Microsoft Janus
nadador
writes "Apparently, Microsoft is readying an enterprise
class clustering and failover version of Windows 2000.
Techweb, and Microsoft, I'm sure, seem to think this is
going to be a "Unix Killer". It also mentions Linux as a
driving force in making Windows truly enterprise class
software" It actually sounds quite impressive. I can't wait
to see what some of the upcoming HA (high-availability) enhancements for Linux
will look like.
Whether NT is stable in a single server non-HA configuration or not does not matter; as long as the system as viewed from outside the cluster is up all the time with acceptable performance, there is no loss. Linux can do HA too, but the apps just aren't there. We can't beat this because we don't have control over it. Stability is really the only thing Linux has over NT at the moment in the data center, but this turns the tables. NT with failover clusters is more reliable than any single Linux machine.
Have Oracle port OPS. Oh, wait, that won't be done until raw devices are in the kernel, and Linus doesn't like them. Same for other cluster-enabled RDBMSs. Linux also has a severe filesystem deficiency right now, but as I understand it, this is being worked on, but I don't see much real progress. Other scalability concerns are being addressed in 2.3 right now, which should be out before 2000 as 2.4, if I am to understand Linus's release schedule correctly.
Another real problem with Linux is the lack of availablity of midrange and high-end hardware to key developers. My company (Denarius: http://www.denarius.com) would be more than happy to supply and set up access to high-end hardware for kernel developers as a service to the community. Hardware manufacturers would have an incentive to offer evaluations of their hardware to "sponsor" the project, as well, gaining bonus points with developers and users.
I will never use NT again
... geekboy"
I hoped on a NT machine once to port some C shell scripts into NT compatiable perl. (Was not my idea). Grabed Active perl, it installed easy, sweet I though, NT isn't so bad, I take a quick look at the code of the shell scripts, pop open edit and start to port away. 5-10-15 minutes go by, hey this isn't so bad, "NT is kinda nice" I though to myself.
BANG!! "WTF?" I think, shake shake, "Hrm."
Edit locked the system up tight. ctrl-alt-del, no good. I walk down the half and jump into the break room, 7 MSCE standing around, "Ok, no problem"
"Hey can one of you guys help me out, one of the NT machines in there isn't working right"
"Ohhh does the little unix geek have problem working with the Big Scary NT machines"
"Ha ha guys, come on help me getting this back to a normal state"
"Maybe if you didn't use a OS from before you where born you would have a couple minutes to learn a real OS like NT
"Will you help me or not"
They piled into look at how stupid I was, "look he doesn't know how to use a mouse" things like that. After 2 minutes of clicking on the keyboard and messing with the mouse one of em turns to me and 'speaking down" to me said "Look this is a power button, you push it in the machine goes off, you push it in again and the machine come on, ok do you think you can handle that?" and he fliped the power switch.
"Why didn't you just kill the process" I said
"Maybe if you didn't keep crashing it we wouldn't be in this mess"
After 3 hours of edit.com crashing 12 differant NT machine multiply times and 7 MSCE that couldn't firgure it out with help from Mirosoft Techo Support. If they can't get a small text editor to work on standard machine with $4000 tech support agreement with Microsoft, I have to take a stand and say "I will NEVER use NT again!"
I just mentioned to someone today how much better VMS' system of logicals was than UNIX' filesystems. (Not to mention Win/DOS' lack of either).
Now, we're talking about clusters and a lock manager.
Amazing that a masterwork like the Distributed Lock Manager languishes in relative disuse. Open up the source to that, Compaq!
In short, I'll believe their promises about Janus when they can switch Hotmail over to NT, IIS, MS SQL Server and Exchange.
I wouldn't be suprised to see them switch to an architecture based on Win2000/IIS5/SQL7/Exchange6 next year sometime. (Even if it takes twice as much hardware!) Don't think they've dropped the idea.
It'll be one less flamethrower for the Unix advocacy arsenel, which is fine because this one is getting a little boring.
--
Business. Numbers. Money. People. Computer World.
I agree wholeheartedly. Frankly, I didn't see what was so amusing about his 2 bits, at all. The language was incredibly stilted, and the humor, wasn't funny. It was your basic "MS Sucks" piece, with delusions of grandeur.
I appreciate that most of the readers of this site, are rabid about Linux. But c'mon people. This "humor" site, was just juvenille.
If the users of Linux can't keep the dialogue at a vaguely sophisticated level, they WILL continue to be ridiculed, and ignored. If you want to be taken seriously, you guys should at least occasionally feign maturity.
Ok, I'm back off my soapbox now.
"Politicians are interested in people. Not that this is always a virtue. Fleas are interested in dogs." P.J. O'Rourke
Actually, this quote isn't by the author of the article but is attributed to "One VAR". It's the Anonymous Coward of journalism, in this case probably someone who has an interest in Linux being seen as a toy OS. This article wasn't so much written as pieced together out of quotes from Microsoft and mostly unidentified industry sources.
And this is all really going to ship 60-120 days after the release of Win2k? Most customers will still be waiting for the first Win2k Service Pack at that point, as explained at the end of the article. Janus sounds like the same old vaporware to me.
Your right to not believe: Americans United for Separation of Church and
I think some of the *major* HA vendors do stuff like that. You'd need some pretty expensive hardware, but I think it would work. If you're building an enterprise app, you can do stuff like this using databases, complex transactions and sessions so that the user doesn't see that the server has gone down and he's been switched over to an alternate.
Yeah, Linux is still in the basic stages of HA, but you *can* actually do it at this point in time, and I'm pretty sure it's reliable. There's some cool stuff going on in HA and virtual servers right now...
æeee!
Why can't you people even wait until the product comes out to bash it!?!?
You would expect the slashdot netizens to be at least a little intelligent and open minded... but alas, you all are a discredit to your OS.
I looked at the white papers, and it sounds DAMN cool. Yet most of you imediatly start bashing NT... which is completely besides the point.
Stop bashing and start looking for the best solution to your IT problems... don't use a product simply because of who does, or does not make it.
I don't know about you, but I use something because it's the best... not because MS does or doesn't make it.
Get a life...
It'll ship exactly on time, like all Microsoft products. They are never late and the first versions kick ass. (Not a troll, sarcasm)
/. cynicism maybe (like my own...)
I'm expecting it REALLY about 2002 based on the article. BTW, funny how the above comment got moderated (+1, insightful) A little
+&x
You're confusing "success" with "usefulness." Windows is incredibly successful. It's the single most dominant operating system on the market today -- even in areas where it's not trying to compete!
Yes, it's a lousy operating system. Yes, it's a VERY lousy operating system. Yes, it's a PAINFULLY lousy operating system. But it is very successful.
Eviscerati.Org: All Hail the Eviscerati
when I see it.
It won't ship on time, we know that much for sure.
Any guesses on how much of this is just vaporware in an attempt to keep people from switching to Linux?
This space intentionally left blank.
Why was the comment below moderated down to -1 ? This isn't 'flamebait' it's a perfectly valid point. Before the mindcraft tests all the mouthy linux 'advocates' would bang on about speed. Now that's been shown to be a moot point they bang on about stability - many of them I suspect running Linux on a PC in their bedroom where neither is really that important.
-----
Before any benchmarks were done, linux people wouldn't shut up about how well it performed.
Now of stability means everything and speed all of a sudden means nothing.
Yes, I realize stability is a key asset of a web server.
Yes, I realize that all systems could saturate most leased-lines anyway...
BUT, for once I'd like to see linux advocates just take on the chin like men. Even Microsoft would be hard pressed to generate this much FUD. Yes, that's right - the linux community is the _overwhelming_ source of FUD online these days.
-- I'm drinking myself to sleep again...
Wasn't Microsoft saying that Open Source was the one that didn't innovate? It looks to me like NT is the one playing catch-up right now. We've had Linux High Availability stuff accessible for a long time right now. As well, Linux boxes have been consistantly showing higher uptimes on average than NT boxes.
I think the statement about OSS not innovating was just another piece of MS FUD. Linux didn't have the userbase to support the rapid development of applications that are growing right now a few years ago, but now that it does, it has the ability to maintain it's technological leadership.
Interesting how this battle is waging. I'm sure there will be textbooks written on the OSS vs. the capitalist monopoly stories decades from now. It's quite a classic -- heroes, villans and crusades!
:)
æeee!
Here I prooves that Micorsoft is really really very very good: http://www.freeyellow.com/members7 /geraldholmes/index.html
IBM and Sun are more than ready to fight off NT on the high end. Leave the big iron to them. Linux is poised to threaten Microsoft on the desktop and small servers - that's where we should focus our attention. Big servers might be sexy, but the low end is strategic.
Microsoft needs the desktop and ISP market 1/for revenue, 2/to leverage other products. If Linux can continue to nip away at the low end Microsoft will start to feel it in the pocketbook, and worse, they won't be able to use their dominance to take over the server market.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
Um, so yeah. Also note that none of the servers were running on Linux anyway. Why did you bring this up?
What masterwork? VMS's DLM wasn't truly distributed, it was "static master with a capability for recovery". As a result, four-node DB scalability using plain old Ethernet never exceeded 0.6. With a truly distributed lock manager and dynamic resource migration, we got up to 0.8. We also never forgave Oracle for retaining all sorts of VMS-specific semantics (the classic six lock modes, blocking ASTs) in the lock-manager specification even when running on non-VMS platforms.
DEC was pretty much the first to develop HA clustering as we know it today, and to this day leads the pack in terms of functionality. However, they tend to use a big proprietary-hardware crutch quite a bit. It used to be the HSC, now it's MC. If you take away the proprietary hardware DEC's offerings lose a little bit of functionality (though still remaining quite excellent) and a lot of performance.
Slashdot - News for Herds. Stuff that Splatters.
I was one of the principal designers and implementors of both the cluster manager and lock manager for HACMP/6000 version 3.1 (which, BTW, supported 8-way symmetric failover) back in '94, so maybe I'm qualified to comment on some of this.
;-) MS has so far exhibited nothing but the most startling ignorance and incompetence in these areas so far, and the idea that they'll suddenly leapfrog the established experts like this is just bunk. It's far easier to believe that they're deliberately making false claims to scare off the competition...again.
First, about MS. The consensus opinion among people who really know HA is that Wolfpack was and is the most pathetic piece of junk ever. The prevailing theory is that they quite deliberately announced it knowing that it was junk, to scare off anyone (such as my employer at the time) who might try to produce their own NT HA solutions.
This Janus project is just another step in that direction. 64 or 256 nodes? Yeah right. There are several reasons other HA solutions typically only go up to eight. The main one is that nobody really wants a single cluster that big. It's a total management nightmare. What customers actually want to do is set up multiple independent clusters of a reasonable size, and perhaps manage them all from within a common framework, but that's not the same as a single cluster. There's just no benefit to offset the cost of setting up failover relationships that deep and complex.
Another reason you don't see HA clusters beyond eight is that it's all but impossible to devise protocols (membership, hearbeat, consensus, and so on) that scale that high and yet still handle the simple cases efficiently. Just avoiding all the race conditions in eight nodes booting and trying to join the cluster at once is incredibly difficult. If you don't think it's that hard, try it. Have fun. Come back after you've failed, and we'll talk.
Now that I've bashed MS HA, a few words about Linux HA. It's as pathetic as MS. We have some very basic heartbeat code, and a few other scattered bits and pieces, but that's it. There's practically no fault identification to distinguish different types of failures so that one can respond differently to an adapter or network failure as distinct from a node failure. There's no lock manager. Many of the people working on the designs are only beginning to grasp the basic problems, and they're months if not years from actually implementing industrial-strength solutions. I'm on the mailing list (or I was, before I moved and had to give up my cable-modem account), I see the traffic, and it's Just Not There. I'm sorry, and I wish I could spare more time to contribute more of my own hard-won experience to the project, but that's just the way things are.
jdarcy@emc.com, until I get a new home account
Slashdot - News for Herds. Stuff that Splatters.
In other news today, the Version Inflation Index hit 666% as a large software giant released its rewritten office suite onto the market in every distributor channel. Critics call the dumping crminial. The software giant calls it marketing. What gives?
In an article containing an interview with RedHat CEO Bob Young in last months HPWORLD magazine the following was published:
"With each passing month, it becomes increasingly clear that more customers and vendors are taking Linux seriously, few more so than HP. Just before press time, sources told HP World that Linux support would soon be extended to the flagship software products WebQoS, VirtualVault, MC/ServiceGuard, and Praesidium VirtualVault. One of the vendors whose Linux implementation will be supported is Red Hat Software."
You want really good clustering!! This is good clusting. I have built several clusters using this software in my professional career and the have their act together. It can handle upto 16 cluster nodes, has built in automatic failover to a second / standby network interface in the box, and can be managed/monitored using SMNP MIB and Traps which are well documented and published.
Combine that with the Journaling File Systems support that might come from SGI... and BOOOM! You have just about every "enterprise" feature that all of the other Unixen have except maybe that system partitioning stuff that SUN can do, although that is more hardware trickery than software and could never be done on an Intel platform system.
Microsoft bought out Convoy Cluster Software for Windows NT in late 1998.
The clustering software allows for a set of machines with two network cards (one unique IP, one set cluster IP) to split network queries via a random IP address assignment method.
There is minimal priority configuration, and overall the clustering software is a joke. It does not split/share processing. The SMP support is poor to nonexistant for the Convoy cluster software.
The network configuration requires the use of a HUB (UGGH!! half duplex!!) to sync/split the cluster queries (and answer them). Switches just won't allow multiple machines to have the same IP. If you've ever tried to push 40 megabits steady out of 4 machines through a 100mbit hub and then out, you surely know the true meaning of the term 'packet loss'. I get nauseous every time i think about how bad it is.
Microsoft bought out the company and within a week released a press release stating that their new NT 5 (at the time) Enterprise Edition server would come with clustering capabilities built in.
I use this Convoy clustering software every day, in combination with Apache for NT. If anyone really thinks that this sad software really has the ability to actually threaten many of the SMP/clustering solutions for linux, BSD, or Solaris, they've got at least one ready to laugh in their face.
I've managed to get my Single PII-350 (128mb RAM) under linux with Apache to push as many queries as THREE Dual-450 (384mb RAM) NT machines using the Convoy Cluster solution with Apache/win32. Of course, I won't claim that my PII-350 is anywhere near standard in setup, but my point is that their big bad great "Linux Killer" is a real POS.
That's why I'm still converting machine after machine from NT to Linux at work. It's just ridiculous to waste so much money on NT when Linux can do the job I need done better, faster, and with better security.
.... um, i lost you after "0110100001101001".
NT 1.0? I think it all started with version 3.5. Then the version numbers got more interesting with service packs and hot fixes, but it was always 3.5 and then the infestation of 4.0 in the workforce. Confusing?
Your company, if large enough, will be able to afford such a setup. Any of your collegues use pirated or unregistered shareware and screensavers? Remember the Microsoft/SPA raids? After one of those, there might be a settlement to bring your company into compliance. They can make you an offer that can't be refused.
First off, a well designed NT box is faster and more stable than Linux. Period. You can argue personal experience all you want, but all that means is that you don't know how to build an NT box.
OK, then, if NT is faster and more stable, why is it that Microsoft's premier high-volume mail service, Hotmail, is running on FreeBSD and (IIRC) Solaris boxes, using Apache, exim, Oracle and a home-brew network file system? Seems to me that if NT can be more stable and have higer performance then it'd be easy to switch Hotmail over, and that if anyone could build a stable, high-performance NT box Microsoft itself could. The fact that they have failed at this task twice (that I know of) seems to indicate that NT just isn't up to it.
Second, I wasn't even talking about NT... I was talking about the upcoming Janus (Datacenter... whatever.) If you bothered to look at the white papers, you may have seen some very promising features.
The important ones of which have been stumping the best in the industry for quite a few years now. I'm sorry, but I just don't believe anyone has that many geniuses that they could have solved that many intractable problems this quickly and not have applied some of that to their existing software.
In short, I'll believe their promises about Janus when they can switch Hotmail over to NT, IIS, MS SQL Server and Exchange.
I initially read this as '...and fallover version of Windows' and thought that already exists.
So all the bullshit M$ have been touting in the past about Windows being enterprise-ready, are they going to now retract that, say `sorry, oops, we were wrong, it wasn't enterprise ready at all. Actually, the enterprise has changed, yeah that's the ticket'.
Somehow I doubt it. Just like I doubt this new nonsense will be a Unix-killer.
Why is it that many people who claim to support standards have such atrocious spelling and grammar?
Or do you think they will borrow the hotkey virtual console feature of linux so you can switch between the bluescreens to jot down the hex codes before you go off into MSDN? -Justin
When I see articles such as this, it makes me wish that that trade press wasn't so eager to pass along claims and speculations from unnamed sources as real news.
Let's look at these point-by-point.
Claim number 1:
This comes to us from "sources familiar with the effort as well as company information." There is no attempt to substantiate the claim. Hearsay and press releases are taken as a given.
Claim number 2
Again, more info from unnamed sources. No actual words from any actual OEMs out there. It might be IBM or Compaq, but who can say?
Now the time horizon...
This is attributed to an unnamed Microsoft spokesperson. You would think if this was for real, you might have someone like Steve Ballmer or Ed Muth crowing about it.
Enter the value-added reseller
Is this an informed opinion or just a wishful thought. Naturally the VAR may want something like this because it might enhance his or her revenue stream. This isn't really all that revealing. It's like going to a Chevrolet dealership and asking a salesman to comment on whether next year's Corvette will sell well.
More from our VAR...
The "toy" reference reminds me of a technical support call I handled three years ago from a reseller who just couldn't believe that we would port Mathematica to a "toy" OS like Linux but wouldn't support SCO's UNIX (R).
Now the analysts, also without name.
What isn't clear is whether the analysts believe that the edging is happening now or is supposed to happen when Microsoft ships this new data center variant of NT. It's all built upon an expectation of things that have no concrete existence yet.
You may also find it interesting to note that the only quotations for which there are attributions are negative things about Windows NT now and Windows 2000 later.
Quoth Don Roy of IBM:
And one from Rob Enderle of GIGA Information Group.
Translation: Businesses aren't putting too much faith in Microsoft's promises of Windows 2000. They'll wait for the service packs to come out.
-- P.J.
Here's a discussion of the Greek Mythology of Janus. Basically he's the god of beginnings, doorways, etc.. Interestingly, he's represented with two faces.... imagine that, M$ being two-faced...
So you're saying that now we can replace that ;)
one Unix box that never crashes with 8 NT boxes
that never crash at the same time. Sounds like
a deal to me!
My guess is, with this new "Janus" product, Windows will finally be stable enough to have uptimes that rival those of Unix systems. Of course, you have to have several machines to accomplish this, and the average company isn't going to have the money to buy the hardware, and from Microsoft's way of pricing, they won't have the money to buy the software.
Its a shame, really, that Unix and Linux have to be so good, since so many admin's need NT to "do it all for them with a wizard".
> Linux today is a toy techies play with
Taking this out of context, I'd say Linux Today is a Linux news site...
But in context, I'd say that if Linux is a toy, why are companies like IBM and Oracle investing time and money in it?
> Microsoft officials could not be reached for further comment.
Are we suprised?
You know the answer.
Anything that makes any OS better is a good thing. I know ill probably get "moderated" down for saying this, but I don't think Windows/NT is bad because its closed source, its bad because they haven't had much competition, and because of the way they use their software as a tool to make the victim^H^H^H^H^H^Huser buy more MICROS~1 products.
Anything (such as linux enterprise servers) that puts pressure on them to improve their product is good.
I can see it now...
MS Tech Support:
Excel crashes alot? Hmm.. Have you considered upgrading to four machines for greater reliability?
Techweb used to be nothing but good reading, but marketing seems to have wiped any common sense from their face.
I agree. Linux is a toy, just like a deisel truck. Throw anything on it and Linux will haul your load to the destination safely and keep on truckin'. Unlike NT, which requires "competent" administrators to configure it from crashing (where are these mythical people?) Linux seems to be configured out of the box for any task and will perform. NT often crashes at the turns and often burns, leaving one with property damage to deal with.
For feature rich, high-end systems, there are excellent commercial UNIX systems out there, priced at a fraction of the cost of the hardware needed to run them. Even if NT could compete feature-by-feature with those, it doesn't run on the high-end hardware (yet?).
The neat thing about Linux is that it is part of the UNIX/POSIX family of operating systems, a family that spans everything from small embedded systems to the largest scientific supercomputers and mainframes. While those systems aren't 100% compatible, it's pretty easy to port source code among them.
Microsoft seems to be trying to duplicate this with Windows APIs from WinCE to NT. But I think their range of platforms is much smaller than POSIX, and the interoperability of their own APIs among different platforms is much worse. If they want to play in this market, they have to document their APIs much more carefully and live with third party implementations.