Re: Internet history (was: We need a better Internet)
On Thu, Apr 8, 2010 at 8:47 AM, Tom Buskey t...@buskey.name wrote: Because we can't keep track of 100 systems what they do in our head. Yah. At $WORK, desktops and laptops have generic names (a static prefix followed by a number), because they're commodities, interchangeable and uninteresting. We only have a few servers, all doing multiple duties, so they have interesting names, like TIGER and MAGNUM. At a previous job, we used Simpsons characters. I remember the nameservers were ITCHY and SCRATCHY, and the firewall was WIGGUM. One benefit to using visual source material is you can print images of the characters and tape them to the machines. :) Fidonet had some cross platform too. Mostly PC DOS, but I remember Z100 (not PC compat, but still MS-DOS) running binkyterm, opus, etc. There was a Linux port of Fidonet stuff as well. There's more than one Fido-style BBS package for *nix. FidoNet's still around. http://www.fidonet.org/ I'm told it's still popular in regions of the world that don't have strong Internet connectivity. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
Tom Buskey writes: Because we can't keep track of 100 systems what they do in our head. But using a naming scheme means you can script it. We don't really care about the names otherwise. Oh, and only one name because if there's another name, we'll get a ticket to fix it by the name we can't script. I used to work on a parallel computer whose compute nodes were named after stars. So, whenever I needed to do something to all of the nodes in the cluster I'd have to write code like: for H in antares atria avior sirius \ regulus becrux pollux mirtak ; do done I would have preferred a more uniform set of names, but that's the way that things go sometimes. --kevin -- alumni.unh.edu!kdc / http://kdc-blog.blogspot.com/ GnuPG: D87F DAD6 0291 289C EB1E 781C 9BF8 A7D8 B280 F24E Wipe him down with gasoline 'til his arms are hard and mean From now on boys this iron boat's your home So heave away, boys. -- Tom Waits ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On Thu, Apr 8, 2010 at 9:31 AM, Kevin D. Clark kevin_d_cl...@comcast.net wrote: So, whenever I needed to do something to all of the nodes in the cluster I'd have to write code like: for H in antares atria avior sirius ... I remember a shared login script at UNH which defined various names, so you could do things like: for H in $DWARVES ; do ... ; done Makes sense to put something like that in /etc/profile or whatever, if you're doing to use the fancy name strategy. :) -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/07/2010 11:42 PM, David Hardy wrote: Yes, md, I remember, as do many or all of us, the same bunch of names for the systems, usually either from the Snow White gang, or Lord of the Rings, or Hitchhiker's Guide. Them were the daze. Now our brilliant successors name them with strings of alphanumeric characters the provenance of which only they, the holy annointed ones, can fathom. Those were good when the number of systems was relatively small. We had the good fortune of having musicians in our group when we built our HPC cluster. We wound up naming it 'Orchestra' and everything is named after an instrument, giving us pretty much unlimited expansion. Each node type or function gets a class of instruments - databases are percussion, compute nodes are woodwinds or string, etc. I have a bit of trouble remembering which Oracle system bass-drum is, but we maintain a wiki page that maintains the list in case we need to look it up. Every system that has external connections has a service name that's not the same as the system name (tomcat, www, jboss, mysql, etc.) so that we don't have users pointing directly to something like 'marimba' and then have to re-educate them when we retire that system in favor of a new one. They just still use mysql. -Mark ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/08/2010 09:31 AM, Kevin D. Clark wrote: I used to work on a parallel computer whose compute nodes were named after stars. So, whenever I needed to do something to all of the nodes in the cluster I'd have to write code like: for H in antares atria avior sirius \ regulus becrux pollux mirtak ; do done I would have preferred a more uniform set of names, but that's the way that things go sometimes. dsh solves that problem -Mark ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
I've always felt that at a minimum servers deserve real names. The first naming convention I saw was on my first Unix account, SunOS boxes in CS at University of Hartford, named after movie computers (hal, skynet). At a previous job, we had HP-UX servers named after characters in Johnny Quest (race, bandit, hadji, etc) and Sun servers named after 70s/80s detective show characters (columbo, baretta, mannix, magnum). I've seen planets, stars, constellations, and Greek or Roman gods used a lot. At a previous job, we used Simpsons characters. I remember the nameservers were ITCHY and SCRATCHY, and the firewall was WIGGUM. One benefit to using visual source material is you can print images of the characters and tape them to the machines. :) When I worked at DSL.net, we named all of our servers after beer or breweries. The benefit of that source material is you can steam the labels off of bottles and put the labels on the servers (well that, and we had a rule for awhile that we couldn't use a name if someone in the group hadn't had the beer). We had machines like ottercreek, chimay, bud, fosters, tooheys, papercity (a printer), oldnick (the workstation of a guy named Nick). My favorite was our backup server, although technically called backup01, it had the beer label for Magic Hat's Blind Faith on it. I still keep the tradition alive today. I have a number of Linux infrastructure VMs named for Lovecraftian places (dunwich, arkham, dreamlands) -Shawn ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On Thu, Apr 8, 2010 at 10:42 AM, Shawn O'Shea sh...@eth0.net wrote: I've always felt that at a minimum servers deserve real names. It really depends on the environment. The more commoditized things are, the less sense it makes to have fancy names. If you've got a 100 node server farm for some massive web site project, everything's an interchangeable part, and it's likely machines are dedicated to single tasks. OTOH, small orgs usually have a small number of multi-purpose servers, and roles get moved around between them a lot, so it makes more sense for the servers to be unique entities in their own right. As Mark Komarinski already mentioned, it's always a very good idea to have generic service names for roles, and alias those names to the machines filling the role. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/08/2010 01:10 PM, Benjamin Scott wrote: As Mark Komarinski already mentioned, it's always a very good idea to have generic service names for roles, and alias those names to the machines filling the role. And, I've noticed that with the rise of virtualization the role names are once again becoming sensible host names for single-purpose VM's. Hardware can be more interestingly named. I name mine after real-world heroes: borlaug, librescu, stevens, kanzius, randi. For me, it's easier to remember that 'smtp' is on 'borlaug' than 'vm3'. I use DNS to look up my own less-used servers' IP's. But I'm not strongly left-brained, so ymmv. -Bill -- Bill McGonigle, Owner BFC Computing, LLC http://bfccomputing.com/ Telephone: +1.603.448.4440 Email, IM, VOIP: b...@bfccomputing.com VCard: http://bfccomputing.com/vcard/bill.vcf Social networks: bill_mcgonigle/bill.mcgonigle ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On Thu, Apr 8, 2010 at 2:41 PM, Bill McGonigle b...@bfccomputing.com wrote: For me, it's easier to remember that 'smtp' is on 'borlaug' than 'vm3'. I would think an alias (CNAME) to smtp would be easier still... ;-) -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
Tom Buskey writes: On Thu, Apr 8, 2010 at 10:42 AM, Kevin D. Clark wrote: The problem that I had was that I frequently had to deal with the situation of this particular problem only really efficiently runs on 1, 4, or 16 nodes in the cluster or this problem only really efficiently runs on 1, 2, 4, 8, or 16 nodes in the clusternow, what nodes were these again, and how do I relate all of the logfiles that I obtained from the last program run? You might have proven my point. Just to be clear, I was trying to illustrate your point, because you an I appear to be in complete agreement on this issue. Regards, --kevin -- alumni.unh.edu!kdc / http://kdc-blog.blogspot.com/ GnuPG: D87F DAD6 0291 289C EB1E 781C 9BF8 A7D8 B280 F24E Wipe him down with gasoline 'til his arms are hard and mean From now on boys this iron boat's your home So heave away, boys. -- Tom Waits ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/08/2010 03:13 PM, Kevin D. Clark wrote: Tom Buskey writes: On Thu, Apr 8, 2010 at 10:42 AM, Kevin D. Clark wrote: The problem that I had was that I frequently had to deal with the situation of this particular problem only really efficiently runs on 1, 4, or 16 nodes in the cluster or this problem only really efficiently runs on 1, 2, 4, 8, or 16 nodes in the clusternow, what nodes were these again, and how do I relate all of the logfiles that I obtained from the last program run? You might have proven my point. Just to be clear, I was trying to illustrate your point, because you an I appear to be in complete agreement on this issue. Maybe I'm not understanding the issue, but isn't the above why queuing systems were made? We're using a dirt-old version of Platform LSF and it already solves the 'running on heterogeneous systems distributed across an arbitrary number of nodes' problem. While returning the output via LSF or shared filesystem. The original problem ($DWARVES) had to do with doing what really looks like sysadmin-type stuff, which dsh already can do. It has the notion of groups so you can have Solaris-specific commands sent to the group of Solaris systems, Red Hat-specific to Red Hat, etc. or have a group that includes all hosts for commands that works across everything. You can have dsh dispatch commands concurrently rather than serially that the for loop does. We can get ~200 nodes updated via systemimager in only a few minutes using this method. -Mark ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
Mark Komarinski writes: Maybe I'm not understanding the issue, but isn't the above why queuing systems were made? We're using a dirt-old version of Platform LSF and it already solves the 'running on heterogeneous systems distributed across an arbitrary number of nodes' problem. While returning the output via LSF or shared filesystem. I don't have much to say about the how to manage heterogeneous systems problem. The parallel computer that I worked extensively with in the past consisted of machines that were all running the same hardware and the same OS (this was deliberate). From my perspective, employing a boring/normalized/easily-scriptable naming scheme would have been advantageous. This isn't the scheme that was put into place ; it wasn't the end of the world for me ; I worked with this. Systems like LSF sound neat but I've never had occasion to use a system like this ; I can't comment on these things. Regards, --kevin -- alumni.unh.edu!kdc / http://kdc-blog.blogspot.com/ GnuPG: D87F DAD6 0291 289C EB1E 781C 9BF8 A7D8 B280 F24E Wipe him down with gasoline 'til his arms are hard and mean From now on boys this iron boat's your home So heave away, boys. -- Tom Waits ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/08/2010 02:51 PM, Benjamin Scott wrote: I would think an alias (CNAME) to smtp would be easier still...;-) There's nothing to alias it to, that's all the vm does, so that's what it's called. The physical host will have a different IP MAC. Besides, MX records need to point to A records. ;) -Bill -- Bill McGonigle, Owner BFC Computing, LLC http://bfccomputing.com/ Telephone: +1.603.448.4440 Email, IM, VOIP: b...@bfccomputing.com VCard: http://bfccomputing.com/vcard/bill.vcf Social networks: bill_mcgonigle/bill.mcgonigle ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Internet history (was: We need a better Internet)
On Wed, Apr 7, 2010 at 12:14 AM, Ric Werme ewe...@comcast.net wrote: The follow on to the ARPAnet, the Internet, started around 1980 with the publishing of the core Internet protocols and porting classics like the new (1973) FTP and Telnet protocols and new ones like NFS and the rest of ONC-RPC. I wasn't there, but as I recall from reading histories: The original ARPANET ran something called NCP (Network Control Program). NCP was sort of like a limited version of TCP-over-IP. It was one-way, i.e., you needed two NCP connections to make a bidirection channel. This is why all the classic Internet application protocols (Telnet, FTP, SMTP, DNS, etc.) have odd numbers -- the even numbers were reserved for the return channel. The ARPANET was cut-over to IP and TCP at some point, and continued to run that way. Since ARPANET was nominally the domain of the US military, other networks (like NSFNET) were started, using the same standards, but with different nominal jurisdiction. Gateways were established. Eventually the various networks converged into the thing we call the Internet today. There isn't a clear line, in space or time, between ARPANET and Internet. Linux didn't appear until 1991 or so. I was off net in 1980, but I think BSD Unix is to the Internet as TENEX and PDP-10s were to the ARPAnet. I believe the original IP/TCP implementation was on a non-Unix as well, although I don't have a reference to hand. RFC-801, NCP/TCP transition plan (1981), documents part of this. Appendix D lists known implementations in alphabetical order (no mention of incept dates). Several Unixes appear, including BSD, but also ATT Research Unix V6 and V6, including what we would today call a distribution by BBN. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On 04/07/2010 09:38 AM, Benjamin Scott wrote: Since ARPANET was nominally the domain of the US military, other networks (like NSFNET) were started, using the same standards, but with different nominal jurisdiction. Gateways were established. Eventually the various networks converged into the thing we call the Internet today. There isn't a clear line, in space or time, between ARPANET and Internet. Actually ARPANET, while a DOD sponsored network, was a way to connect universities so they could share research. I'm over generalizing, but it wasn't strictly an internal military thing. -- Jerry Feldman g...@blu.org Boston Linux and Unix PGP key id: 537C5846 PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB CA3B 4607 4319 537C 5846 signature.asc Description: OpenPGP digital signature ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
On Wed, Apr 7, 2010 at 12:39 PM, Jerry Feldman g...@blu.org wrote: Actually ARPANET, while a DOD sponsored network, was a way to connect universities so they could share research. I'm over generalizing, but it wasn't strictly an internal military thing. Right, I didn't mean ARPANET was used by DoD only, I meant it was under their jurisdiction. IIRC, ARPANET was later split into MILNET (which *was* DoD only) and NSFNET, or something like that. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
*And let's not forget EasyNET, people, at DEC, back in the glorious '90s. * On Wed, Apr 7, 2010 at 1:28 PM, Benjamin Scott dragonh...@gmail.com wrote: On Wed, Apr 7, 2010 at 12:39 PM, Jerry Feldman g...@blu.org wrote: Actually ARPANET, while a DOD sponsored network, was a way to connect universities so they could share research. I'm over generalizing, but it wasn't strictly an internal military thing. Right, I didn't mean ARPANET was used by DoD only, I meant it was under their jurisdiction. IIRC, ARPANET was later split into MILNET (which *was* DoD only) and NSFNET, or something like that. -- Ben ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/ ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
True enuff; and I made a little boo-boo: EasyNET was around in the glorious '80s, too! We were running VMS, natch, and little old me got jammed up behind a couple of posts I made there concerning contemporary international politics, about which I will say no more, as I don't wish to have the repercussions I got back then. EasyNET had a slew of various threads on all kinds of interests and was the predecessor, at least as far as I'm concerned in my own experience, of places like AOL's and CompuServe's forums, and Salon's Table Talk (where I also got jammed up a couple of times) Oh Lordy, now the memories are flooding back: USENET, and in my own neck of the woods in suburban Beantown back then, the Boston Computer Society's BBS, which I accessed through a DEC Rainbow, the machine I also used to login to DEC VAXen at work to monitor various jobs and processes from home. Ain't it great to be one of the surviving dinosaurs? cheers, from rainy northern Vermont today... On Wed, Apr 7, 2010 at 3:52 PM, Benjamin Scott dragonh...@gmail.com wrote: On Wed, Apr 7, 2010 at 2:13 PM, David Hardy belovedbold...@gmail.com wrote: And let's not forget EasyNET, people, at DEC, back in the glorious '90s. Heck, back then, everyone had a cool name for their own network. ;-) -- Ben Scott @ FidoNet 1:324/127.4 ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/ ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
EasyNET was around in the glorious '80s, too! uucp(1) - Unix to Unix Copy decvax!maddog - who needs any stinking domain names? And surely you *name* your computer systems! shaman, guru, shamet, wicca - my systems all have *names* Snow White and the Seven Dwarfs! (sneezy, dopey, doc, bashful, grumpy, sleepy, happy)and I did not even have to look up the names...not because of the movie, but because of the eight systems in engineering. Now I go back to sleep. md P.S. Do not forget Fidonet! Urf! ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
Yes, md, I remember, as do many or all of us, the same bunch of names for the systems, usually either from the Snow White gang, or Lord of the Rings, or Hitchhiker's Guide. Them were the daze. Now our brilliant successors name them with strings of alphanumeric characters the provenance of which only they, the holy annointed ones, can fathom. And, if memory serves, FidoNet ran a whole lot of those BBS thangs. Sleep sounds good on this rainy, foggy and thunder-stormy night here in Vermont...cheers to our sister state of NH and the folks down in the tropics of Massachusetts. On Wed, Apr 7, 2010 at 10:51 PM, Jon 'maddog' Hall mad...@li.org wrote: EasyNET was around in the glorious '80s, too! uucp(1) - Unix to Unix Copy decvax!maddog - who needs any stinking domain names? And surely you *name* your computer systems! shaman, guru, shamet, wicca - my systems all have *names* Snow White and the Seven Dwarfs! (sneezy, dopey, doc, bashful, grumpy, sleepy, happy)and I did not even have to look up the names...not because of the movie, but because of the eight systems in engineering. Now I go back to sleep. md P.S. Do not forget Fidonet! Urf! ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
Re: Internet history (was: We need a better Internet)
Walt, Mickey, Minnie, Sleepy, and friends... Oh, yes, and I remember that somehow Satan (ran out of Disney names...) had a vote to make a quorum for the cluster, which was a huge headache to find. Satan, of course, was on the far end of the building from Walt (obviously only Walt was *supposed* to have a vote), on a user's desk, over thicknet. The rest of the workstations were tapped off that on 10Base2, and of course from time to time someone would run the end of the cable straight in to an Ethernet card instead of using a T and a terminator. *shudder* The network runs were *weird*, too, since the classified and unclassified cables couldn't come within a meter of each other. You are in a maze of twisty little passages, all alike. It took us months to figure out what was going on, since we had to keep the system up as much as possible during working hours... and they kicked us out and locked the vault door on the front of the building at 6 sharp. That and there were only two of us, and we kept getting sent out to do temporary cluster installs on subs. I am a BIG fan of star topology (including in non-network electrical systems) as a result. --DTVZ ___ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/