Slashdot Mirror
DECnet Isn't Dead
Ronald Dumsfeld writes "The odds of folks under the age of 25 on Slashdot having heard of DECnet are pretty slim. This article over at Datamation gives some insight into people who've not given up on it. Poke around and find the documentation for the OSI-compliant version, or download the Linux version of the older DECnet IV and bask in the Security Through Obscurity."
It's been ages since I've heard of that. I thought it was confirmed to be dead, unlike our BSD friends whom netcraft seems to confirm to some /.ers... :D
I guess if you wanted to keep really sensitive stuff from being seen, you could PIX your application server and behind that, everything runs on DECnet behind the application server.
But, if you're going to do that anyway... you could just as well use Appletalk, VINES or NetBIOS (w/o TCPIP) instead of DECnet... neither of which would be visible to outsiders.
None of this will save you from VBS attacking desktops. Email is email, whether it came in over IP, a floppy disk, or DECnet.
That's because it is a Windows box running on an old PC. Indeed, PCs in general cannot be trusted with essential data. That is because PCs are a commodity item designed to be low in cost, and to work just well enough to outlive their warranty.
Systems from DEC and IBM, from the 1970s, are known to work very well even today. That is because they were engineered for reliability, quality, and extremely long lives (40+ years). That is why they can be trusted with critical data, even decades after they were manufactured, while a seven year old PC is most likely sitting in a closet broken, leaking mercury.
Cyric Zndovzny at your service.
The Telephone System (i.e. communication channels)
"All you have to do is be fragile and grateful. So stay the underdog." Chuck Palahniuk, Choke
You've obviously never used real computing systems engineered for quality and reliability by true engineering firms like DEC and IBM.
These systems are like the Roman aqueducts. Sure, they're ancient, but they function to the point of still being very usable today. That is because they were designed to last. I'd trust my 25 year old VAX cluster over any PC-based system or cluster any day.
Cyric Zndovzny at your service.
Yeah, I remember DECnet. The coolest thing about it is that it required you to have a special DECnet MAC address for every Ethernet port on each host. The good news is that this led to widespread Ethernet MAC reprogrammability...
Contrast with the PCs of today when we often do not enough computers for the entire class because so many of then are broken. I am not advocating going back to big iron, but when one factors in the cost of redundancy to compensate for the unreliable PC, the PC solution is not nearly so cheap.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
It all depends on what you need to do. Sure the VAXCluster was great in its day. There are still engineering and cluster management aspects that are nicer than anything new. So long as your load fits into the VAXen then that's great. Scaling your cluster is getting harder. It's getting harder to justify having a VAX filled room when the work can be done on a *single* PC. Aside from that, isn't it getting hard to find TK80 boot tapes?
What properties make it better suited than TCP/IP for critical stability? I admit that I don't know much about many of the "old" protocols, so I probably lack the perspective to really appreciate one over another (kind of like functional programming seems goofy until you try it - then enlightenment!).
Put another way, why is VAC/Alpha-over-DECnet better than VAC/Alpha-over-TCP/IP?
Dewey, what part of this looks like authorities should be involved?
Then you're completely disconnected from what it means to run computing systems that need to be reliable.
Who the heck CARES how many "Moore" generations a computer is if it does the job it was designed to do?
If your a gamer or just absolutely have to have the latest and greatest then reliability doesn't mean diddly to you. If your system is supporting space missions (for example) then it needs to be utterly reliable.
The VAX systems certainly are that - there's more engineering represented in those systems than Dell or HP or Apple can even IMAGINE applying today.
Today's desktop systems are, frankly, bare minimum crap compared to the old iron. Today's systems are cheap, throw away, crap.
Being a Christian means saving yourself until the right woman comes along.
No, it doesn't.
VAX/VMS had 64-bit computing, seamless virtual memory management, no root superuser, granular permissions, clustering, and all the other stuff *nix is just getting now (thanks to Hans Reiser, Ted T'so, Linus Torvalds and friends) decades ago. VMS was also the first POSIX-compliant system, didja realize that?
The problems with VMS were that it was expensive and closed source, and it was unfriendly to people whose native language was not English (which is why Torvalds hated it, incidentally).
I left unix for VMS because the unix geeks were condescending and unable to admit their OS had flaws, which made it impossible to fix them. I left VMS for linux because the linux geeks were actually addressing the fundamental flaws of unix.
to my mind, the situation isn't as simple as all that. There are a lot of issues to consider before giving up on a solution that is currently working just fine.
1) migration ain't cheap
2) the original post said clearly that 100% uptime was required. Correct me if I'm wrong, but most migration swap-overs, there is typically at least a few seconds of dropped client requests is there not? This would drop the uptime into the high nines.
3) I haven't heard of a migration yet that didn't have at least *one* application that needed to be modified or was broken outright
4) doing this swapover requires a lot of free space and power capacity to run both old and new sets of machines for a considerable length of time. (if this were in my server room, I'd want to keep the old hardware around for at least a business quarter "just in case".) I dunno about where you work, but I don't really have enough rackspace and power to run 10 times as many servers as I have in there now, which is what you seem to suggest.
5)Sure the newer hardware will provide faster capability, but if the current setup is running just fine at say 40% capacity, what is all that extra power going to do? I gather from the article that the server in question is used internally. Client load and load growth seem to be well in hand and easily managed. The extra hardware you suggest would just sit around warming the cages.
6)seasoned admins (and I cheerfully admit I am not one of them) don't come cheap. Someone is going to have to pay to retrain or replace these guys in order to manage the new system(s)
7) old hardware, old license, fully amortized costs vs brand-new hardware, possible licence fees, additional "hidden costs" for new backup solutions, monitoring software etc etc....to me that's no contest...
I need a wheelchair van for my son. Help me get the word out. https://www.gofundme.com/wheelchair-van-for-jj
Yes, it certainly does.
"Christian" doesn't mean following the rules you like, and ignoring the ones you don't. Many people like to call themselves Christians and ignore the rules, but that's completely different.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant