Reputation: 2654
Requirements are as follows :
To elaborate a little more...Data is generated by a software...One user can have many instance of this software and our system supports multiple users.
Our System >> Users >> their software (all relations are one to many - left to right)
Upvotes: 1
Views: 367
Reputation: 2497
For Data storage , use Hadoop HDFS (Hadoop File system)
If data is insert only (data structure not getting modified) , use Hive
Use Sqoop as a connector between HDFS and Hive.
If data needs to be search ( something like Search engine ) , use Solr
Lastly for maintaining log data, try Flume
Upvotes: 1
Reputation: 77505
Seriously, the only way to do real-time is to not store it at all beyond maybe writing it to a dated log file for recovery. Use stream processing for the analysis, not a database.
Upvotes: 1