| Friday, July 12, 2002 |
12:15 - Fuzzy Blue-Green Numbers
http://www.apple.com/xserve/performance.html
|
(top)  |
The competitor servers listed at Apple's Xserve specs page are from Dell, IBM, and Sun. It seems clear to me, at least at a cursory glance, that the reason for this choice was to pit the Xserve against the most comparable equipment from the most high-profile manufacturers in the business. Dell and IBM, for instance, are the only server makers that my company will buy boxes from; we've standardized on the PowerEdge line, and somehow I don't think we're statistical outliers.
But the sticky point is that Dell's 1U machine, the PowerEdge 1650, in its most maxed-out configuration has non-DDR SDRAM and comes to about $4300. The Sun Fire V100 also only supports SDRAM, and it's a lot more expensive too. I haven't checked the IBM x330, but I'd wager that it too doesn't do DDR, and probably comes to a similar price point as the Dell does.
Strictly speaking, the only points on which the Xserve (which in its tested configuration comes to about $6000, or $5000 if you were to outfit it with the smaller 60GB drives, which they have no reason not to do for the purposes of these tests) competes with the Dell and IBM servers is their 1U profile. The Xserve is really a 1U unit competing against 2U units from these companies, and with 1U units from third-parties who produce a better package than Dell and IBM do.
For instance, as has been brought to my attention in an e-mail discussion, APPRO makes a 1U Xeon-based server with four drive bays and support for a whole lotta DDR SDRAM. (No easily accessible price, though.) And PogoLinux makes a similar box based on AMD Thunderbirds that will run you about $4500, including 1GB of DDR SDRAM.
It seems reasonable to me that Apple should not have configured their entire competition lineup from machines like these, though they do match the Xserve's specs more closely; their purpose in positioning the Xserve is to compete with the status quo of Dell and IBM, and who would have guessed that these companies' offerings would be significantly less studly than those from relative unknowns?
But would it have killed them to have included one of these kinds of boxes in their tests?
Maybe it would have, and that's why they didn't do it.
The trouble with Apple-- okay, yes, I'm boiling down a lot of different concepts lately into "the trouble with Apple"-- is that they're falling prey to the old Dilbertian admonishment that "one lie is always more convincing than two lies":
COW-ORKER: Why weren't you at the meeting? I left you a message.
DILBERT: I didn't get any message.
COW-ORKER: Oh...
DILBERT: And when I tried to call you back, you weren't there.
...Or at least a variation on this. See, Apple has two tiers to its marketing strategy: the products themselves, and their PR department. One or the other of these would make an effective sales plan. But both together can result in trouble.
Most companies' marketing departments are weaselly. That's sort of expected. It's generally assumed that if a company is saying anything more grandiose about their products than simply listing the specs, there's probably a lie in there somewhere. Microsoft's touting of the handwriting recognition functionality in their Tablet PC as "the best in the industry", for instance, is blatantly untrue-- especially considering that they spent their pre-bake-off time educating David Coursey and others about how "handwriting recognition doesn't matter". But... people forgive them that, beacause people expect the products to be shoddy and the marketing to be weaselly. It's a stable system. It meets our expectations. There's no cognitive dissonance, and we buy the products.
But Apple's problem is that the products themselves, if shorn of all their PR, if placed in a darkened room with their direct competitors and tested by an unbiased subject in isolation from societal pressures, are in fact better. Any open-minded tester would find the user experience on the Mac to be more satisfying. This is widely accepted truth even among my PC-using friends.
But on top of that, we have Apple's legendary PR machine. Steve Jobs is still the master showman; his speeches at MacWorld cause the audience to scream and swoon like girls at an Elvis concert. When he smirks and tells us that a machine is "insanely great", or that it blows away the competition in performance specs, we believe him-- but to PC users, it sounds ludicrous. They're used to weaselly marketing. They're used to being told lies, and assuming that whatever those lies are about is probably nowhere near as great as it looks in the magazine photos. So they assume that the more grandiose Jobs' claims, and the more the apple.com website gushes, the more crappy Apple's actual products must be.
The products themselves, in an ideal world, would sell themselves. There would be no need for marketing. And in a world where Apple's products really did suck, the weaselly marketing would be looked at as par for the course.
But the two in combination make a volatile brew.
Am I saying that Jobs is Apple's worst enemy? Well, in part-- that's the conclusion that the board of directors reached in the late 80s, when they forced Jobs out. He was just too idealistic, too blinded by his own vision, and he was hurting sales. So Apple marketing from then on, until Steve's return, was fairly demure and benign-- and yet that didn't save the company from continuing to lose market share.
Does this invalidate my point? I don't think so, because what I'm talking about is a new, post-Jobs-return phenomenon. Ever since he's come back, the new product announcements and MacWorld keynotes have been showpieces like they were of old, and the website has become such a fairyland of promise and light. Before Jobs' return, the products were lackluster too, and failing to capture the imagination of the public is what really made Apple dwindle in the last decade.
That's not a problem anymore. Apple design makes waves again. People are proud to put the decals on their cars. Apple's market share is growing again, the company is turning a profit and expanding its physical presence back into its old buildings-- they're healthy again, and ready to send out new exploratory feelers, like the Xserve.
But when the products themselves are good, and the marketing is grandiose, it unsettles customers. They can't see an obvious small lie, so they assume there must be an even bigger, hidden lie. Nothing can be that good... can it?
No, I'm not going to claim that Apple's products are every bit as good as Jobs' rhetoric claims it to be. But it is damned good. OS X does in fact kick ass. iTunes is, dare I say it, insanely great. iMovie and iPhoto are excellent tools that not only allow people to do things, but that invite them to. Jaguar is the result of uncompromising effort from some of the best engineers in the industry, and when it's out we'll have a multithreaded Finder, network services that configure themselves through Rendezvous, handwriting recognition in any app, the spring-loaded folders that everyone's been clamoring for, and even redone interface widgets that indicate that someone looked at Aqua and said, "No, it doesn't quite look good enough yet".
Apple has a whole lot going for it. It's a company that does more work for less thanks than almost anybody else in the industry. And their products do in fact display superiority in a whole lot of ways. But, unfortunately, because of the grandiose marketing-- which is just a corporate extension of the evangelism that people like me feel compelled to do by way of spreading the word-- those benefits of their products may as well be null and void.
I'm the target audience of the keynote speeches and the marketing websites. I eat it up. I relish it. It's written by people like me, for people like me-- and so naturally I can't resist basking in it a bit.
But Apple does need to realize how they appear to the public at large; what successfully sells new Macs to existing Apple fans is not going to have the same success selling to the Other Side. The "Switch" ads are a step in that direction, but...
I'm operating from one simple basic principle, one with which I realize others may not agree: Apple having financial success and gaining market share is a good thing. From that axiom, it's logical for me to try to defend Apple's decisions, to try to dispel FUD and slander about them, to try to figure out new ways to pitch the company to those who wouldn't otherwise consider it. Right now, that means looking at stuff like the Xserve specs and giving Apple the benefit of the doubt as far as fairness against the competition goes; I want to believe their tests were conducted fairly, and that their choices in competitor hardware were informed by realistic analytical goals.
This may not be the case. Maybe Apple is being weaselly about it, and we won't know that until we start seeing some independent anecdotal numbers.
But I'm of the mind that regardless of how the numbers are sliced, Apple has produced a very compelling entry into a very demanding market, and I think it would be a shame to see it dismissed out of hand because the numbers don't appear trustworthy.
UPDATE: As I'd kinda hoped he would, Paul Summers weighs in with comments on the competitive boxes mentioned above.
Just a quick point-- the reason Apple didn't include any boxes such as the two you note from Pogo Linux and APPRO, is because they're not comparable, nor are the competition.
Firstoff, the hardware-- the pogo linux box comes to $4,588 in an XServe-like configuration. BUT, you can only put three IDE drives in it (due to the fact that they are using cheap ass on board IDE controllers, and the cdrom eats one channel up-- whereas Apple has a separate controller for each device). You can only get one GB ethernet port, to boot.
Upgrading that box to four SCSI drives brings the total to $6805, and it's still not as feature-packed as the XServe. This of course, ignores the fact that you get a piece of shit OS in place of OS X.
And, a final point on that ridiculous box. Four 10k rpm drives, and two athlon processors, in a 1U case? Remember how hot ONE 10k rpm drive got in lionking.org, which was a full ATX case? Fill a rack with those things, and not only will you be sucking enough power to light up Chicago, you'll be calling the fire department in short order. Hell, the boxes on top of the stack would probably get so hot that the solder on the boards would melt, to say nothing about thermal failure of the drives.
The APPRO thing is even worse. Same lack of 5 IDE device support. Same heat issues with two high voltage procs and four hot drives. Same lack of dual gigabit. Same lack of three PCI slots. Same lack of an AGP slot. The list goes on.
I've said it before, and I'll say it again. Computers are tools, and the measure of success of a tool is how well it -works-, not the raw specs.
You can fill an entire rack with XServes and use far less power, generate far less heat, and as a result of the reduced heat, have a much longer MTBF on the boxes and their parts. This of course, is with all of the things the competition doesn't have. Dual Gigabit. Firewire. Four IDE drives AND a cdrom. A slide-open case. A smaller, cooler power source.
And, let us not forget the primary advantage. OS X and it's bloody -insane- remote management apps. Unfortunately, as is always the case with Apple, some people will always overlook the simply amazing engineering in their products, and assume the raw numbers are the end all be all.
...The tweakers [e.g. APPRO and PogoLinux] aren't competition, as they don't have any (or very little) accountability or support. That may be all fine and good for a linux geek who likes to sit on the sidelines and quote numbers-- but no corporation with an actual IT department is ever going to buy such machines.
Whereas they will from Dell and IBM, and hopefully Apple.
Not that this really addresses the problem-- it still looks as though these machines are more closely competitive with the Xserve than the Dells and IBMs are, and I'd have liked to see at least lip service or some acknowledgment of the weird fact that they out-compete the Big Boys. But it does get some more facts thrown onto the table.
|
|
