Home > Software engineering >  How can we implement Cassandra File System(CFS) in a Java Project
How can we implement Cassandra File System(CFS) in a Java Project

Time:02-21

I am fairly new to Cassandra. I was trying to chunk files to be stored in Cassandra DB and came across the term CFS. But I was unable to find any implementation of the same and some places even saw it as deprecated. The files expected around max of 100 MB. Could anyone help for he same.

CodePudding user response:

CFS was an attempt to store binary data in the Cassandra, and it's failed miserably. The better approach was DSEFS, but it stored only metadata in Cassandra, not actual data that were stored as binary blocks on the disks (similar to HDFS).

But you need to reconsider your decision. Cassandra isn't optimized to store big binary blobs, and although you can chunk them into smaller blocks, you'll get all sorts of the problems with repairs, bootstrap of the new nodes, etc.

The better approach would be to store in Cassandra only metadata (including location of where data is stored), but store actual files on something like, AWS S3, Azure Blob Storage, or if you're on-premise, on something like Minio.

  • Related