Slashdot Mirror
Red Hat Not Seeing Microsoft, Ubuntu as Threats
Ian Price writes "Red Hat is shrugging off Microsoft's entry into the cluster computing space after Microsoft announced that it has completed the code for its Windows Compute Cluster Server 2003 targeting high-performance computing. From the article: 'Scott Crenshaw, general manager of enterprise Linux platform at Red Hat, dismissed Microsoft's entry into cluster computing. "They're playing catch-up," he said. "Linux is often associated with high-performance computing, but Windows has never achieved that on a large scale."' Crenshaw also commented with respect to Ubuntu: 'Their user base is still small, so we're not seeing the impact of it [Ubuntu] so far.'"
I think Netscape was quoted as saying something similar shortly before Internet Explorer utterly destroyed their marketshare. If nothing else, don't underestimate Microsoft's ability to leverage their monopoly into new markets.
I am sure Microsoft said the same thing about Red Hat. Pride goes before a fall Red Hat.
Philosophy.
Fix your package manager!
I am sick of downloading packages from weird websites, version conflicts, and typing this stupid and overly long command into the shell over and over, hoping - nay, praying - that RPM won't spit out another conflict error this time. YUM seems tacked on, and I've never gotten it to work properly.
I switched to Ubuntu, even though it had less polish and was so deep in development, simply because application management actually worked, and things were in a logical order (supported, unsupported, universe, multiverse).
Maybe it's not practical, maybe I'm talking out of my ass having not used a Red Hat operating system since Fedora Core 3, but it's the only thing that prevents me from using Fedora at home or on a server, and the only thing that prevents me from recommending it to friends.
"Joy is contagious," he said, peering into the microscope.
Years and years and years.
They've had several clusters into the Top500 several times.
A couple examples are a NSCA self-made cluster of NT machines that reached rank 207 in June 2000 top500 list. It consisted of 256-processor production supercluster, which consists of 128 Hewlett-Packard machines with dual 550-MHz Intel Pentium III Xeon processors.
These early efforts were typified by statements like:
"Couldn't barely get the benchmark done before the entire cluster would go done"
"If one node failed the entire cluster would go down"
And stuff like that.
That's the first time NT posted a top500 standing. They had earlier efforts going back several years.
About every single top500 list since then had a Microsoft-based cluster somewere.. Until recently.
Now Linux, which started gaining ground about the same time that Microsoft started with clustering research, now dominates the top500 list.
Good luck on that one, MS. I also like how their P.R. stuff always makes it sound like Microsoft just started getting into clustering.
Keep in mind this is just a PR piece. I'm sure Redhat is all too aware of the threat from their competitors. But they'd be idiots to go to the media and say, "Yes, we're really worried about the new Microsoft offering because it is superior to ours in so many ways." They will (and should) always talk down their competition externally. It is internally that they need to react to and manage the threat.
- Then they fight y...
Oh, wait... You say it's RedHat ignoring Microsoft and not the other way around??For instance, they need to achieve a "critical mass" of users before hardware and software vendors certify their products against any Linux distribution, he explained.
Ubuntu market and RHEL market is totally different. Ubuntu is "now" heading toward Enterprise desktop environment with support, but Ubuntu had and always has been about average joe's Desktop PC while RHEL had and always has been about heavily toward Enterprise customers. So I think, by reading the article, it looks like RH is taking Ubuntu as not a competitor, but rather as a grassroot movement trying to reach that "critical mass". And to be fair, Crenshaw did point out a very good point here. That is, popularity doesn't count for the vendor certification which is the industry embracing OS distro with hardware and software for better customer support and that is what Enterprise customers look for.
Microsoft being in cluster market so late in the game, it's fair to say that MS had failed to grab the market share early on. So the statement in the article is accurate. Who knows if MS will monopolize cluster market share in coming years? But this statement is on the bull's eye.
"Linux is often associated with high-performance computing, but Windows has never achieved that on a large scale."
This has been the case for Microsoft. When Win2k Data Center edition was coming out, I was hoping better support for complete cluster suite, but wasn't satisfied with MS's offering with half baked solution and limitations. Besides, call me crazy, but 200+ cluster nodes, there is no way single Windows cluster node installation will be easier than a kickstart/NFS/bash script of RHEL cluster node. I don't know, maybe there is similar thing for Windows... I'm not a Windows guy, so I'm not sure. Please do correct me.
"Don't let fools fool you. They are the clever ones."
I run Beowulf Clusters for a living.. Three to be exact. Two run Gentoo and one runs Mac OS. I see Mac OS as a far more likely product in clusters than Microsoft. And even then Mac OS is missing huge chunks of functionality in the cluster world. Checkpointing is broken using Condor and there is no third party apps for Grid Engine. Most programs fail to compile without some massaging. Often programs attempt to compile against native libraries rather than X11. This prevents remote users from using the apps.
Even with all of this though programs can be made to work. I have something like 100 custom programs that needed installed on my clusters. NCBI tools, Bio apps, stuff like that. All of them are coded to Unix environments. Compiling them on windows would be a total pain in the butt! I keep hearing that new programs will be made to work but I don't see that happening all that much. Most new programs are forks of old programs. (At least in the Bio/Geo worlds.) I still have TONS of fortran stuff out there. Lots and lots of stuff that only compiles against GCC 2.95. These things need modified in order to work with a newer version of the SAME OS.. you think a total change is going to happen?
Plus.. The cost of the OS can be killer. When you are talking $1200-$3400 a node an added $500 is huge! Our Mac OS cluster cost us $50k in software licenses. And its 50 nodes. Even if Microsoft drops the price to $100 a pop that is still REALLY expensive. $100 a pop across 50 nodes pays for a bunch more nodes!
So I guess what I am saying is that unless Microsoft starts writing tons of its own apps it won't break into the cluster world very fast. They will be luck to grow as fast as Apple has (%1 of the top 500 list in 4 years).
1) The right way of spamming the world: A top 100 cluster under Microsoft OS control gets a I.Love.U2 virus
2) An interactive assistant - Microsoft PaperClip - grows fast and takes the world under its control
3) Finally Vista runs at decent speed. Modest Min Sys Req - a cluster
4) The_Big_Bang_simulation.vbp
I agree with most of your comment. Here's where I take a different view --
Given that I'm not Microsoft, or Red Hat, I'd rather be a Red Hat stockholder than a Microsoft stockholder.
Also, I'd rather be monetizing services for rapidly spreading open-source software, than trying to get developing nations to pay for my proprietary software.
I urge you to focus on the direction and rate of the change, rather than the magnitude of the status quo.
There are too many people in the world not using computers yet. Eventually, most will. But if everyone paid Windows licensing fees, many developing nations would have to hand over most of their GNP to Microsoft. That's absurd!
In my humble opinion, it makes sense for India, China and several other developing countries to throw their collective might behind internationalized open-source software running on commodity hardware. When there are literally a million eyeballs scouring OSS for bugs, we'll see phenomenal changes in this playing field!
If intellectual property were enriched Uranium, intellectual property law would be the mechanism in an atomic bomb that prevents critical mass, and an economic boom.
-- Subvert the dominant paradigm. Repeat as desired. http://ownlifeful.com/
This time RedHat != Linux.
...
...hum... wonder if this windows flavor could be subverted as an even cheaper Windows to be installed on desktops. (I don't mind missing all "wonderful" features available in other flavors like the ActiveX-bugged IE or the DRM-laden Media Player. Just want a kernel that is compatible with games. I'll fix the gaps with OSS and stick to linux for the rest)
If RH starts loosing market share, it'll more likely be to other Linux distributor or other opensource os, like suse,ubuntu,debian,openbsd,etc.
It's not the whole Linux community of developppers ingoring they adversaries, it's only *a* specific solution vendor.
You can kill distribution, but it's much harder to kill Linux as a whole.
Netscape Navigator almost disappeared back then, because it depended on a sinle company and that company failed to notice the threat and lost market shares.
That and I'm sure Microsoft will manage to build something that sucks in terms of scaliability, reliability and above all : possibility of customisation and reasonnable per-CPU license price.
Some labs build huge clusters, this new Windows flavor must cost less than the "Windows Beginners Edition [a.k.a. 3rd world edition]" (*) and provide impeccable service, otherwise it can't compete with opensource softwares.
Plus, unlike in the browser case, Microsoft can't try to leverage its desktop OS monopoly : you can bundle a browser on a widely deployed OS, but you can't "bundle a cluster" inside the OS - that sentence doesn't make sense.
Clusters are mostly custom build to specific needs, by people who have enough technical knowledge to assemble whatever they need. Windows Cluster-flavor must attract them by its qualities, not because laziness drives them to choose whatever option came with the box...
(*):
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
What benefit? Not wiping out your entire system every 6 months to keep it running at a usable pace. Predictable reliablity, etc.
I just don't have to reboot anymore.
It's worth the driver hell that one often has to go through on a new system. Systems shipped with Linux? Probably a great idea.