I’m not considering the cost of compression in relation to the transfer savings because the size of the files is huge (several million lines of text) that compress really well. Pkzip/gzip seems to get well over 80% compression. Then yes, after the mvs job step runs, the ftp target is in another city or even continent, and the ftp traffic is encrypted inflight using ftps.
My goal is to to compress the text file prior to ftp. Can bpxbatch programs like tar read/write from/to dd names, or fully qualified dataset names instead of Unix-like file paths? Don On Sun, Jun 30, 2019 at 09:19 Steve Thompson <ste...@copper.net> wrote: > If this file is being sent inside your firewall, the time and CPU cycles > will cost more than the ftp. This is based on experiences using MFT > products. (Basically what Gadi said). > > We found in testing that compressing was really only useful with small > pipes. Of course, there is a ratio between number of bytes to transfer and > bandwidth in determining the effectiveness of the compression (and > compression method). > > Now, if this is confidential data, and is going outside of your firewall, > you have to consider encryption. Compress first, then encrypt, because > encrypted data is generally uncompressable. > > HTH > Steve Thompson > > Sent from my iPhone — small keyboarf, fat fungrs, stupd spell manglr. > Expct mistaks > > > > On Jun 30, 2019, at 12:05 PM, Gadi Ben-Avi <gad...@malam.com> wrote: > > > > If both systems are on the same physical computer, it might not be worth > it. > > The time and cpu cycles it would take to compress and uncompress might > take longer than transferring the un compressed file. > > > > > > > > -----Original Message----- > > From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> On > Behalf Of Donald Russell > > Sent: Sunday, June 30, 2019 6:58 PM > > To: IBM-MAIN@LISTSERV.UA.EDU > > Subject: Using bpxbatch to compress an MVS dataset > > > > I have a batch process in zOS 2.1 (soon to be 2.3) that creates a large > text file I want to FTP to a zLinux system. > > > > How can I use bpxbatch tar or compress (or ?) to create a smaller file I > can ftp instead instead of the original file? I don’t want to use pkzip > unless that’s the only choice. Terse is no good because Linux can’t unterse > it. > > > > Is there a way to specify a DD name for the input and output files, > similar to how FTP allows put/get //DD:<dd name> > > > > Part two... the text in the file is EBCDIC, but Linux wants ASCII. I > don’t see an option to do the conversion. > > > > I’ll have to check tr, but maybe there’s a way to use more traditional > Unix syntax like > > > > cat //dd:in | tr ... | tar -cv //dd:out > > > > Cheers, > > Don > > > > ---------------------------------------------------------------------- > > For IBM-MAIN subscribe / signoff / archive access instructions, send > email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > > > ---------------------------------------------------------------------- > > For IBM-MAIN subscribe / signoff / archive access instructions, > > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
