Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Linux Software

What happened to the Linux SNA Project? 14

Posted by Cliff from the where-did-it-go? dept.
Blip asks: "Last time I visited the Linux SNA Homepage, it was still there. But now it seems that some commercial guy has taken over. What happened to the GPLed code? Is it all Commerceware now?"
This discussion has been archived. No new comments can be posted.

What happened to the Linux SNA Project? More

What happened to the Linux SNA Project?

Comments Filter:
  • True, SNA is still being used a lot more than many people probably suspect. There are many banks, insurance companies, and other generally conservative institutions with large investments in SNA networks that they aren't going to just throw away. They continue investing, too: adding new equipment, upgrading software, and so forth, so it would make sense that some of that new equipment could be running Linux.

    On the other hand, it seems unlikely that many people are setting up new SNA networks, and it's getting harder and harder to find people who know how to run one, which over time will make running SNA expensive or impossible. It'll be a long time before all SNA is dead, but the handwriting is on the wall.
  • At the county court house where I used to work, our AS/400 speaks both SNA (Twinax Cables for dumb terminals) and TCP/IP (on Token Ring -- soon changing to Ethernet) for PC workstations. *However* on OS/400 Release 3, (and up to R4V2) Device Naming is not supported over TCP/IP. This kills a number of important features, and makes printing harder (requires LPD). So they run special software (SNA Router) to access the SNA features. So yes, SNA is still needed, even when it runs encapsulated over TCP/IP.
  • Actually, my experience with SNA is that it is actually far less reliable than TCP/IP in practice. This is at least true when you are dealing with SNA over WAN connections, and more true when you start mixing in 3rd party SNA products like Brixton and SunLink. One of its biggest failings of SNA is that it doesn't automatically recover from line outages and requires messy 'reaquisition' of connections which often requires manual interactions. Some of these failings may be in the way that SNA works with CICS. I don't really understand the mainframe side of things that well. In comparison, once the hardware problems on a line are fixed, TCP/IP networks just start talking again.
  • http://www.snipix.freeserve.co.uk/linux.htm

    http://www.blarg.net/~mmadore/

    These 2 pages have lots of links to stuff.

    From what I understand, linux-sna.org was true vaporware... never had working code. So an old version is pretty worthless.
  • Well, I'm glad that so many people have commented here. I mean, who cares if something previously freely available and useful just poofs away into a very expensive and non-free project? Let's talk about geek profiling and evolution and cruelty to penguins.

    That being said, from what I can tell, most (if not all) of the Linux-SNA code was written by the one person. So if he wants to privatize it, well, it's only his work he's doing it to. But surely there most be a version of the old code floating around out there?

    Speaking of which, what happens when a GPL project with multiple members wants to get taken commercial / closed by its "owner"? In most cases, I assume, development would fork, with the last free version going its way, and the non-free version going its own way. But what happens when there are no copies of the free code around?

    (One also wonders what the IceLinux-Samba 'software' he sells really is...)
  • I believe SNA is a dead protocol..most mainframes can do TCP/IP, and no one seems to have a need for SNA anymore. its like the Sun i386 machine - obsolete and dead.
  • I've following the Linux/SNA project for a while now and was quite interested in it.

    I looked at the web site and they're wantin better than $3K for it? Way too much if you ask me. Sounds like they know they they don't have any competition for Linux/SNA networking software and want to get what they can from it. But, at least you get an unlimited license, and they provide the source with it.

    Anyway, I think a GPL'd SNA stack would be better. It's IBM protocol, and they're supposed to be interested in Linux now so.....

    Shatter
  • SNA is far from dead, in fact it's probably being used more now than ever. Many banks use IBM mainframes and AS/400s, and SNA is used to link them together. Several credit bureaus use SNA. As far as being obselete, IPv6 is supposed to have several features that SNA has had for years. SNA is one example of classic IBM over-engineering (kinda like those heavy guage, double-shielded serial cables I have around here somewhere, vintage IBM), it has error correction on almost every level, so SNA packets are almost never lost or damaged as they sometimes are in TCP/IP. In the last few years, IBM has added more and more TCP/IP features to their systems. But given a choice, most sys admins in charge of IBM systems will choose SNA to link them together. Shatter
  • SNA is far from dead, in fact it's probably being used more now than ever. Many banks use IBM mainframes and AS/400s, and SNA is used to link them together. Several credit bureaus use SNA.

    As far as being obselete, IPv6 is supposed to have several features that SNA has had for years. SNA is one example of classic IBM over-engineering (kinda like those heavy guage, double-shielded serial cables I have around here somewhere, vintage IBM), it has error correction on almost every level, so SNA packets are almost never lost or damaged as they sometimes are in TCP/IP.

    In the last few years, IBM has added more and more TCP/IP features to their systems. But given a choice, most sys admins in charge of IBM systems will choose SNA to link them together.

    Shatter
  • If a GPL project has multiple contributors, each contributor owns the code they submitted. (Unless they specifically give up that ownership) If the project director or whatever wishes to release it under a different license, he or she would need the written permission of each contributor. All of the contributed code for which permission could not be obtained would have to be removed. The GPL is designed to keep things GPL.
  • I've seen a couple of Linux SNA implementations besides this one, all where commercial. Unfortunatly I don't have links..

    As for the price, I imagine, since most people don't have an IBM mainframe in their basement, this is almost exclusivly for business use, $3G isn't alot of money for businesses that can afford Mainframes.

    However judging from my experience, I used to work in a shop that used to use an SNA stack on HP (SNApplus 2). I would have possibly used the Linux SNA stuff on a proof of concept project that may not have be able to get the $3G funding needed. A lower cost or free SNA kit would be useful

  • Honest question - I thought most modern mainframes could run TCP/IP and the older SNA protocols were encapsulated.


    --

Slashdot Top Deals

What good is a ticket to the good life, if you can't find the entrance?

Close