1

I am doing research regarding NoSql. I have found two DBs that are close to what I require for implementation.

just wondering how the file system of these Nosql interacts with Windows/Linux file system?

for instance, Cassandra uses Cassandra file system, so is it the case that it is a virtual file system that with help of JRE/JDK, it interacts with the platform file system?

Community
  • 1
  • 1
madiha malik
  • 827
  • 2
  • 8
  • 15

2 Answers2

1

Yes, Cassandra File System just sits on top of the platform file system. All of these databases (Cassandra, MongoDB, etc.) are going to use a typical underlying filesystem provided by the OS, e.g. ext3/ext4, XFS, NTFS.

helmy
  • 9,068
  • 3
  • 32
  • 31
1

The Cassandra File System is actually not required to use Cassandra. CFS only comes with DataStax Enterprise and isn't available in Apache Cassandra. It is not required or integrated in any way into normal Cassandra operations.

CFS was built as a HDFS compatible filesystem for use with batch analytics software. It is implemented as a set of C* tables but will eventually be represented as on disk data. Because it is a set of C* tables it will be distributed about your cluster and act as a distributed fault tolerant filesystem.

RussS
  • 16,476
  • 1
  • 34
  • 62
  • now thats something Interesting! Because on web, Cassandra is paired with Datastax, so it gave impression that this is how CFS deals! So apart from Datastax, who else are backing Cassandra? – madiha malik Nov 14 '14 at 05:33
  • And what if one is using Only Cassandra without Datastax, SO HDFS is the system I should opt for big files? or with Hadoop, any compatible file system can be used? – madiha malik Nov 14 '14 at 05:38
  • You don't need a distributed filesystem to use Cassandra but if you want to use pig or map reduce on your Cassandra data having HDFS or cfs is almost a necessity – RussS Nov 14 '14 at 06:13