Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Rackmount Server Rumors Resurface
- Thread starter arn
- Start date
- Sort by reaction score
What would be the big advantages of a rack-mount Mac server over, say, a rack-mount Linux server? Seems to me the Linux server would have a much better price/performance ratio, and a dedicated server would nullify all the advantages that the Mac and the Mac OS have on the desktop. Who would want a rackmount server with low performance and a non-journalling filesystem from a company that historically has been a joke in the server arena? I just don't understand.
Alex
Alex
OS X Server
It's true that you get more bang for your buck with a Linux/Intel server, and I doubt Apple seriously expects their rackmounted server (if they release one) to be used as a serious Web server.
Where OS X Server really comes into it's own, however, is as a server for OS X clients - right now nothing else is satisfactory. For example, to my knowledge, you can't transfer files with long file names from an OS X client to an NT or Linux server - if anyone has cracked this, please let me know!
So schools, design studios or any other places where they have a lot of Mac clients are likely to be interested.
This may also be a sign that Apple intends to build on OS X's potential for clustering.
Strip out fancy graphics cards, firewire ports, and Superdrives and Apple might actually achieve a decent price on this one!
It's true that you get more bang for your buck with a Linux/Intel server, and I doubt Apple seriously expects their rackmounted server (if they release one) to be used as a serious Web server.
Where OS X Server really comes into it's own, however, is as a server for OS X clients - right now nothing else is satisfactory. For example, to my knowledge, you can't transfer files with long file names from an OS X client to an NT or Linux server - if anyone has cracked this, please let me know!
So schools, design studios or any other places where they have a lot of Mac clients are likely to be interested.
This may also be a sign that Apple intends to build on OS X's potential for clustering.
Strip out fancy graphics cards, firewire ports, and Superdrives and Apple might actually achieve a decent price on this one!
Macs...
...kick **** for clustering. My friend and I (both 16) set up a cluster in half an hour, 20 minutes of that was getting the right version of carbonlib and OS8 for the powerbook. Basically a rackmount G4 would have ram, a hard drive, 1/2 processors, and a gigabit ethernet port. Nothing else (maybe an optional ADC port for hooking a monitor to it). I bet you could fit it in a 1U case pretty easily, and it would be MUCH cheaper than the current G4s. ([all these are guesses]-$500 for the superdrive, -$100 for the case, -$100 for the simplified motherboard [no FW, USB, PCI])
...kick **** for clustering. My friend and I (both 16) set up a cluster in half an hour, 20 minutes of that was getting the right version of carbonlib and OS8 for the powerbook. Basically a rackmount G4 would have ram, a hard drive, 1/2 processors, and a gigabit ethernet port. Nothing else (maybe an optional ADC port for hooking a monitor to it). I bet you could fit it in a 1U case pretty easily, and it would be MUCH cheaper than the current G4s. ([all these are guesses]-$500 for the superdrive, -$100 for the case, -$100 for the simplified motherboard [no FW, USB, PCI])
as alluded to earlier, a rackmount server by Apple would be truly a godsend to Schools looking to have Apple "workstations" and a viable server.
Doesn't that seem like such a dirty word for a Mac??
Doesn't that seem like such a dirty word for a Mac??
...but what if the eBay item was actually a rack-mount server board, and the "new" connector was actually just the same VGA output connector that the iBook has? As such, it would allow monitor connections for "headless" servers....
Let's not go there again...it's been done to death. Besides, the debate is if they are going to do it, not if it would be better than "x" server.
I think with their forays into 3D development lately (Nothing Real, etc.) it makes sense for them to provide machines that are rack mountable for server farms/ render farms. Bring it on Apple!
Whether or not a rackmount Mac server would be better than "x" server will directly influence whether or not Apple would release one, so that's why I think the pluses and minuses of a rackmount Mac server deserve debate. You don't want to see Apple release an utter failure, do you? It's just that the idea of a rackmount Mac makes me nervous, because I know that Apple is no stranger to utter failures...
Alex
How about this new mobo, then? Does it quell any of your fears about Apple releasing about an "utter failure"?
I think this is a form factor issue first, performance issue second. Once they get the design down for a 4U, or whatever, chassis they will have already made an inroad in situations where clustering/stacking is a key issue.
I don't want Apple to put out a rackmount version of their server just for the hey of it. It had better be a great performer with substantial new features in order for people with x86 based systems to go for it. I think DDR is one step in that direction...as long as they're not still playing catch up.
I haven't read up on the leaked motherboard yet. New motherboards are great, but then again the competition comes out with new motherboards all the time too, so...
Whether they can release something competitive is indeed the big question. I feel in order for Apple to succeed in the rack-mount server space, they must A:
- Make it cost-effective
- Make it low-power (which affects TCO in #1)
- Make sure it performs fairly well
- Make sure it is well-marketed and not seen as another "uh oh, here they go again" type move.
OR, they must B:
- Make sure it is new and different in some way which would cause those in the market for rack-mount servers to switch over, regardless of cost or performance.
I think B seems like a more logical choice for Apple... they do very well when they find their niche, exploit it, and slowly grow from there. I feel that Apple would get creamed if they tried to do A.
And it depends on which server market they're aiming for... they may be better off building better I/O into their systems than gunning for mad MHz. Sun is making a killing off their 500MHz $999 blade, just because it offers better I/O performance than any PC... so perhaps Apple could use something like that - a weakness of x86 - to their advantage.
Alex