Poll: Should Apple enter the high-end server market?

[MacRumors]

Vote: Poll: Should Apple enter the high-end server market? (http://www.macpolls.com/?poll_id=229&ref=forums.macrumors.com)

[macfreek57]

i say uncertain


because the mac is known for being a "user-friendly" machine (even though i would not qualify my use of the mac as a need for user-friendliness).

high-end server is a big concept for "the iMac company" and "the Aqua company"
to introduce as part of their line-up. people might tend (even if it's a small tendancy) to want to stray away from that.

on the other hand,
Apple is becoming more and more "power-user-friendly". that's good for big companies. if I were a big company i'd do my research and get the best machine. and Apple is right on their way to having the fastest server in the world.
just as long as it's not Blueberry

of course we all know Pixar would buy :p :p

[jkojima]

Yes, I feel that Apple should offer a high-end server. That's my gut reaction, anyway. I know little about the size and nature of the market for such machines, but if Apple does decide to go for it, I'm sure they will have done their homework.

One major barrier to entry is that they'd be going directly against well established incumbents like SGI and HP. That's a tall order.

It would surely be interesting to see if Apple does enter this market!!

[BigJayhawk]

i think Apple should definitely enter this market (if they have a DOMINANT PRODUCT). Even if this wasn't a great money-maker, it could serve the same purpose that the Dodge Viper serves for Dodge. We have this absolute COOLEST, FASTEST, BADEST sports car and if you buy our mini-van (which is in the same commercial, of course) you will sort of feel a part of the "Viper Experience."

The Viper is a bigger ad for OTHER Dodge Products than it is for itself.

[MrMacMan]

yes

[mac15]

It depends if high end buisness's want them. IF they big vendors want them then by all means yes do it but if they don't. Then some G5 Xserves will do

[macnews]

I said yes because I think it can fill a nitch the others haven't - primarily a nice GUI for a unix box/server.

I also said yes because I am hoping it will make Apple do a better job at there implementation of OSX server.

[macnews]

I said yes because I think it can fill a nitch the others haven't - primarily a nice GUI for a unix box/server.

I also said yes because I am hoping it will make Apple do a better job at there implementation of OSX server.

[bluedalmatian]

I wish Apple would also make an effort to enter the corporate server market.

I notice they never release sales figures for XServe....

[Wuddel]

I don't think they should enter this market. Selling Xserve and Xserve RAID is ok, because they complement the usage of a G5 (PM G5 + Xserve RAID) by "creative professionals".

I think the market does not need another Apple-product. Remember: there is already a Xserve cluster-configuration. A full rack of Xserve-nodes gives you a nice high end server with 84 CPUs.

Apple should concentrate on fewer hardware-products and make them better. (the inevitable Xserve G5)

[Lanbrown]

Originally posted by jkojima
Yes, I feel that Apple should offer a high-end server. That's my gut reaction, anyway. I know little about the size and nature of the market for such machines, but if Apple does decide to go for it, I'm sure they will have done their homework.

One major barrier to entry is that they'd be going directly against well established incumbents like SGI and HP. That's a tall order.

It would surely be interesting to see if Apple does enter this market!!

The real incumbents are Sun and IBM. SGI is a shadow of its former self and HP has three different platforms.

The Itanic is slow to take off, one major reason is software. Apple would face the same battle. The Power4 is the better chip for the job then the 970 is as well. There are huge margins at the high-end, but customers are also very picky. Besides major programming changes, you have to train as well. Solaris and AIX are heavily entrenched and trusted.

The R&D money could be better of spent on consumer devices.

[Lanbrown]

Originally posted by BigJayhawk
i think Apple should definitely enter this market (if they have a DOMINANT PRODUCT). Even if this wasn't a great money-maker, it could serve the same purpose that the Dodge Viper serves for Dodge. We have this absolute COOLEST, FASTEST, BADEST sports car and if you buy our mini-van (which is in the same commercial, of course) you will sort of feel a part of the "Viper Experience."

The Viper is a bigger ad for OTHER Dodge Products than it is for itself.

Look at how well they are doing financially.

[sedarby]

Absolutely not. OS X Server is no where near ready to compete against more mature OS's offered by SUN, IBM, etc.

[Freg3000]

I say no. Even though Apple has been met with much success recently, it can't just go out their and recklessly swing for the fences. They problably could afford a big screw up with 4 billion in cash, but why take the risk.

If Apple has a market here, then go for it I guess. I just don't see it. I love Apple and all, but seriously, if you wee a small business or cooperation, why would you want to go with Apple?

[idea_hamster]

I often hear people note that Apple thinks of itself as a hardware company first and a software company second. (Hence, the Marklar project is, what, real? not so much?) So, in that sense I think it's a bit odd to admittedly have a range of hardware that doesn't account for a major need of most (prospective) major clients.

However, maybe Apple's message is that their computers can run in conjuction with any high-end server system that you choose, so they've decided to focus on making the human-computer interface the best it can be.

Even so, I think that having a high-end server product -- even if it doesn't sell well -- would encourage small- to mid-sized companies embarking on the major computing step to see Apple as a top-to-bottom solution. Being able to tell people that they will never have to buy from another company if they don't want to is important, I think. Right now, buyers starting out buying networked computing solutions (maybe with unrealistic optimism) take note that an Xserve is not the last answer -- that they'll have to switch to something else eventually.

[wdlove]

I say yes, when Apple puts the G5 into the Xserve they will be well on the way!

[trebblekicked]

they are using a proc that would be effective in use as a high end server/render farm. they should investigate. like someone else said, the high end market is finickey (sp?). but apple does have friends in high places at ILM, Pixar, etc. Don't cancel the picnic before checking the weather.

[Lanbrown]

Originally posted by trebblekicked
they are using a proc that would be effective in use as a high end server/render farm. they should investigate. like someone else said, the high end market is finickey (sp?). but apple does have friends in high places at ILM, Pixar, etc. Don't cancel the picnic before checking the weather.

Pixar uses at least two platforms. One of which is a cluster setup. Apple wouldn't need to be in the high-end market to sell in the cluster portion. Just sell the X-Serve, the cluster version more specifically.

[Lanbrown]

Originally posted by idea_hamster
Even so, I think that having a high-end server product -- even if it doesn't sell well -- would encourage small- to mid-sized companies embarking on the major computing step to see Apple as a top-to-bottom solution. Being able to tell people that they will never have to buy from another company if they don't want to is important, I think. Right now, buyers starting out buying networked computing solutions (maybe with unrealistic optimism) take note that an Xserve is not the last answer -- that they'll have to switch to something else eventually.

What small to mid-size business uses a server with up to 64 processors? Large corporations do for back-end stuff. A system that can go up to the size is at least one million dollars.

Building a system that can handle two processors is harder then building one with one. A system with four is harder then one with two. Apple would spend a lot of money on R&D to build a system that can handle 64 processors. If they really wanted to get into that arena, they would be better off porting OS X over to the Power4 platform. Get OS X to run on it and sell rebadged IBM servers.

[Vonnie]

Let IBM do it, Apple has not a reputation in the high end server market, which is pretty important if you buy +10.000$ pieces of hardware.

[caveman_uk]

Apple's strong points are cool design, ease-of-use and quality. Sadly the first two don't really come into play when the machine's going to be stuffed in a back room somewhere. x86 running Linux on commodity hardware is much more cost efficient. Now what would be an option is an Apple cluster controller that seemlessly works in Linux clusters and allows people to easily create and manage tasks for the cluster. The Linux clustering stuff is really powerful but hardly user-friendly. Here is Apple's niche. As for what would a small business need with a 64-processor server. Where I work we have about 80 employees total. We have a 10 node Linux cluster used by the molecular modelling and bioinformatics guys (about half a dozen active users). 10 Nodes is nowhere near enough as jobs take days to complete.

[medea]

If they can come up with a system capable of outperforming the other big names in the field then I say go for it, a high-end G5 server could interest many.

[Lanbrown]

Originally posted by caveman_uk
As for what would a small business need with a 64-processor server. Where I work we have about 80 employees total. We have a 10 node Linux cluster used by the molecular modelling and bioinformatics guys (about half a dozen active users). 10 Nodes is nowhere near enough as jobs take days to complete.

Would your business pony up the millions it would cost for the system? You can get 64 processor machines now, yet your company hasn’t bought one.

[Lanbrown]

Originally posted by medea
If they can come up with a system capable of outperforming the other big names in the field then I say go for it, a high-end G5 server could interest many.

Performance doesn't mean everything. The Itanic is one of the best performing chips around, but yet it doesn't sell. Sun and IBM outsell it easily, as the Alpha and PA-RISC do and they are being killed off.

Buying a few extra processors for a box is cheap compared to changing code, testing and training people. Not to mention vendor relationships that have been established. You could spend a lot of money to get something that is faster, but what about down the road. You know what to expect with your current vendor. Compared to the unknown of the other. What happens if you switch a few years later, your previous vendor has a processor that just kills the competition. You could have upgraded with no code changes if you would have stayed. Define outperforming. Do you mean that SPEC scores that everyone uses? What about the chip the 970 is based of off, the Power4. The 970 has a faster clock speed then the Power4, but the Power4 is engineered to run at full utilization without failures, which is why it has a slower clock speed. Back to those pesky SPEC numbers. Most of those do not really test what the system is sued for, some of the SPEC tests do. Wait until Sun releases their new boxes. While the CPU will be relatively unchanged, the performance in REAL world applications will be much improved. There is more to performance then just the CPU.

OS X is not setup for 64 processors as BSD isn't either. There would have to be substantial changes to the OS to support 64 processors efficiently. Hardware design plays a major role as well. Sun can get over 98% efficiency on their systems, even in a clustered setup. The Pentium III dual processor machines got far lass then that. Basically you got 100% out of the first CPU, 50% out of the second, 25% out of the third, etc. The hardware wasn't designed for multi-CPU and neither was the OS. Both most work together and efficiently. Apple would have a long road ahead of them. Brute force from a processor doesn't necessarily mean much. As the adage says, works smarter, not harder.

Interest and sales are two different animals.

[ldjessee]

Hello,

I say they do not need to make a 'large server'.

Apple's strength would be to sell a whole rack of dual processor XServes, and have Rendevous aware software.

They even put distributed rendering in their programming interface, lots of cheap machines can be more flexible than one large machine.

Just my thoughts and opinions.

[j33pd0g]

Overall, I think big business owners are poor decision makers when it comes to technological purchases. Really. I think they will always go with what is cheaper. Apple will never be able to compete against cheaper. And I love them for that reason alone. You really do get what you pay for. Big business are just not economical, and lack the vision of lighter workload innovations. Apple will sell some big servers, but will gain no notoriety. They can enter the market just as long as they don't bank on it.

[Keno]

Impressive as they may be, the XServe is frowned on in our server room because it lacks the mundane feature of redundant power supplies.

It's just not OK to down the server because a single UPS flames out or circut loads need to be balanced.

Even the cheapo Dell 1650 1U rack server has redundant power.

Come on Apple, give us this basic server feature before you wander off in to 'high-end' land.

[scem0]

i am undecided - but I guess it couldn't hurt.

Well, it COULD, but I doubt it would.

Apple shouldn't concentrate on the high end market though, the money is with the consumers. And many consumers, including me, have trouble justifying spending more than $2,000 for a home computer.

Apple needs to be lowering prices. They should concentrate on that before all other things.

scem0

[xelterran]

I dont see why not.. maby they could use the Power5 in it!

[bousozoku]

I said absolutely not.

As difficult as it is to do somewhat difficult things that expensive server operating systems take for granted, Apple would need to focus in an area where ease of use has nothing to do with a GUI.

I would like to see them do it because only the non-UNIX world makes it easy to operate large machines, but I doubt that they would haev a clue.

[MrMacMan]

I say if apple can work on this and not screw it up, i believe they might have a shot at places that need high-end servers, but aren't commited to a company.

Apple would have to pull off quite a feat to get one up and running though, seriosly this is no joke. Dual processors are easy, clusters are easy.

A high end server is a totally different game.

[nslyax]

I don't know. I cant see why Apple would build massive servers that relatively few companies would need. Why not stick with the Xserve/Xserve RAID that they have now, and offer a nice discount on volume sales? This allows a small company to buy just a few units or a larger one to buy several and cluster them, and without forcing Apple to make several different types of server to meet everyones needs. If Apple decided to produce a server meant exclusively for video rendering or a similar task though, I can see that working out, since its one of Apples focused fields. Besides, don't you think IBM would be upset if Apple took an IBM chip and made a Unix based enterprise server, just like what IBM does?

My 2 cents, though I will be first to admit its not worth much.

[Sayhey]

If Apple really wants to expand into this business they should look to a merger/buyout with Sun. The old "Snapple" rumored alliance would have lots going for it. Not that I think this is really going to happen. McNealy and Jobs together? The size of rooms needed to handle those two egos staggers the mind. Just kidding, Steve, if your listening! For Apple to commit the kind of capital that would be needed to make it as a new business would also drain resources from the areas it needs to concentrate on.

[skirtley]

By always trying to produce the best product, We will see shared improvements in all products.

[iJed]

Would Apple not just be better off making Mac OS X Server run on a high-end IBM server and then rebranding it with a few extra options and selling it as their own? Why reinvent the wheel?

[yzedf]

Apple does not understand the customer service and support to do such a thing.

If they try, it will flop. Big IT types don't look for Apple.

[jefhatfield]

apple should stick with what they know best

and that is making home computing products for us and high end graphics boxes like the g5

[legion]

This would be a tremendous waste of money for Apple. High end servers are more than just hardware. There is actually a lot more money involved in support than hardware and Apple does not have even the beginnings of a company infrastructure to handle those demands.

On top of which, IBM has announced a low-end server with 4 970 chips running linux starting at $3500. Apple would never be able to compete.

A major part of the retail business is understanding where and who your customer is and I think that is where Apple tends to falter. They think they've got one type of customer underfoot and then they can branch out without doing the proper research on the new customer (while at the same time losing touch with the old customer).

As for the Xserve numbers and the RAID units, it so far has been a failure for Apple. There are so many design flaws (once again from not understanding the customers needs). For instance, just looking at the drives: ATA drives in a server and RAID?! the RAID unit is based on units of 7?! (btw, the correct answer is ultrawide scsi drives, and based on units of 5).

In conclusion, Apple should stick with front-end equipment. They should spend their time and money getting higher graphics on board (take a chunk out of SGI's market) and getting developers on board for 3-D modeling.