Is Open Source Insecure?

Roaring Penguin Software ^ | 6-10-2002 | David F. Skoll

[PatrioticAmerican]

"Oracle and Sun haven't taken the TPC benchmarks seriously for at least 4 years"

"Otherwise, the benchmark means nothing. "

Is that why they used the TPC benchmarks in their advertising, because it meant nothing?

[PatrioticAmerican]

Actually, I doubt Harr has ever designed a system that needed the 100,000+ TPMs easily achieved by SQL Server and Oracle. The whole point of speed is moot for his usage.

[Knitebane]

LOL! Microsoft wrote OS/2. I guess you think that MS Word exists because of "competition" from Notepad?

LOL! Get real!

You know, I often enjoy my jousts with Bush2000. He can make a concise, intelligent arguement over the various facets of the Microsoft/Open Source discussion. Granted, he's often poorly informed about various aspects of the debate, but that's largely due to his lack of exposure to the Open Source environment.

That's simply ignorance, and ignorance is curable through education.

Having not had the dubious pleasure of locking horns with you over these issues, I'm prepared to give you the benefit of the doubt and surmise that you too are ignorant.

Given your complete lack of knowledge of the history of OS/2 and the history of the Microsoft Word/WordPerfect battles, your clueless statement is either due to ignorance of the subject or due to stupidity.

Please go do some research on the subject before making such comments. Once you've actually been exposed to the facts of the matter, ignorance will no longer be an excuse and we can all just dismiss your rantings as the works of an idiot.

[Bush2000]

You know, I often enjoy my jousts with Bush2000. He can make a concise, intelligent arguement over the various facets of the Microsoft/Open Source discussion. Granted, he's often poorly informed about various aspects of the debate, but that's largely due to his lack of exposure to the Open Source environment. That's simply ignorance, and ignorance is curable through education.

Ouch. I'm hurt. ;-p

[Bush2000]

No, I'm a developer. If Oracle is better, I use it. I have no reason to "prove" it to anyone else. I don't like Oracle enough to "sell" it.

Let's rephrase that: You're a developer who's trying to sell the snakeoil that Oracle is better than SQL Server on equivalent hardware. I'm really getting bored with you, Harr. You really offer no challenge anymore. You make idiotic statements, I call you on them, and when you're asked to provide evidence to backup your statements, you ramble on about "not being a salesman", etc, when clearly, your comments are intended to promote your own tools. Whatever.

Do us both a favor and don't respond. I don't have any more time to waste on your nonsense.

[Bush2000]

Actually, I doubt Harr has ever designed a system that needed the 100,000+ TPMs easily achieved by SQL Server and Oracle. The whole point of speed is moot for his usage.

Hey, don't try to talk sense with the guy: He simply likes paying more for less.

[Knitebane]

See what happens when I try to compliment someone?

Ok, I've learned my lesson. I'll never say anything nice about you again. :P

[Knitebane]

Speaking of snakeoil...

...trying to sell the snakeoil that Oracle is better than SQL Server on equivalent hardware...

And where, pray tell, does one find hardware capable of running Windows that will compare to a Sun E10000, HP Superdome or IBM RS/6000?

Microsoft had the opportunity to run Windows on high-end hardware but killed it off after NT4.

The only way to get close to the performance of a real server is to cluster Windows servers, but there is a price to pay there too.

As any aircraft mechanic can tell you, doubling your pieces of hardware doubles your number of mechanical failures. And therein dies the "bang-for-your-buck" argument when you factor in having to support multiple Intel boxes versus one rock solid big iron box.

[PatrioticAmerican]

"Microsoft had the opportunity to run Windows on high-end hardware but killed it off after NT4."

ES7000, ring a bell? Considering the high-end capabilities of even a single pocessor compared to just 10 years ago, I'd say we have achieved a serious level of performance few will ever need. Just 5 years ago a 400Mhz processor and a 266Mhz bus was considered extreme computing. Today, we have 2.4+Ghz processors and 533Mhz busses. The SMP capabilities of the chipsets has even went ballistic compared to yesteryear. There will always be the needy application, but, for the most part, a dual or quad does nicely for most applications, and clustering achieves almost the rest.

[Knitebane]

The Unisys ES7000, while a quite powerful box in comparison to, say, a Dell Poweredge, is still just an Intel multi-processor machine, with all of the liabilities that that brings. It rates as a mid-range server and still not in the same league as a Sun E-box or a IBM RS/6000.

This kind of thinking is common among people who only do Windows. Those who have worked on only bittyboxes think that a really big bittybox is as good as it gets. If you can plug it into a standard wall socket, it's still just a PC.

And clustering, while effective up to a point, has it's own problems, as I mentioned earlier.

[PatrioticAmerican]

I have mainframe & mini in background with DoD. While the ES series isn't an an e-box, it is far more powerful than what 99.9% of applications require. Clustering adds a near infinite expandability and reliability. Even mainframes are clustered for reliability, especially geographically, and mainframes do crash.

[Knitebane]

While the ES series isn't an an e-box, it is far more powerful than what 99.9% of applications require.

How nice. I guess that's why IBM went out of business. Obviously all those mainframes had no use.

Oh, wait....

And the military is using Windows boxes for SOSUS, submarine flow simulations, stealth radar cross-section mapping and nuclear weapons simulations.

Oh, wait....

And isn't it nice how all of the universities were able to jettison the big iron from their research departments.

Oh, wait...

And remember when all of the big accounting firms went all Windows, and even Fortune 500 companies were able to do their accounting with PeachTree?

Oh, wait...

As an ex-military person, you should remember your basic training.

"This is a rifle. It is used for killing people. Please do not try to shoot a tank with it."

Maybe there should be boot camp for Windows weenies.

"This is a PC. It is used for desktop and mid-range server functions. Please do not try to do mainframe work with it."

Or perhaps I'm just being insensitive. Did you receive a head wound while on active duty?

[Bush2000]

How nice. I guess that's why IBM went out of business. Obviously all those mainframes had no use.

IBM's mainframe business is primarily to old guard companies who upgrade existing systems. They don't get a whole lot of new deployments of mainframes.

And the military is using Windows boxes for SOSUS, submarine flow simulations, stealth radar cross-section mapping and nuclear weapons simulations.

That's primarily because the technology is 30 years old. You can accomplish much of the same thing with good clustering. But that would mean rewriting the software -- something the feds have little inclination (and probably existing expertise) to do.

And isn't it nice how all of the universities were able to jettison the big iron from their research departments.

Please tell me you aren't using universities as a model for technology deployments.

And remember when all of the big accounting firms went all Windows, and even Fortune 500 companies were able to do their accounting with PeachTree?

None of the reasons you've provided obviates PatrioticAmerican's fundamental point: That clustered PCs can replace stodgy mainframes. Clustering is relatively new in the PC world. It hasn't had an opportunity to displace mainframes primarily because it hasn't existed previously. But that will change over time as the costs of big iron can't compete with the economy of scale of clusters. Don't get me wrong: Mainframes will not die. They will always have their place. But anyone who thinks that they can't be displaced is kidding himself. Today's clusters are far more capable and fault tolerant than you are giving them credit for.

[Bush2000]

Incidentally, I must say ... you're remarkably less strident than usual. If I didn't know better, I'd say you were almost ... well ... human now. We can almost carry on something resembling a rational debate. ;-p

[PatrioticAmerican]

The IBM mainframes are on their way out. Many have already been replaced. Most of the mainframes sold are to replace existing installations. Few are sold into new applications. Most minis have also been replaced.

If you think that the PC architecture hasn't evolved into a far more reliable box than the original 20 years ago, you might want to look at the hardware available, now. Just because a box isn't tied into a 480VAC MG doesn't mean that it isn't powerful. How much juice it sucks up is no indicator of it capabilities.

Even for mobile systems, laptops, of all things, have replaced the 370/390 systems used to process data at stationary installations. When a 390 had 300MIPS and 50GB storage, and now the average laptop has 1,000+MIPS and 60GB storage, what is the diff, except for reliability and failover, and we have that in server boxes.

P.S. IBM mainframes use clustering and node clustering, so I wouldn't try to pick on Itnel architecture for using it. :>

[ShadowAce]

That clustered PCs can replace stodgy mainframes.

Sorry. That is not correct. As a progammer with mainframe experience, I can tell you there is no way in the world a cluster of PCs is gonna replace a mainframe.

Wanna know a secret? The mainframe's power lies in its I/O--not its CPU. A cluster of PCs just isn't gonna match the I/O of a mainframe. A cluster of PCs just ain't gonna process 170 million transactions every night, with 100% uptime--and I don't care what OS you're running.

[Bush2000]

I don't want to disturb your illusion of beauty so we'll leave it at that.

[Dominic Harr]

Is that why they used the TPC benchmarks in their advertising, because it meant nothing?

When was that, 5 years ago?

Look it's obvious you have no DB experience now, so other than trying to sell me on a product you've never used, do you have anything else to add to this conversation?

[Dominic Harr]

Let's rephrase that: You're a developer who's trying to sell the snakeoil that Oracle is better than SQL Server on equivalent hardware.

Fact: You came in here trying to sell SQLServer.

Fact: You're claiming something that any experienced developer knows isn't true.

Fact: I called you out on it.

You lose. You did not make the sale.

[Demidog]

Apache actually works. Hope that helps.