“The server operating system wars never seem to slow down. Last week it was Red Hat’s turn with the announcement of Red Hat Enterprise Linux 5, which incorporates the Xen open source hypervisor. Naturally there’s also the endless market speculation about the final feature set and likely arrival date of Windows Server 2007. And then there’s Solaris, which with its nice value-add features like DTrace and its new status as open source software is making something of a comeback, it seems.”
The last paragraph tries to sums up the article but fails miserably.
It’s basically a long winded moan about A: The servers wars do not seem to be slowing down and how can a person be expected to keep up with all these changes?(my advice is to get a new job and let the kids take over from where he left off). B: RedHat is trying to beat Sun at it’s own game while Microsoft is battling it out with Oracle for the database market(please tell me something we all didn’t know) and C: While everybody is kicking the crap out of each other in the Unix sphere, MS is cashing in with Exchange and SQL sever 2005 (another situation that most of us are well aware of).
Now, taking all that into consideration, tell me how the hell did he reach the conclusion stated in the final paragraph quoted above? Is it just me or does anybody else fail to see the chain of thought that led to the above wide reaching, earth shattering conclusion? Surely, when making such statements, it’s a good idea to actually include the chain of thought that led to them? And how, pray tell, does this have anything to do with RedHat’s choice of customer?
The more I hear people bitching and wining about the moral foundation of Linux (and FOSS in general) the more I want to hop on the next flight, grab them by the collar and rattle they’re thick heads about while screaming at them to wake up! Arguably, Linux is where it’s at because of it’s moral foundation. How you choose to interpret or use that foundation does not in anyway subtract from the overall benefit of that foundation!
IMHO, this article is to be avoided. Well unless, like me, you have absolutely nothing else to do 😉
I found the article a bit meaningless. Yer, duh, people pick their systems based on what they can run on it and what they can get away with, cost-wise, but he then goes on to do a pretty inaccurate estimation of Linux and Solaris servers. It’s just impossible to estimate and compare, and it’s not even useful as an approximation.
He also goes on to fall into the trap that Microsoft leaves for silly analysts when they try and compare Windows Server market share to everyone else. He regurgitates, verbatim, how Windows Server has a greater market share based on revenue. Not shipments, which he seems to get mixed up, but revenue, which can mean and include almost anything:
Windows Server commands about two thirds of total server unit shipments and drives almost three times the dollar value in server system sales as Linux ($5.3 billion vs. $1.8 billion in Q5 2006).
There’s also a few other pot and kettle comparisons:
Compare that to Microsoft’s estimate (on the same slide cited above) that the total world installed base for database servers is three million machines, or Oracle’s statement in its marketing materials that it has around 250,000 customers. Of course Oracle’s database business is larger in dollar terms than Microsoft’s, because its database is a lot more expensive.
OK. So Oracle gets more revenue from databases because it is more expensive, but Microsoft doesn’t get more revenue from Windows Server (and applications, which can be included) because it is more expensive? Bizarre. There’s not an awful lot of reasoning going on there.
Red Hat, has forgotten to pursue a volume strategy and has painted itself into a premium-priced corner that it is busy fighting over with Sun.
That, I can agree with. There is are just too many companies like Red Hat, most of them in more niche markets around data centres, who want to steadfastly pretend that they do not compete with Windows Servers of Microsoft’s applications. They are sitting on what they have, rather than looking at things like Exchange usage and trying to attack it.
But MySQL and Postgres aren’t viable replacements for Oracle or SQL Server
That’s pretty debatable. An awful lot of companies are needlessly using SQL Server, and certainly Oracle.
and PHP and Ruby aren’t viable replacements for enterprise Java or .Net.
Java is very much at home on a Linux system, and companies are doing it.
And nobody seems to be doing the work needed to make Linux scale on the new multiprocessor scale-up server hardware…
That’s pretty meaningless. He’s going to have to define what that is.
and PHP and Ruby aren’t viable replacements for enterprise Java or .Net.
Well he has a point there. Nothing takes out a server quicker than a big fat JVM humping on it’s RAM. For now ruby will will have to settle for being cleaner, lighter, and far more functional than Java.
I guess your one of those types that believes fortune 500 companies should be running gentoo on their servers?
Of course not, but since you’re into making radical assumptions based on your lack of experience with the subject at hand, might I make one of my own. You’re one of those people who buy into the whole “Enterprise” buzzword crap, aren’t you?
You see, I’ve actually been working with web servers for many years and have seen all too often what Java can do to a perfectly good machine. The fact is, no matter how pretty your web app is, it’s not going to matter if your machine has to be rebooted three times a day.
You see, I’ve actually been working with web servers for many years and have seen all too often what Java can do to a perfectly good machine. The fact is, no matter how pretty your web app is, it’s not going to matter if your machine has to be rebooted three times a day.
Strange; I have never found that.
Do you think Java would be so widespread in the industry if this really were the case? Rebooting three times a day… don’t be ridiculous.
Quite frankly a little java app in your browser is ok to some degree, but beyond that it is just a useless memory hog that could easily be replaced in many instances by other technologies.
Well, that’s about the only place I NEVER want to see a Java app(let). The JVM starts up slow (slower than by starting it manually), and stays in memory until the browser is closed. Sun failed here big time; they should have made it lighter, like flash (but preserving the speed).
And as for the article: it really isn’t worth reading.
Sun dropped the ball on client side java in a general way, and have only started groping for it since the advent of .net. To make matters worse, IE shipped with its own quirky and archaic JVM, and never upgraded, so if you want to write an applet your are writing for java 1.1 (AFAIK). To put the final nail in the coffin, flash appeared and solved virtually every issue in the space applets were trying to address, just better in every conceivable way.
My first job out of school was working on a J2EE project, in my 4 years there we only used an applet once, and it was non visual to handle some rather intense client side calculations. When AJAX came out, that was the first thing we used it for, to replace the applet.
Java does take alot of memory, and doing things The Sun Way includes massive amount of infrastructure, and considering those two points, java is not really appropriate for anything but large projects on big hardware.
BUT, java also has a fantastic security record, and has functionality that can’t be matched things like ruby. Ruby on Rails is fantastic for what it is, but to say it replaces everything is the same kind of statement as Gentoo should be run everywhere.
I think you miss the point so far:
In enterprise the less count is the “lighter” of the plataform, don’t you think ?
“For now ruby will will have to settle for being cleaner, lighter, and far more functional than Java.”
Why are stupid comments like these rewarded with mod-ups? This site use to have a following of people interested in information but now it’s a tap for the same kind of ludicrous fanaticism displayed by G.W. Bush and militant Islamic theocracies. Explains why I haven’t been here in a while.
Ruby is lighter than Java but even Ruby fanbois admit Ruby’s support of UTF leaves something to be desired.
I mean are we talking standard libraries? If so Java kicks Ruby in the teeth repeatedly with cleats on.
As far as languages themselves go … I haven’t actually done much research to compare the two but the right tool for the right job and I’m sorry but Ruby doesn’t cut it as the right tool for i18n’d and l16n’d projects. The web is multilingual these days. I swear half these people could just go back to CGIs with iso-latin character encodings and still be happy.
Where do these articles come from? To say that MySQL and Postgres are not viable replacements for Oracle and SQL Server shows exactly the ignorance that persists throughout the IT world, especially in upper management. These people, much like this author, simply form ideas based on ignorance. Because they are not familiar with MySQL, thus it is not equal. The reality is way too many IT managers have chosen SQL simply out of ignorance when they very well would have done much better with MySQL or Postgres.
He does somewhat have a point in that users do decide which applications are needed first, then chose the OS. But beyond the large enterprises, and a few SMEs, Windows has already been chosen by default because it is a known networked/workgroup OS, where as Linux is still proving itself to them, if they are even familiar with it in the first place.
Finally, yes PHP/Ruby can do much better than Java/.Net. As fskit points out “Nothing takes out a server quicker than a big fat JVM humping on it’s RAM. For now ruby will will have to settle for being cleaner, lighter, and far more functional than Java.”
I for one prefer just to avoid Java all together. Quite frankly a little java app in your browser is ok to some degree, but beyond that it is just a useless memory hog that could easily be replaced in many instances by other technologies.
Where do these articles come from? To say that MySQL and Postgres are not viable replacements for Oracle and SQL Server shows exactly the ignorance that persists throughout the IT world, especially in upper management. These people, much like this author, simply form ideas based on ignorance. Because they are not familiar with MySQL, thus it is not equal. The reality is way too many IT managers have chosen SQL simply out of ignorance when they very well would have done much better with MySQL or Postgres.
But thats always been the case; take Sybase for example, in the TPC non-clustered benchmarks, up to 300GB, Sybase 12.x eats Oracle for lunch when running on Solaris x64 – and yet, we have people going for Oracle on Solaris. It doesn’t make sense.
But all has to do with, what I keep saying, the Oracle mystic; the same mystic that makes people buy Microsoft’s products; there is nothing exactly great about the product, they don’t actually *need* it as there are cheaper alternatives, but it makes them feel good that they’re using a tool from what they deem ‘the winning team’.
Its like Microsoft Office and collaboration tools as mentioned in the OpenOffice.org thread in a previous article; sure, they *could* collaborate using another method, but they think that because its included with Office, obvious that is the superior way, there for all office suites should do it the same way.
I’m not so sure that’s the main reason people turn to MS or Oracle products. I think the main reason companies turn to these products is simply because they base their marketing on:
a) compatibility, and
b) support.
While I’m not denying that other solutions are just as valid as MS / Oracle – when you have trained consultants, dedicated forums and hundreds of publications dedicated to the set up, maintenance and debugging of these systems, it does ease corporate minds that their product wont (shouldn’t) fail, regardless of how good they are in a market comparison.
True, but at the same time, alot of it is cluelessness; for example, I deployed a fleet of computers for a highschool; FreeBSD + GNOME + OpenOffice.org running ontop.
Skeptical at first, but once I set a mock up desktop, they moved around, loaded up OpenOffice.org, they realised they could do everything they needed, and it wouldn’t require them shelling out large amounts of cash for licences.
I think the issue is that management should be less quick to judge alternatives out there; and if there is an alternative, for christ sake, don’t use a consultancy firm, get your IS/IT staff to investigate, get the company to send out a free trial version, setup a stand alone server, and test it to see how successful it is.
Having seen some of the stupid things I’ve see management spend money on in the past – first class plane trips, expensive food and so forth, all on the company card, its a small price to pay to be a little more open minded on managements part.
Maybe, but if you know you have a massive resource base (expensive or not) to rely upon during setup or in case the worst happens – then it makes sense to go with that company. After all, a payroll server (for example) crashing could prove very very damaging to a company if no-one is there with specialise knowlage to suport the system.
I appretiate your example, but my comments were more directed at companies that offer top-end solutions for business critical systems. While I agree in princible with what you’ve achieved, it wouldn’t (in my opinion) be as practical to set up a payroll server (to use a previous example) “just to see what happens”.
We use Oracle for our HR and payroll systems and while I don’t particularly like the application itself, the support level for their systems is superb.
If one has no specialised knowledge to setup and maintain it, then it would be the perfect organisation for outsourcing, for example. In New Zealand, apart from small businesses with around 6 employees, I can’t remember a single one which actually do their own payroll – most of the time its handled by Pay Global and a few other companies; there are also some banks who are now offering payroll services as part of their business packages.
Who said that? keep the current system running, and setup another system as a *MIRROR* to the first one – damn, does *ANYONE* read. I didn’t say go out and simply just replace it, and pray for the best, I said, get/obtain/find ANOTHER server not being used, and load it up, and test it on a seperate server.
True, but my point is; its when people turn down the alternatives simply because it isn’t as ‘cool’ as what they’re used to; for me, I can float between 4 different office suites and complete the same task with ease – why? because firstly I know the basic fundamentals of office tasks, which can be transposed onto any number of office suites, and secondly, because ultimately my focus is on outcome, not whether my feng shui is inlinement or whether the GUI is bogged down in eye candy.
Thats the point I’m getting at; its not about the fact that a company *chooses* a certain application, but the method of which they came to that conclusion – “we need to use Office because it has this feature [feature]”, for example, whilst ignoring that OpenOffice.org/StarOffice can accomplish the same thing using a different method – then the compounded failure of asking whether the extra $200 they pay for Office is really worth doing something in that way; is the organisation going to get back that $200 in terms of productivity improvements through the use of the new feature vs. doing in the way one would on OpenOfice.org/StarOffice?
Edited 2007-04-02 22:57
Ease of use. Make something simple and powerful and people will migrate to it. Linux installs without much effort, while Solaris requires some effort (although it’s getting better). Microsoft tried to create a server that a reasonably trained capucin monkey can administer.
Cost is also a major driver. Your pointy-haried boss walks in and says we need a system to do X, which needs to be done today, and you have no budget. You’re going to pick a good, quick, easy to use tool like LAMP. I don’t know too many developers that will reach for Oracle and Oracle Application Server as the “quick and dirty” way of getting something done.
Eventually, that simple system that was developed to be thrown away morphs into an enterprise tool. Or what starts as a simple project to make blogging easier becomes good enough for corporate use.
I keep reading the name of this article as “It’s the Slack, Stupid!”
And from there, expecting my Discordian friends to jump out and sing unto me the praises of Eris…
If you could solve the problem with “AJAX” today you could have solved it with Javascript five years ago. How can this possibly be true? Five years ago there was no standardized method of making asynchronous HTTP calls. Any web application that relied on asynchronous data transfer in the last two to ten years was broken by design.
Given that Java does “AJAX” too one is left wondering what you really replaced the applet technology with. Frankly the statements made above lead me to believe “intense client side calculations” could be defined as the addition of two and two. Java can do AJAX too! zogmf! Alert the press!
So what was your major, marketing?
roflcopter