How would you go about transferring a 500GB file from an MVS system to a Sun Solaris UNIX system every 10 minutes? We are currently using tapes and it takes too long. We need a fully redundant and highly available (and dependable) solution. Our current storage environment is primarily IBM and Sun.
To achieve the throughput you are talking about you will require a sustained transfer rate of greater than 853 MB/s. This is not practical for the multi-platform transfer software such as CNT's UltraNet Gateway or IBM's InfoSpeed.
But back to the 500GB issue?
One way to tackle the problem is to use a filesystem abstraction software method. This is accomplished by exploiting a common filesystem (such as NFS) and using an agent to handle lock management between the MVS and open system.
EMC's InfoMover product provides this functionality. InfoMover allows data that is accessed normally by the MVS systems to also be accessed by an open system. Therefore, the data is not transferred, but accessed in place. The software agent on the open system controls the EBCDIC to ASCII translation and locking. Detailed information can be found at EMC InfoMover.
In your situation, you would stop updates to the MVS 500GB file, the file would be accessed by the Sun system, and updates could resume immediately after use by the Sun system. By using this technique, you can reduce your ten-minute cycle to only the time needed to access the file from the Sun system.
You said that using primarily IBM and Sun but a requirement for this technique is a storage platform such as an EMC Symmetrix that allows logical volumes to be addressed by both a mainframe and open systems.
InfoMover is both highly available (multiple channel paths) and dependable. I have successfully deployed InfoMover and it has been transferring files for a few years now at a client site.
Hope this helps.
Dig Deeper on Data storage strategy
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.