Can't comment on Macs speed but when USB v 2.0 trinkets hit the shelves USB connected external hard drives will be a realistic option (they exist now but are limited by USB's current 12 Mb/sec transfer rate). USB 2.0 speeds are supposed to be 350+ Mb/sec. The big BUT in this equation is that USB is just what it's name implies, SERIAL so your SCSI chaing is going to be faster for i/o from multiple devices on the chain.
I especially liked the 4 industrial strength, rubber-coated casters - being near the top of the list I'm guessing they're more important to the functioning of the cluster than the switch, the hard drives, video cards etc. Woohoo parallel processing finally comes to PC land. This is news?
To my way of thinking, this is where *BSD/Linux acceptance and inroads will make it's largest gains. Company XYZ decides it's not paying the licencing costs for $commercial_OS anymore, does some research and finds that it might actually be feasible to switch to *BSD/Linux and gives it a try, possibly in one department first to see how the users manage (this is on the desktop, users rarely know or care what's on the servers). If the experience is positive, the whole company switches.
Even if they have to hire a consultant to set up the guinea pig department or hire a reasonably decent sysadmin (assuming a small company that has done mostly it's own IT stuff using MS) to get everything in place and answer the inevitable questions during the transition, the cost savings should be there. I know a few people who wouldn't mind being a part-time sysadmin for several companies and charging each of them a portion of a decent yearly salary, if the small companies really felt they didn't need a fulltime admin.
And as the users became comfortable exploring and learning *BSD/Linux they would become more open to installing and using it at home.
It seems that what most commercial OS'es have going for them are a large number of applications that users are familiar with and the inertia the apps and the data associated with them create. It can be fairly easy to change the OS behind the scenes, but a huge pain in the ass to change your apps, especially if you have mountains of data that would be unusable or incompatible with the new OS & apps and have to retrain users to a completely new interface and process to perform their job. This is one reason why big shops still run all manner of legacy systems. If it ain't broke don't fix it, right? Especially when you have thousands of home-grown programs (and business process surrounding them) that produce exactly the reports that your employees require to perform their duties etc.
But as the business case for *BSD/Linux in terms of cost becomes stronger and as the base of apps & usability increase, it will be harder and harder for corps to ignore their attractiveness, especially with the commercial support that Linux is enjoying and the evidence of scalability and reliability that huge sites like Yahoo, Hotmail and Slashdot provide.
We still use a Stratus where I work, it is our POS network switch. From what I understand, VOS is on it's way out and will be replaced by some other OS that Stratus doesn't have to maintain. Yes, it does support many 'obscure' networking protocols, many of which are still in use on POS networks in North America (among other networks). If I'm not mistaken, IBM's version (which the previous poster mentioned) is S/88 (the hardware), I'm not sure what OS this hardware runs. I think Tandem is another hardware manufacturer of these types of redundant, highly-available systems (or perhaps that's another name for the IBM machines?) And that's what we use the Stratus/VOS combo for: submitting transactions to our OS/390 systems for authorization or switching them to another authorization provider.   When I see the guys that work on our Stratus I just shake my head looking at the bizarre commands that they're typing in.
Um... VM is one operating system, OS/390 is another. OS/390 can run on a host VM system or directly on S/390 hardware, but not the reverse because OS/390 has no virtualizing capabilities. But you are quite right that S/390 hardware is specifically designed to assist OS virtualization and multiple instances of multiple OS'es, at a 'BIOS' level in PC terms.
Yourself and a previous poster are quite correct that IBM was the first to fully implement the idea, on System 360 hardware. VM is still around today and is still widely used on big iron. And you're quite right, IBM is very aggressive in protecting their patents. This "Flash Vos Super O/S- Virtual Flash Based Operating System" (this tripe pasted right from their site) doesn't sound like much different (but from what I could gather on their website there are probably some differences that might allow them to secure at least some of the patents) than what VM does on big iron. Just to confirm what I believe someone else posted, VM is the OS that IBM used to run 25,000+ instances of Linux on one S/390 server. Based on the broken graphics tags and totally amateurish implementation of their website I wouldn't expect much of their product or their ability to defend patents that might conflict with IBM's... but what do I know I'm just an (O)S/390 guy.
I gotta agree, (apologies to the authors) this article is really quite lame, even though I appreciate what it's intentions were, namely informing computer enthusiasts about the cooling capacity of various fans. However, this subject was covered much more thoroughly in a more well-written and humourous piece on Ars Technica, it's either a full piece comparing fans or just what's in their product reviews section. I concur with Brento, I really don't think this merited being posted on Slashdot either.
Overall I agree with your well thought-out (esp. for an empty stomach) post Jeff and there's a couple of points I'd also like to mention. I have to agree that for the really big jobs (ie enterprise critical or enterprise scale for established large companies) OS/390, VM or a commercial *NIX is the way to go, but IMHO Linux is less than two years behind some of these, in features at least.
AIX's JFS does provide the same capabilities as you mentioned for AdvFS.
Red Hat has recently released a high-availability clustering package that allows multiple types of OS'es to take part in the cluster (although I haven't looked at the details to compare capabilities to say an IBM S/70 HA system running AIX, which I'm familiar with, or OS/390's Parallel Sysplex capabilities, which sound almost identical to the ones you spoke about for DEC/Compaq. The S/390 servers in the cluster can be physically separated by a distance of 10-20 miles and still participate in the dynamic load balancing and failover responsibilities).
Kernel level auditing - I can't recall for certain if Linux has kernel level auditing which you alluded to with the system auditing point, but I believe that it does (as do all enterprise level OS'es, and NT).
As you mentioned, several vendors are jumping in with hot-swappable core system hardware components, this brings Linux (interestingly also probably NT) into the arena with E10K's and S/390 hardware.
Scalability is being addressed, but I don't expect that Linux will hit what OS/390, VM or Solaris offer (and AIX 5 is slated to introduce) regarding scalability and running multiple instances of the OS on one box. (unless you consider VM Ware, which while it may be a fine product I seriously doubt is at the capability or stability level of the other systems mentioned - anyone with solid info to the contrary please post a reply).
Multiple physical and logical paths to one physical device, including logical connections to the device from multiple instances of the OS. Let me try to explain this further for those unfamiliar with OS/390 & VM architecture. Let's suppose you have a box with a SCSI drive. Now imagine that you have more than one SCSI cable that attaches to that drive. Further imagine that each physical cable can support multiple logical connections, from more than one instance of your OS (all of which are running on one physical box). I don't know if there are any other OS & hardware systems out there that have this ability (forgive my ignorance of other hardware and OS'es of which I have only fleeting knowledge).
The ability to control the priority of processes and threads with fine granularity (255 levels IIRC). Under OS/390 you can control each thread or process running under the OS, so even if there are various types of threads that are spawned by the Webserver (just for instance - could be any thread running on the system from any app or process), I can specify what priority each has (ie I could give the thread that serves up the order form a high priority - say 255 - and the thread that serves up the complaint page low priority - say 1) for when the CPU's are at 100%.
And it's becoming ever more difficult to determine where Linux ends and commercial *nix begin... Gnome will soon be standard on numerous commercial OS'es (which is difficult for the unwashed masses to differentiate from the OS if they see the same interface) and AIX 5 will offer extensive compatibility with many core Linux / open-source apps & utilities. While this obviously doesn't mean that AIX is using the Linux kernel, the cross-pollination is occurring much faster than anyone would have dared guess even 2 years (or 1 year?) ago.
Flamers please note that I am a staunch Linux booster and have used it at home for approximately 3 years. I can't wait to see what promise it holds, but nonetheless could not go to upper management and suggest that we run our enterprise processes on it yet. Our enterprise still relies on OS/390 databases, with AIX taking on an ever larger role, and NT doing it's best to try and weasel it's way in. NT has been largely relegated to an OS for the desktop and user frontend where I work, with the backend being on AIX or OS/390 (with a couple of HP boxes hanging on by their fingernails because of apps that we can't seem to kill no matter how hard we push!)
Following info copied verbatim from Netcraft. Can't say exactly when this news is from, page info in Netscape doesn't give a last modified date but the latest results shown in the graphs are from July 2000. Not that I love to blow MS' horn for them but...
Microsoft announced a new tool aimed at giving hosting companies the capability of producing Windows 2000 based web
sites more quickly. The tool has been codeveloped with interland.net who are currently the largest hoster of Windows 2000
sites in the US, with around 2,500 sites. Interland also have around 30,000 sites hosted on Linux, and some 60,000 on NT4.
Windows 2000 has yet to see significant takeup with hosting companies, but this was clearly set out as a goal for Microsoft
by Steve Ballmer in the advent of the Windows 2000 launch.
Dogfood
HotMail has commenced its much awaited migration to a Microsoft operating system. Some Windows 2000 machines have
recently been moved into the load balancing pool, with currently between 90-95% of requests being served by the
established FreeBSD/Apache platform, and 5-10% from Windows 2000. The Hotmail site infrastructure is enormous, and
even if everything runs smoothly, a migration will likely take several weeks. LinkExchange, the other prominent FreeBSD
site owned by Microsoft, still runs FreeBSD but redirects users to BCentral.
Personally I'll still stick to solutions that use open standards and are standard-compliant when at all possible.
In a somewhat off-topic aside, when I took a C++ programming course, our instructor was very big on MS products, even though their implementation of C/C++ breaks the ANSI inheritance standards (which rightly bothered him as well as righteously pissing me off). We were learning the basic mechanics of OOP design and methodology but he kept telling us that "you'll never use a lot of this stuff (text-driven menus etc) in real life because you will most likely be doing GUI based programming in the REAL WORLD (our progs ran in a DOS window). I tried to point out to him that often you may be part of a team and your piece is the backend logic of the app and some GUI team designs the frontend or that you could be designing a program module for UNIX or a mainframe and thus would never have to deal with GUIs whatsoever, but he didn't want to hear any of it. So much for higher ed.
What is really sad is an uninformed person such as yourself who doesn't know what mainframes have become and spreads misinformation to all of the Slashdot multitude. REXX and CLIST are still the dominant scripting languages, but PERL is also available for OS/390. And JCL is a very flexible framing language that allowed modularization & reuse of code long before it became fashionable (although I grant you that it is rather arcane and unintelligible until you've worked with it for awhile). Furthermore, OS/390 currently has a posix compliant (ie Unix) shell with hierarchical file systems like your beloved PC's, run Java, C, C++ (oh and did I mention Perl?) PASCAL, FORTRAN, PL1, REXX (which is not just a simple scripting language, there is object-oriented REXX as well), and support thousands of users and thousands of batch processes concurrently. Here are the numbers of users logged on to our mainframes currently (two S/390 processors, one with 7 Gigs of memory and one with 6):
N13509 USERID TERMINAL TYPE
N13512 ------ -------- ----
N13511 13 USERS LOGGED ON, 0 DISCONNECTED.
(EASI) N13511 8 USERS LOGGED ON, 0 DISCONNECTED.
(NMQ) N13511 54 USERS LOGGED ON, 44 DISCONNECTED.
(ENET) N13511 15 USERS LOGGED ON, 1 DISCONNECTED.
(NMV) N13511 182 USERS LOGGED ON, 57 DISCONNECTED.
(NMB) N13511 403 USERS LOGGED ON, 36 DISCONNECTED.
(NMT) N13511 297 USERS LOGGED ON, 9 DISCONNECTED.
(NMU) N13511 206 USERS LOGGED ON, 75 DISCONNECTED.
(NMF) N13511 383 USERS LOGGED ON, 157 DISCONNECTED.
(NMD) N13511 650 USERS LOGGED ON, 204 DISCONNECTED.
(NMC) N13511 761 USERS LOGGED ON, 164 DISCONNECTED.
The number of users logged on reflects all of the people actually doing work (2959 to make it easy for you)... the disconnected sessions (747) have timed out, so there could be a great deal more (also keep in mind that this is Friday before one of the last weekends of the summer so there are probably on the order of 15% more people on other days). Those two processors also run approximately 12 LPARS (like running 12 copies of Windows on two PC's) consisting of 6 production images and 5-6 test images (including a Linux LPAR strictly for testing & diddling around - it also runs an instance of Apache) and an LPAR dedicated to Lotus Notes (which uses the Posix shell & HFS) which has more than 3000+ users defined. Our Development LPAR also runs Lotus Domino Go Webserver, so we have two different webservers on the same physical machine, and could potentially have 100 or more if we chose to run that many instances of Linux.
And BTW IBM does add more features to JCL from time to time as they find necessary. Add to that the fact that we run 6+ instances of DB2 UDB (supporting dozens of concurrent inquiries from stinkin' PC desktops via Datajoiner - these are people not even logged onto the mainframe, and aren't even aware that the data they see delivered to them comes from the mainframe) as well as an instance of IMS (another DBMS) and their associated interface regions... most of your sad little PC's just don't stack up. Sure some of the stuff seems archaic, but it's often actually more flexible in many ways than more current competing technologies in use today, just too difficult for some lamers to wrap their minds around. The real reason that PC's took over from mainframes is that processor technology advanced to the point that the computer could fit onto a desktop and support one user instead of thousands, and those desktop size computers were cheaper. Just go to your bank or insurance company and ask them what they use to process the bulk of their important corporate data, or the DOD which is the most secure OS they know of. IBM also offers cryptographic co-processors (CPUs specifically for encryption) on S390 hardware. These cryptographic coprocessors are the only security devices of their kind to be awarded the Federal Information Processing Standard (FIPS) 140-1 Level 4 validation by the U.S. government (from the S390 website).
OS/390 (formerly MVS) has had dataset directories for over a decade, something that MS is just now implementing with Active Directory. Can your little PC have more than one image of your OS access the same hard drive (OS/390 can have up to 5 or 6 OS images access the same hard drive), and has multiple redundant physical and logical connections to any given piece of hardware. Can you pull one of your processors in your SMP PC out and not have it crash or even hiccup? I thought not. Can you separate your clustered CPU's by 5-10 miles so some natural or man-made disaster doesn't interrupt your processing? I thought not. Well, that's too bad because OS/390 and S/390 can.
----
Re:It's ironic as hell that a Mac user said it.
on
Is UNIX An OS?
·
· Score: 1
Dead-on response! While computers and UI's have gotten more and more sophisticated, if you are really a purist in computing tech you must distinguish between the OS which provides some of the basic services and the applications that utilize those services and simplify the experience for the end user, making them more productive (unless you're like me and spend 4 hrs a day playing Hearts:-) This is the fundamental problem that geeks have with Windows, because it throws all of the OS and GUI plus pieces of (or sometimes the entire thing) apps into inscrutable DLL's, and thus it's very difficult to analyze where problems are when they arise. Combine this fact with buggy software that you feel the need to try to tinker with... At least Apple got the OS & utilities stable. But I still believe that the ability to break back to a command-line and kill off aberrant processes is extremely valuable. And as noted, many things can be done faster (particularly for fast typists with a low typo rate) via CLI than a GUI, especially when one becomes familiar with the various switches etc. that provide extreme flexibility and speed which require multi-level config popup panels in Windows. That said, when I'm unfamiliar with all of the options of a command I'll still sometimes use smit (an AIX character-cell based UI) to do configuration of various things. So for an end user of their own PC who needs/wants to do minimal system administration, it's a good thing (GUI), but for setting up multiple machines or when one becomes proficient with one's OS, it can be a hindrance to not have a CLI and be able to script some actions. But back to the original point --- Repeat ten times: Apps are NOT part of the OS!
Doh! Oh well - there's no info on Slashdot about what the site in question is anyways... still should be more careful ensuring that the Post Anon checkbox is actually checked!
Long live the blowfish! I agree, secure by default is the reason I chose to adopt OpenBSD for my webserver. For those who have never installed a *BSD (I've only dealt with OpenBSD so forgive me if this is not accurate for all the *BSD systems out there), the default is to NOT install ANY unnecessary services that could be exploited. The assumption being that if you need to have the service available on the system, you will install it. Also IIRC, it does not ship with telnet available, you must use SSH or install telnet on your own. I also rootkitted the whole damned thing as soon as I installed it (took me a fuck of a long time but I feel it's worth it). So now none of the default commands etc. do what an unsuspecting intruder thinks they might. Anyone interested should look into SecurityFocus for advisories and tips on how to secure your system (whatever OS you choose to run). For those who like Linux, you could also check out bastille-linux.org for a secured Linux solution (sorry haven't actually used it yet myself so I can't give firsthand accounting). --posting anon to protect my site's security!
At the risk of responding to a big-time troll, here I go... Why don't we concentrate on actually making the most of our current software/hardware before we decide to make bigger and better stuff first By your logic researchers should never have tried to find any tech better than vacuum tubes, in which case you'd never be able to post on Slashdot. You say vacuum tube computers were for the super rich... not really true, they were very complex and not capable of being mass-manufactured at the time (each computer of the time being the size of a small house) and consumed literally hundreds of megawatts of electricity, so they were only for the institutions that felt they held enough promise of usefulness in the future. They were initially purely for research (research into computing science mind you, not research into other things utilizing computers). You should always have visionaries looking into the future and dreaming up extremely long-term future tech whilst others are working on gradual improvements to current tech. Without such a balance society would never have looked into gasoline (internal combustion) engines, maybe we would have just constantly tweaked and refined the horse-and-buggy setup! And definitely never would have invented airplanes or dirigibles - why not perfect the automobile before worrying about such nonsense as contraptions that can travel through the air? In theory what we have now is good enough. You don't see rocket powered cars or personal helicopers in mass production and in wide use do you? Ignoring the fact that the two technologies you mention are even lesspractical than our current automobile-based existence, I personally do need more powerful hardware if I'm ever going to see a robotic butler/maid in my house while I'm off travelling to some far-away cosmos! If you are happy eating your raw meat and grains and bathing in cold water, fine, but I'm still going to use the Plasma-Stove(tm). And how do you know that the interface of the future isn't going to be one that comes as a result of new hardware tech that some engineer dreams up? I can envision a helmet-like device that reads neural activity in my brain (patent pending in too many countries to list - IOW piss off I thought it up first) and types stuff on the page so I don't have to sit here for 5 mins and respond to your post, correcting all my damned typos! But that interface is dependent on some hardware whose technology isn't available yet. And maybe we need smaller, more powerful processors to embed into the helmet to decode all of the neural activity in my head (believe me it's a hotbed of synaptic storms)... you getting the picture?
And IBM is incorporating the NUMA technology it acquired when it bought Sequent into it's server lines. The new server will be called Regatta, with two 1GHz + CPU's strapped together in a single chip. CNET has a couple of good articles about this here and here
Errrr.... and OS/390 & OS/400. They're not going anywhere anytime soon, especially considering how entrenched in corporate datacenters they are. At the company I work for we run Lotus Notes on OS/390 (4000+ users on one S/390 server with two Lotus Notes server instances) vs. approx 300 users on AIX (on an RS/6000 SP/2 node). OS/390 is a pretty big workhorse that many companies rely on heavily, although the situation is changing with the hardware improvements that have come on the other hardware platforms.
AIX (for the benefit of those who may not know AIX is IBM's flavor of Unix) is doing pretty well actually since it ships with all of their RS/6000 servers (which have been getting a larger chunk of the market since the introduction of the S80 and HA70 models), but I'm sure that IBM has realized that with the huge interest and potential that Linux has it would better serve their interests to integrate with Linux than come up with this half-baked Monterey nonsense. I never understood why IBM was going in this direction (Monterey) anyways, given that some of the Linux development teams were stating that they would probably have the first OS available for the Itanium at least 6 mos ago. Plus by making a relatively common codebase and ease-of-recompile for Linux apps onto AIX, IBM will be getting a lot of applications ready-made that they can point out to potential customers. Hell, I can't wait to be able to run some of the Linux apps on our RS/6000 server farm. Another happy announcement from Big Blue from my perspective!
Thanks for mentioning that point about the service contract (big oversight on my part)... we have three S/390's and a fair whack of RS/6000's (four SP2 frames with three Control Workstations, an HA70 and an S80 plus SSA DASD) plus an IBM tape library so yeah, we have a pretty expensive 24/365 IBM onsite support contract (lots of swag without even having to complain!) which is not nearly what you'd get with just buying the box(es). However, we got a bit of a BS runaround with Sun for a production box with a less than Super-Duper support contract - their response was quite disappointing as was the length of time that they were going to take to get a tech onsite, especially as the box is less than 6 mos old! I agree though, it baffles me why someone would feel the need to put in a different distro when VA has done a bangup job of optimizing the RH version they ship preinstalled.
Um... if IBM didn't have their ass onsite in less than 2 hrs for a major problem with any of our RS/6000 boxes the prez of IBM Canada would hear about it. But those suckers are a wee bit more than 12 large. One time we had a malfunctioning backplane in one of our systems... within 24 hours we had had the problem escalated through 3 support levels to some of their top North American specialists and replaced practically the entire node, not to mention having two techs onsite - one of them was here for 15 hrs straight. (The error that was coming up in POST didn't point to what the actual problem was so it took 'em awhile to find it). Point being if you are buying expensive hardware you can and should expect top notch service. However IBM would've told us to go blow ourselves if we'd tried to put Linux or BSD on their hardware and were calling for support, just because that is not part of their support contract with us. I expect that pretty soon we'll be trying Linux on some of those boxes as we are on our S/390, and then IBM would be on the hook for any problems even if they were running IBM OK'ed Linux version. Just my 2 cents (worth 1.2 US cents).
Completely agree with the DriveCar()... TalkToManager() was installed and enabled by default on my RH system, but on my OpenBSD system they had performed an audit on it a year and a half ago and corrected the buffer overflow exploit, but it's not installed as a default.
{:-)
While Linux has a way to go, the Linux project is getting help from all comers and all corners... including many corporations with extensive knowledge of developing OSes and related elements (not least of all SGI to whom the Linux-using public at large owes a debt of gratitude) and this should allow it to come up to the level of commercial OSes more rapidly than those commercial OSes were developed themselves. Look at it this way... it's now like IBM and SGI are co-developing a capable OS that can be run on many platforms, with a third party moderating so (hopefully) no nonsense occurs like the OS/2 - Windows dustup.
This is another RECYCLED Slashdot story... at any rate I came across the Tiqit website months ago either from a direct link on/. or from this site. I assure you I was led to those sites from some story posted on/. Still pretty cool, but not new.
Actually, it looks like the JPEG image at the top of the page is the culprit, it's doin' a helluva job at screwing up Netscrape. If you stop loading the page before the image gets loaded you're OK.
Dude, I don't know what that website is doing, but it's fscking up my Netscrape (running on NT at work here) ROYALLY! If I have a single Netscape window running, the text & hypertext get written over whatever window I was last in (imagine a page of text superimposed over top of explorer). It's not pretty! Even worse, when I scroll down the page it repeats the text like an effect from Photoshop (er the Gimp). Anyone elso out there experience this effect? Too bad too, it looks like an interesting site but the only thing I could really see on it was the direct link to the picture of the NeXT cube.
The thing I don't get is this: Cobalt got sued by Cube Computer Corp and settled, this would imply that Cube Computer Corp has some kind of trademark rights to cube shaped computers... so how come Cube Computer Corp is not now also suing Apple? Unless of course the NeXT cube predates the founding of Cube Comp. Corp. I say throw the lawyers for all three corporations into a wrestling ring and have a 'cage match' style down-and-out last person standing lawyers brawl. Whomever wins gets to sue the other two for $50M.
Slashdot Mirror
Can't comment on Macs speed but when USB v 2.0 trinkets hit the shelves USB connected external hard drives will be a realistic option (they exist now but are limited by USB's current 12 Mb/sec transfer rate). USB 2.0 speeds are supposed to be 350+ Mb/sec. The big BUT in this equation is that USB is just what it's name implies, SERIAL so your SCSI chaing is going to be faster for i/o from multiple devices on the chain.
----
I especially liked the 4 industrial strength, rubber-coated casters - being near the top of the list I'm guessing they're more important to the functioning of the cluster than the switch, the hard drives, video cards etc. Woohoo parallel processing finally comes to PC land. This is news?
----
To my way of thinking, this is where *BSD/Linux acceptance and inroads will make it's largest gains. Company XYZ decides it's not paying the licencing costs for $commercial_OS anymore, does some research and finds that it might actually be feasible to switch to *BSD/Linux and gives it a try, possibly in one department first to see how the users manage (this is on the desktop, users rarely know or care what's on the servers). If the experience is positive, the whole company switches.
Even if they have to hire a consultant to set up the guinea pig department or hire a reasonably decent sysadmin (assuming a small company that has done mostly it's own IT stuff using MS) to get everything in place and answer the inevitable questions during the transition, the cost savings should be there. I know a few people who wouldn't mind being a part-time sysadmin for several companies and charging each of them a portion of a decent yearly salary, if the small companies really felt they didn't need a fulltime admin.
And as the users became comfortable exploring and learning *BSD/Linux they would become more open to installing and using it at home.
It seems that what most commercial OS'es have going for them are a large number of applications that users are familiar with and the inertia the apps and the data associated with them create. It can be fairly easy to change the OS behind the scenes, but a huge pain in the ass to change your apps, especially if you have mountains of data that would be unusable or incompatible with the new OS & apps and have to retrain users to a completely new interface and process to perform their job. This is one reason why big shops still run all manner of legacy systems. If it ain't broke don't fix it, right? Especially when you have thousands of home-grown programs (and business process surrounding them) that produce exactly the reports that your employees require to perform their duties etc.But as the business case for *BSD/Linux in terms of cost becomes stronger and as the base of apps & usability increase, it will be harder and harder for corps to ignore their attractiveness, especially with the commercial support that Linux is enjoying and the evidence of scalability and reliability that huge sites like Yahoo, Hotmail and Slashdot provide.
Just my 1.4 cents worth (Canadian eh?).----
We still use a Stratus where I work, it is our POS network switch. From what I understand, VOS is on it's way out and will be replaced by some other OS that Stratus doesn't have to maintain. Yes, it does support many 'obscure' networking protocols, many of which are still in use on POS networks in North America (among other networks). If I'm not mistaken, IBM's version (which the previous poster mentioned) is S/88 (the hardware), I'm not sure what OS this hardware runs. I think Tandem is another hardware manufacturer of these types of redundant, highly-available systems (or perhaps that's another name for the IBM machines?) And that's what we use the Stratus/VOS combo for: submitting transactions to our OS/390 systems for authorization or switching them to another authorization provider.
  When I see the guys that work on our Stratus I just shake my head looking at the bizarre commands that they're typing in.
----
Um... VM is one operating system, OS/390 is another. OS/390 can run on a host VM system or directly on S/390 hardware, but not the reverse because OS/390 has no virtualizing capabilities. But you are quite right that S/390 hardware is specifically designed to assist OS virtualization and multiple instances of multiple OS'es, at a 'BIOS' level in PC terms.
----
Yourself and a previous poster are quite correct that IBM was the first to fully implement the idea, on System 360 hardware. VM is still around today and is still widely used on big iron. And you're quite right, IBM is very aggressive in protecting their patents. This "Flash Vos Super O/S- Virtual Flash Based Operating System" (this tripe pasted right from their site) doesn't sound like much different (but from what I could gather on their website there are probably some differences that might allow them to secure at least some of the patents) than what VM does on big iron. Just to confirm what I believe someone else posted, VM is the OS that IBM used to run 25,000+ instances of Linux on one S/390 server.
Based on the broken graphics tags and totally amateurish implementation of their website I wouldn't expect much of their product or their ability to defend patents that might conflict with IBM's... but what do I know I'm just an (O)S/390 guy.
----
I gotta agree, (apologies to the authors) this article is really quite lame, even though I appreciate what it's intentions were, namely informing computer enthusiasts about the cooling capacity of various fans. However, this subject was covered much more thoroughly in a more well-written and humourous piece on Ars Technica, it's either a full piece comparing fans or just what's in their product reviews section.
I concur with Brento, I really don't think this merited being posted on Slashdot either.
----
Overall I agree with your well thought-out (esp. for an empty stomach) post Jeff and there's a couple of points I'd also like to mention. I have to agree that for the really big jobs (ie enterprise critical or enterprise scale for established large companies) OS/390, VM or a commercial *NIX is the way to go, but IMHO Linux is less than two years behind some of these, in features at least.
And it's becoming ever more difficult to determine where Linux ends and commercial *nix begin... Gnome will soon be standard on numerous commercial OS'es (which is difficult for the unwashed masses to differentiate from the OS if they see the same interface) and AIX 5 will offer extensive compatibility with many core Linux / open-source apps & utilities. While this obviously doesn't mean that AIX is using the Linux kernel, the cross-pollination is occurring much faster than anyone would have dared guess even 2 years (or 1 year?) ago.
Flamers please note that I am a staunch Linux booster and have used it at home for approximately 3 years. I can't wait to see what promise it holds, but nonetheless could not go to upper management and suggest that we run our enterprise processes on it yet. Our enterprise still relies on OS/390 databases, with AIX taking on an ever larger role, and NT doing it's best to try and weasel it's way in. NT has been largely relegated to an OS for the desktop and user frontend where I work, with the backend being on AIX or OS/390 (with a couple of HP boxes hanging on by their fingernails because of apps that we can't seem to kill no matter how hard we push!)
----
Microsoft announced a new tool aimed at giving hosting companies the capability of producing Windows 2000 based web sites more quickly. The tool has been codeveloped with interland.net who are currently the largest hoster of Windows 2000 sites in the US, with around 2,500 sites. Interland also have around 30,000 sites hosted on Linux, and some 60,000 on NT4. Windows 2000 has yet to see significant takeup with hosting companies, but this was clearly set out as a goal for Microsoft by Steve Ballmer in the advent of the Windows 2000 launch.
Dogfood
HotMail has commenced its much awaited migration to a Microsoft operating system. Some Windows 2000 machines have recently been moved into the load balancing pool, with currently between 90-95% of requests being served by the established FreeBSD/Apache platform, and 5-10% from Windows 2000. The Hotmail site infrastructure is enormous, and even if everything runs smoothly, a migration will likely take several weeks. LinkExchange, the other prominent FreeBSD site owned by Microsoft, still runs FreeBSD but redirects users to BCentral.
Personally I'll still stick to solutions that use open standards and are standard-compliant when at all possible.
In a somewhat off-topic aside, when I took a C++ programming course, our instructor was very big on MS products, even though their implementation of C/C++ breaks the ANSI inheritance standards (which rightly bothered him as well as righteously pissing me off). We were learning the basic mechanics of OOP design and methodology but he kept telling us that "you'll never use a lot of this stuff (text-driven menus etc) in real life because you will most likely be doing GUI based programming in the REAL WORLD (our progs ran in a DOS window). I tried to point out to him that often you may be part of a team and your piece is the backend logic of the app and some GUI team designs the frontend or that you could be designing a program module for UNIX or a mainframe and thus would never have to deal with GUIs whatsoever, but he didn't want to hear any of it. So much for higher ed.
----
The number of users logged on reflects all of the people actually doing work (2959 to make it easy for you)... the disconnected sessions (747) have timed out, so there could be a great deal more (also keep in mind that this is Friday before one of the last weekends of the summer so there are probably on the order of 15% more people on other days). Those two processors also run approximately 12 LPARS (like running 12 copies of Windows on two PC's) consisting of 6 production images and 5-6 test images (including a Linux LPAR strictly for testing & diddling around - it also runs an instance of Apache) and an LPAR dedicated to Lotus Notes (which uses the Posix shell & HFS) which has more than 3000+ users defined. Our Development LPAR also runs Lotus Domino Go Webserver, so we have two different webservers on the same physical machine, and could potentially have 100 or more if we chose to run that many instances of Linux.
And BTW IBM does add more features to JCL from time to time as they find necessary. Add to that the fact that we run 6+ instances of DB2 UDB (supporting dozens of concurrent inquiries from stinkin' PC desktops via Datajoiner - these are people not even logged onto the mainframe, and aren't even aware that the data they see delivered to them comes from the mainframe) as well as an instance of IMS (another DBMS) and their associated interface regions... most of your sad little PC's just don't stack up. Sure some of the stuff seems archaic, but it's often actually more flexible in many ways than more current competing technologies in use today, just too difficult for some lamers to wrap their minds around. The real reason that PC's took over from mainframes is that processor technology advanced to the point that the computer could fit onto a desktop and support one user instead of thousands, and those desktop size computers were cheaper. Just go to your bank or insurance company and ask them what they use to process the bulk of their important corporate data, or the DOD which is the most secure OS they know of. IBM also offers cryptographic co-processors (CPUs specifically for encryption) on S390 hardware. These cryptographic coprocessors are the only security devices of their kind to be awarded the Federal Information Processing Standard (FIPS) 140-1 Level 4 validation by the U.S. government (from the S390 website).
OS/390 (formerly MVS) has had dataset directories for over a decade, something that MS is just now implementing with Active Directory.
Can your little PC have more than one image of your OS access the same hard drive (OS/390 can have up to 5 or 6 OS images access the same hard drive), and has multiple redundant physical and logical connections to any given piece of hardware. Can you pull one of your processors in your SMP PC out and not have it crash or even hiccup? I thought not. Can you separate your clustered CPU's by 5-10 miles so some natural or man-made disaster doesn't interrupt your processing? I thought not. Well, that's too bad because OS/390 and S/390 can.
----
Dead-on response! While computers and UI's have gotten more and more sophisticated, if you are really a purist in computing tech you must distinguish between the OS which provides some of the basic services and the applications that utilize those services and simplify the experience for the end user, making them more productive (unless you're like me and spend 4 hrs a day playing Hearts :-) This is the fundamental problem that geeks have with Windows, because it throws all of the OS and GUI plus pieces of (or sometimes the entire thing) apps into inscrutable DLL's, and thus it's very difficult to analyze where problems are when they arise. Combine this fact with buggy software that you feel the need to try to tinker with... At least Apple got the OS & utilities stable. But I still believe that the ability to break back to a command-line and kill off aberrant processes is extremely valuable. And as noted, many things can be done faster (particularly for fast typists with a low typo rate) via CLI than a GUI, especially when one becomes familiar with the various switches etc. that provide extreme flexibility and speed which require multi-level config popup panels in Windows. That said, when I'm unfamiliar with all of the options of a command I'll still sometimes use smit (an AIX character-cell based UI) to do configuration of various things. So for an end user of their own PC who needs/wants to do minimal system administration, it's a good thing (GUI), but for setting up multiple machines or when one becomes proficient with one's OS, it can be a hindrance to not have a CLI and be able to script some actions.
But back to the original point --- Repeat ten times: Apps are NOT part of the OS!
----
Doh! Oh well - there's no info on Slashdot about what the site in question is anyways... still should be more careful ensuring that the Post Anon checkbox is actually checked!
----
Long live the blowfish! I agree, secure by default is the reason I chose to adopt OpenBSD for my webserver. For those who have never installed a *BSD (I've only dealt with OpenBSD so forgive me if this is not accurate for all the *BSD systems out there), the default is to NOT install ANY unnecessary services that could be exploited. The assumption being that if you need to have the service available on the system, you will install it. Also IIRC, it does not ship with telnet available, you must use SSH or install telnet on your own. I also rootkitted the whole damned thing as soon as I installed it (took me a fuck of a long time but I feel it's worth it). So now none of the default commands etc. do what an unsuspecting intruder thinks they might. Anyone interested should look into SecurityFocus for advisories and tips on how to secure your system (whatever OS you choose to run). For those who like Linux, you could also check out bastille-linux.org for a secured Linux solution (sorry haven't actually used it yet myself so I can't give firsthand accounting).
--posting anon to protect my site's security!
----
At the risk of responding to a big-time troll, here I go...
Why don't we concentrate on actually making the most of our current software/hardware before we decide to make bigger and better stuff first
By your logic researchers should never have tried to find any tech better than vacuum tubes, in which case you'd never be able to post on Slashdot. You say vacuum tube computers were for the super rich... not really true, they were very complex and not capable of being mass-manufactured at the time (each computer of the time being the size of a small house) and consumed literally hundreds of megawatts of electricity, so they were only for the institutions that felt they held enough promise of usefulness in the future. They were initially purely for research (research into computing science mind you, not research into other things utilizing computers). You should always have visionaries looking into the future and dreaming up extremely long-term future tech whilst others are working on gradual improvements to current tech. Without such a balance society would never have looked into gasoline (internal combustion) engines, maybe we would have just constantly tweaked and refined the horse-and-buggy setup! And definitely never would have invented airplanes or dirigibles - why not perfect the automobile before worrying about such nonsense as contraptions that can travel through the air?
In theory what we have now is good enough. You don't see rocket powered cars or personal helicopers in mass production and in wide use do you?
Ignoring the fact that the two technologies you mention are even less practical than our current automobile-based existence, I personally do need more powerful hardware if I'm ever going to see a robotic butler/maid in my house while I'm off travelling to some far-away cosmos! If you are happy eating your raw meat and grains and bathing in cold water, fine, but I'm still going to use the Plasma-Stove(tm).
And how do you know that the interface of the future isn't going to be one that comes as a result of new hardware tech that some engineer dreams up? I can envision a helmet-like device that reads neural activity in my brain (patent pending in too many countries to list - IOW piss off I thought it up first) and types stuff on the page so I don't have to sit here for 5 mins and respond to your post, correcting all my damned typos! But that interface is dependent on some hardware whose technology isn't available yet. And maybe we need smaller, more powerful processors to embed into the helmet to decode all of the neural activity in my head (believe me it's a hotbed of synaptic storms)... you getting the picture?
----
And IBM is incorporating the NUMA technology it acquired when it bought Sequent into it's server lines. The new server will be called Regatta, with two 1GHz + CPU's strapped together in a single chip. CNET has a couple of good articles about this here and here
----
Errrr.... and OS/390 & OS/400. They're not going anywhere anytime soon, especially considering how entrenched in corporate datacenters they are. At the company I work for we run Lotus Notes on OS/390 (4000+ users on one S/390 server with two Lotus Notes server instances) vs. approx 300 users on AIX (on an RS/6000 SP/2 node). OS/390 is a pretty big workhorse that many companies rely on heavily, although the situation is changing with the hardware improvements that have come on the other hardware platforms.
----
AIX (for the benefit of those who may not know AIX is IBM's flavor of Unix) is doing pretty well actually since it ships with all of their RS/6000 servers (which have been getting a larger chunk of the market since the introduction of the S80 and HA70 models), but I'm sure that IBM has realized that with the huge interest and potential that Linux has it would better serve their interests to integrate with Linux than come up with this half-baked Monterey nonsense. I never understood why IBM was going in this direction (Monterey) anyways, given that some of the Linux development teams were stating that they would probably have the first OS available for the Itanium at least 6 mos ago. Plus by making a relatively common codebase and ease-of-recompile for Linux apps onto AIX, IBM will be getting a lot of applications ready-made that they can point out to potential customers. Hell, I can't wait to be able to run some of the Linux apps on our RS/6000 server farm. Another happy announcement from Big Blue from my perspective!
----
Thanks for mentioning that point about the service contract (big oversight on my part)... we have three S/390's and a fair whack of RS/6000's (four SP2 frames with three Control Workstations, an HA70 and an S80 plus SSA DASD) plus an IBM tape library so yeah, we have a pretty expensive 24/365 IBM onsite support contract (lots of swag without even having to complain!) which is not nearly what you'd get with just buying the box(es). However, we got a bit of a BS runaround with Sun for a production box with a less than Super-Duper support contract - their response was quite disappointing as was the length of time that they were going to take to get a tech onsite, especially as the box is less than 6 mos old! I agree though, it baffles me why someone would feel the need to put in a different distro when VA has done a bangup job of optimizing the RH version they ship preinstalled.
----
Um... if IBM didn't have their ass onsite in less than 2 hrs for a major problem with any of our RS/6000 boxes the prez of IBM Canada would hear about it. But those suckers are a wee bit more than 12 large. One time we had a malfunctioning backplane in one of our systems... within 24 hours we had had the problem escalated through 3 support levels to some of their top North American specialists and replaced practically the entire node, not to mention having two techs onsite - one of them was here for 15 hrs straight. (The error that was coming up in POST didn't point to what the actual problem was so it took 'em awhile to find it). Point being if you are buying expensive hardware you can and should expect top notch service. However IBM would've told us to go blow ourselves if we'd tried to put Linux or BSD on their hardware and were calling for support, just because that is not part of their support contract with us. I expect that pretty soon we'll be trying Linux on some of those boxes as we are on our S/390, and then IBM would be on the hook for any problems even if they were running IBM OK'ed Linux version. Just my 2 cents (worth 1.2 US cents).
----
Completely agree with the DriveCar()... TalkToManager() was installed and enabled by default on my RH system, but on my OpenBSD system they had performed an audit on it a year and a half ago and corrected the buffer overflow exploit, but it's not installed as a default. {:-)
----
While Linux has a way to go, the Linux project is getting help from all comers and all corners... including many corporations with extensive knowledge of developing OSes and related elements (not least of all SGI to whom the Linux-using public at large owes a debt of gratitude) and this should allow it to come up to the level of commercial OSes more rapidly than those commercial OSes were developed themselves. Look at it this way... it's now like IBM and SGI are co-developing a capable OS that can be run on many platforms, with a third party moderating so (hopefully) no nonsense occurs like the OS/2 - Windows dustup.
----
This is another RECYCLED Slashdot story... at any rate I came across the Tiqit website months ago either from a direct link on /. or from this site. I assure you I was led to those sites from some story posted on /. Still pretty cool, but not new.
----
Actually, it looks like the JPEG image at the top of the page is the culprit, it's doin' a helluva job at screwing up Netscrape. If you stop loading the page before the image gets loaded you're OK.
----
Dude, I don't know what that website is doing, but it's fscking up my Netscrape (running on NT at work here) ROYALLY! If I have a single Netscape window running, the text & hypertext get written over whatever window I was last in (imagine a page of text superimposed over top of explorer). It's not pretty! Even worse, when I scroll down the page it repeats the text like an effect from Photoshop (er the Gimp). Anyone elso out there experience this effect? Too bad too, it looks like an interesting site but the only thing I could really see on it was the direct link to the picture of the NeXT cube.
----
The thing I don't get is this: Cobalt got sued by Cube Computer Corp and settled, this would imply that Cube Computer Corp has some kind of trademark rights to cube shaped computers... so how come Cube Computer Corp is not now also suing Apple? Unless of course the NeXT cube predates the founding of Cube Comp. Corp. I say throw the lawyers for all three corporations into a wrestling ring and have a 'cage match' style down-and-out last person standing lawyers brawl. Whomever wins gets to sue the other two for $50M.
----