Art, This goes back to the idea of Rapid Exchange. The unloaded dataset is on a Mainframe volume(s) in the array that can be mounted on Unix and Windows systems through an API provided by HDS. The volume can be read-only to Open, read-only to mainframe, or RW for both - you choose.
The dataset can accessed as a Binary, converted through the supplied tables, or you can write your own C routine. It's not an FTP replacement, though some shops use it that way, it is file level data sharing between Mainframe and Open, and it's been around for 6 or 7 years that I know of. This also cuts out =all= the intermediate IO. And the main point is that it all happens at 2-4Gb speed on the SAN, and not 1Gb on your network. I wouldn't suggest it for shops that move a couple of GB a week to/from Open, but if the 100s of files a day are very large then it worth considering, or checking if the current DASD vendor has something similar. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On > Behalf Of Art Celestini > Sent: Thursday, 16 February 2006 10:21 PM > To: [email protected] > Subject: Re: Disk vs Tape scenario > > So, from the info Ron and Bruce provided, it sounds like multi-GB single- > threaded disk output can get 30-40 MB/sec these days. Impressive! I > don't > know if even the newest and fastest tape can compare with those numbers. > > However, for the original poster's situation, I still believe his best > solution is to write directly to his NFS-mounted Unix file system from > Fast Unload. In that way, he cuts out =all= of the intermediate I/O. > > > > ================================================== > Art Celestini Celestini Development Services ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
