoop.I am going through the Professional Hadoop Solutions book to gain some knowledge on Hadoop and ecosystem. I want to clarify what is main difference between HDFS and HBase. The way i understood is like both are storage System. They differ only in accessing the data.HBase access Data through NoSQL and HDFS process data with Computational Framework(MapReduce). If this is case why can't we can have only one storage either HDFS or HBase. Based in the requirement they will plugin and plug out the frameworks. My assumption might be wrong, It would be great if any one comment and provide correct relation between them.

Renukeswar
  • 567
  • 1
  • 6
  • 14
  • 1
    HBase is build on top of HDFS to access the data at runtime. HBase stores the data on HDFS only. You can understand more when you use it at practical problem. – Kumar Dec 03 '14 at 06:59
  • 2
    Possible duplicate of [Difference between HBase and Hadoop/HDFS](https://stackoverflow.com/questions/16929832/difference-between-hbase-and-hadoop-hdfs) – Suraj Rao Aug 28 '17 at 06:36

2 Answers2

2

HDFS

  • Hadoop distributed File System for storage.
  • Follows write-once read-many architecture.
  • Cannot use it for real time as map-reduce takes some times to complete.

HBase

  • Key-pair Column oriented NoSQL database.
  • Built on top of HDFS.
  • Storage can be your local file system or HDFS.
  • Can use for real time data access.
  • See HBase official page for HDFS s HBASE

Look at this Answer

Community
  • 1
  • 1
Kumar
  • 3,782
  • 4
  • 39
  • 87
0

HBase has properties like inserts and updates (puts) which HDFS cannot provide. HDFS simply supports appends.

There are many reasons for using either HDFS and HBase.

Krishna Kalyan
  • 1,672
  • 2
  • 20
  • 43
  • Are they both Storage systems. Why can't they keep only one in Hadoop eco system. I think these 2 are tightly coupled we can't take out any one of them from Ecosystem.as per our business requirement (if we have more writes /reads) we can choose any of them right.Say for example if i want to store some data files , where should i put those in hdfs or hbase(my requirement would be more writes) – Renukeswar Dec 03 '14 at 06:34