As it says, it is meant for backup and data could be compressed if required and needed only for emergency.
~100 GB, should be expandable on demand.
Able to mount it as a remote drive.
High read IO is not required.
Solution: Northwestern vault. But we need to buy more space.
Alternate: Amazon S3 or similar cloud based solutions.
Web accessible data storage
The data should always be available and will be constantly accessed the web applications.
~100 GB or more.
Able to mount it as a remote drive, any network protocol(CIFS/NFS etc) will do.
Good read IO is needed.
Should reside in the same network as the web application, less network overhead.
Solution
Currently we have NUBIC windows server mounted as shared drive with CIFS. It is working reasonably well for GBrowse. The space is still no concern.
Concern
For multiple user access, a single shared network server might be a bottleneck. Moreover, random disk access for bam files is relatively slower than sequential access.
Alternate
Cloud disks, however then the web application has to reside on same network.
Shared network disks through glusterfs or lustrefs.
Analytical storage
The data will be used by analytical tools and will also generate temporary data files as needed.
~150GB or more.
At least access though ssh.
Should be tied to a computing cluster.
Good read/write IO.
Solution: If we use quest for computing cluster, then have to buy additional vault storage that is tied to northwestern quest.
Alternate: Any cloud computing(amazon/google/rackspace etc).
Here are our needs
Data storage and backup
As it says, it is meant for backup and data could be compressed if required and needed only for emergency.
Solution: Northwestern vault. But we need to buy more space. Alternate: Amazon S3 or similar cloud based solutions.
Web accessible data storage
The data should always be available and will be constantly accessed the web applications.
Solution
Currently we have NUBIC windows server mounted as shared drive with CIFS. It is working reasonably well for GBrowse. The space is still no concern.
Concern
For multiple user access, a single shared network server might be a bottleneck. Moreover, random disk access for bam files is relatively slower than sequential access.
Alternate
Analytical storage
The data will be used by analytical tools and will also generate temporary data files as needed.
Solution: If we use quest for computing cluster, then have to buy additional vault storage that is tied to northwestern quest. Alternate: Any cloud computing(amazon/google/rackspace etc).