C:D has supported non-mainframe systems: e.g. Unix, Open VMS,
Windows, and a few others (can't remember all the systems I had
to touch with the function test system for the z/OS product.
It has checkpoint restart, the ability to restart once it detects
the partner system is now available. It supported different VTAM
connections (I removed a few that were deprecated), but it
definitely used LU6.2, and TCP/IP (IPv4 and IPv6), along with UDP
(sp?). It also has scheduling where you can submit a transfer
that will not start until a certain time.
And you can pull from a partner as opposed to only being able to
push data to a partner.
With PLEX, one could set up a server (manager) and define several
servers for that manager to control, and one could conceivably
have a manager run 300 simultaneous transfers from one server.
(at least I remember experimenting with such a number). The
problem revolves around block size of the file to be sent or
received so that one did not go SOS because of needed I/O
buffers. And with Unix systems services files, I know that was
being worked on, but I don't know if it got implemented. Then
there was GDG support. I knew the guy that handled that and he
retired a few years ago.
Regards
Steve Thompson
On 8/14/2026 4:26 PM, Rick Troth wrote:
When I trip over a thread like this, I naturally think of NJE,
which for me was the original. (And then came the Internet.)
It's reactionary because I "felt" the demise of NJE-like file
transfer and there are few alternatives, even now.
MFT as the name of a particular solution/product seems to apply
to _file transfer over MQ_, which is cool in its own right. (I
did MQ work for BMC back in the day. Loved it.)
I've been trying to learn more about Connect:Direct which
presents its own learning curve, but seems to be a better
mousetrap than FTP.
Never been a fan of FTP, and I continue to hear colleagues
complain about it. I major factor with FTP (and others) is the
need for credentials on the receiving end. This is sadly also
true of SCP/SFTP (of which I *am* a fan, it's good stuff).
And then in the consumer world, people attach files to email.
Don't get me started.
Please keep UFT in mind. Forgive me, Steve, because I don't
mean this as a hijack.
When you compare UFT against NJE, it's operationally very much
the same but uses an open protocol. UFT also requires less for
"client mode" operation. (You don't have to establish an NJE or
RJE node just to send files.) One aspect of UFT is when sending
to /user/@/node/, the "user" might actually be just a queue
(shades of MQ) and not an actual user/ID.
I've been using UFT on my home systems for years. Love it! Just
send the file ... done! Ahhhhhh......
The C client in the UFT project supports a proxy, so one could
theoretically send UFT traffic via MQ. But I have not done any
work with MQ products or utilities in too many years. [sigh]
The happy news in *my* playground is that my employer picked-up
UFT as the basis for their (not yet GA) VFILE product. If
you're interested, send email to support@.
(And to the list admins, this is not advertising. I am again
mentioning the *protocol* which was designed from the get-go to
support Z operating systems.)
On 8/14/26 8:42 AM, Steve Thompson wrote:
I know that they acquired Connect:Direct, I was at Sterling
when we got acquired by IBM.
Does IBM have any other MFT products for z/OS or z/VM?
--
Regards,
Steve Thompson
MaGA:
Make Mainframes Great Again:
Mainframes have been running AI products since the 1980s!!
Mainframes were doing Cloud before "Cloud" was a thing:
Java, scripting, HA, MQ, Linux, client-server...
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN