Apple Discontinues Xserve, Only Available Until January 31st

[crackpip]

Have you ever even worked with servers or seen rank servers up and running.

MacMini zero redundancy (no go automatically) add in very little horse power and ineffective use of space.. Mac pro are a waste of space and power. Very little redundancy to them.
They would need multile Mac Pro acting as a server in any large scale deployment so waste of a lot of space and power.

It is pretty clear you do not know what you are talking about. KnightX works in the field understand this stuff.

OK let me use short sentences with small words to make myself clear, because you evidently have some trouble with reading comprehension.

OSXS bad.

For real computing, XServe cost too much and OSXS bad.

For big deployment, XServe cost too much and no good admin tools and OSXS bad.


Now for the big words:
Obviously, the situation may be more nuanced for smaller specialty deployments. But, for large deployments, I have never felt that there is a compelling reason to use Apple products in the server space.

What about larger data storage? Are you telling me that admins are chaining together XServes for data storage and redundancy purposes instead of dedicated RAID arrays?

In cluster environments, admins often customize the kernel to optimize for hardware (e.g. low latency interconnects, etc) and topology. Does anyone do this for OSXS? No.

Do we know of any large deployments at Apple that uses Apple server hardware and OSXS?

You're right, though, I do not maintain a large deployment of Mac's. I do research requiring high performance computing. I use massively parallel clusters, and I have allocations on 4 clusters that are in the top 100 to run my own source code. None of the machines in the latest top 500 use Apple hardware, and even when Apple hardware was in use, they didn't use OSXS-- they used Linux. I also maintain a small deployment for development and teaching. The clusters we use are not Apple, but the servers and workstations are. Though I love OSX client, I am not happy with OSXS. I can't imagine the nightmare of trying to use it in a large deployment. God help you if you try to upgrade the server to the latest 10.X release before the version hits 10.X.4.

crackpip

 

[glyaflya]

Think about the new data center in NC

At first I was puzzled by the decision to cancel the server hardware, but upon more thought I realize Apple is investing in a huge data center in North Carolina. This isn't just for MobileMe customers.

My sense is Jobs & Co see enterprise storage moving from local computer rooms to the cloud, to data centers. If that's so, why spend billions designing and supporting small form factor, proprietary hardware just for data centers? Buy'em from IBM, HP etc.

 

[chatin]

The cloud belongs to Apple not some dude with an Xserve.

The elegance of the design was lost in basement equipment rooms or co-locations.

I installed a few but never where they could be seen or admired. The reason: iNoise!

 

[Stella]

At first I was puzzled by the decision to cancel the server hardware, but upon more thought I realize Apple is investing in a huge data center in North Carolina. This isn't just for MobileMe customers.

What business is going to trust Apple, especially after this? Your very small business might, everyone else is going to steer very clear of any Apple "Enterprise" solution... Apple could / would discontinue any service in an instant.

 

[crackpip]

What business is going to trust Apple, especially after this? Your very small business might, everyone else is going to steer very clear of any Apple "Enterprise" solution... Apple could / would discontinue any service in an instant.

Do you know anyone who trusted Apple for a medium to large deployment of servers prior to this announcement?

 

[KnightWRX]

OSXS bad.

For real computing, XServe cost too much and OSXS bad.

For big deployment, XServe cost too much and no good admin tools and OSXS bad.

You can repeat that ad nauseum all you want, but it doesn't make it true. OS X Server is a niche, made to manage big OS X deployments with great tools for client management. It also serves as a great platform for studios that run off Final Cut Studio.

The fact is, shops that depend on both these functions are now left without any real data center grade hardware to run it off of.

 

[crackpip]

You can repeat that ad nauseum all you want, but it doesn't make it true. OS X Server is a niche, made to manage big OS X deployments with great tools for client management. It also serves as a great platform for studios that run off Final Cut Studio.

The fact is, shops that depend on both these functions are now left without any real data center grade hardware to run it off of.

But that is my point, it's not meant to manage large OSX deployments. It is a niche product meant for small OSX deployments. What specific services do you require from OSXS?

crackpip

 

[KnightWRX]

But that is my point, it's not meant to manage large OSX deployments. It is a niche product meant for small OSX deployments. What specific services do you require from OSXS?

I don't, I don't have OS X clients nor do I work for a video editing firm. But I can see these people needing things like Netboot, Update Server, NetInstall, NetRestore.

We have equivalents for all those and wouldn't be able to manage our 20,000 something Windows clients without all this automation.

Managing 10-20 PCs doesn't really require these tools, managing upwards of 1,000 does. And you don't run the infrastructure required for such a deployment off a Mac Pro.

Not to mention, again, Final Cut Server. How do you even run that without OS X Server and the Xserve ? Ridiculous.

 

[yadmonkey]

At first I was puzzled by the decision to cancel the server hardware, but upon more thought I realize Apple is investing in a huge data center in North Carolina. This isn't just for MobileMe customers.

My sense is Jobs & Co see enterprise storage moving from local computer rooms to the cloud, to data centers. If that's so, why spend billions designing and supporting small form factor, proprietary hardware just for data centers? Buy'em from IBM, HP etc.

Why do people who don't know what servers are for keep posting here?

The graphic artists and video editors I support use their Xserve/RAID combo to serve/share massive files very quickly over a fast network. Cloud technology is nowhere near fast enough or reliable enough to handle this duty.

 

[crackpip]

I don't, I don't have OS X clients nor do I work for a video editing firm. But I can see these people needing things like Netboot, Update Server, NetInstall, NetRestore.

We have equivalents for all those and wouldn't be able to manage our 20,000 something Windows clients without all this automation.

This is what I'm talking about. The processes you are talking about are most certainly critical. The tools that you use by MS and others are incredibly good. The problem is that Apple's implementation is half-assed and finicky. Maintaining clients requires an inordinate amount of manual interaction with the clients. Take a look through some of the online discussions on this. A place that manages 20,000 mac clients would be using the same server infrastructure that you probably do. Exchange, Active Directory, etc. They wouldn't be deploying Xserves.

crackpip

 

[KnightWRX]

This is what I'm talking about. The processes you are talking about are most certainly critical. The tools that you use by MS and others are incredibly good. The problem is that Apple's implementation is half-assed and finicky. Maintaining clients requires an inordinate amount of manual interaction with the clients. Take a look through some of the online discussions on this. A place that manages 20,000 mac clients would be using the same server infrastructure that you probably do. Exchange, Active Directory, etc. They wouldn't be deploying Xserves.

crackpip

WSUS doesn't work with Mac clients I'm afraid, nor does Tivoli. I never mentionned the Mail server or Opendirectory either so I don't see where you got Exchange and Active Directory from and again, you have failed to even start addressing Final Cut Server.

Face it, you're just an apologist. The fact is, there's no migration path for people that are using this infrastructure and Apple just made the biggest mistake of their history in the enterprise sector.

 

[crackpip]

WSUS doesn't work with Mac clients I'm afraid, nor does Tivoli. I never mentionned the Mail server or Opendirectory either so I don't see where you got Exchange and Active Directory from and again, you have failed to even start addressing Final Cut Server.

Face it, you're just an apologist. The fact is, there's no migration path for people that are using this infrastructure and Apple just made the biggest mistake of their history in the enterprise sector.

20,000 clients do not need to access Final Cut server. Final Cut server installs are likely much smaller deployments.

I brought up Exchange and Active Directory as examples of services that a large institution would deploy instead of Apple's offerings.

I'm saying Apple's server and remote admin products are **** especially for large deployments. I'm saying they have never given a **** about enterprise and this latest announcement is nothing new. How is that being an Apple apologist?

crackpip

 

[Hal Itosis]

Telling any server team that you're discontinuing support for a platform with only 2 month's notice and no upgrade path will lose you any chance you have of ever coming back to their server room.

You're overhyperventilating again. Try actually reading post #1 this time. Those who so desire have two months in which to purchase new units. Standard support goes up to a year, and optional support for three years total.

Even if they do announce it at this point, they failed. You don't leave enterprise customers hanging and don't EOL products on a Friday without anouncing a migration path.

To which "enterprise" customers do you refer? (i.e., name them). Or perhaps i should say: define "enterprise".


Apple's role in the server room has always been niche at best. This is simply part of an inevitable cultural shift, which will allow them to focus more on the client side. [it's much easier to see that when not perched on such an elevated soapbox.]

 

[Rodimus Prime]

20,000 clients do not need to access Final Cut server. Final Cut server installs are likely much smaller deployments.

I brought up Exchange and Active Directory as examples of services that a large institution would deploy instead of Apple's offerings.

I'm saying Apple's server and remote admin products are **** especially for large deployments. I'm saying they have never given a **** about enterprise and this latest announcement is nothing new. How is that being an Apple apologist?

crackpip

Umm I know of some deployments of FC servers that have well over a 100 people who need access to them. Clearly you are showing your faults here. Once you cross about 20-30 computers you really are going to be going to server grade stuff. It just is no longer cost effect not to have a sever handle all those processes.

Break a 100 you almost have to have them. Crack 1000 it is a requirement. You also have companies that will put all those servers off to a 3rd party. That 3rd party might have 20-30 smaller clients and they manage all the software. So they will need those servers.

Apple FC server is also SOL here.

 

[KnightWRX]

20,000 clients do not need to access Final Cut server. Final Cut server installs are likely much smaller deployments.

I brought up Exchange and Active Directory as examples of services that a large institution would deploy instead of Apple's offerings.

Now you've just shown how much you don't know about this topic.

I can run 20,000 clients off of about 15 VMs for Exchange. However, a render farm for 10 simultanous projects of video editing is probably going to require about 50 times the processing capacity.

Apple and oranges really. Final Cut Server doesn't just serve small text files from a database like Exchange does (yes, even with attachments, a 8 MB file served off of Exchange's EDB backend is nothing).

 

[alexkhan2000]

The Server Market

Let's look at the big picture of global server market. I used to follow the enterprise server market in the late-90's because that's where the hot growth used to be and I had a small number of shares of Sun stock (which are absolutely worthless now).

It had been a while since I checked on this industry and I knew the server market was bad but I didn't know it was quite this bad.

The hardware server market is a sinking ship that has been completely commoditized. The margins must be horrible. Apple's share as part of "Others" in this chart doesn't even figure to be 1% based on Apple's breakout of their revenues by product categories.

Apple's Mac business is growing at 25%+ rate and is now worth $22 billion (which would rank as #65 on Fortune 500 as Tim Cook recently pointed out), which is more than half of the entire server market which is declining at nearly 20% rate. Why should Apple invest and put their resources into this market? It's no wonder that the likes of IBM, HP and Dell are focusing on things like storage, networking equipment, software, and services like consulting and systems integration. They're not making enough money selling servers and there's no growth there.

Who still thinks that Apple should have bought Sun? There was a time when Sun was nearly $20 billion in annual revenues and had a market cap of over $120 billion during the dot com boom about a decade ago. After the dot com bubble burst, Sun went through the decade losing billions and got snatched up by Oracle for the bargain basement price of around $7 billion and now the revenues are below $4 billion and sinking fast.

Does anyone here truly think that Apple can compete in this space with razor thin margins just to gain a few market share points in the "Others" category in a shrinking market?

 

[Rodimus Prime]

Let's look at the big picture of global server market. I used to follow the enterprise server market in the late-90's because that's where the hot growth used to be and I had a small number of shares of Sun stock (which are absolutely worthless now).

It had been a while since I checked on this industry and I knew the server market was bad but I didn't know it was quite this bad.

The hardware server market is a sinking ship that has been completely commoditized. The margins must be horrible. Apple's share as part of "Others" in this chart doesn't even figure to be 1% based on Apple's breakout of their revenues by product categories.

Apple's Mac business is growing at 25%+ rate and is now worth $22 billion (which would rank as #65 on Fortune 500 as Tim Cook recently pointed out), which is more than half of the entire server market which is declining at nearly 20% rate. Why should Apple invest and put their resources into this market? It's no wonder that the likes of IBM, HP and Dell are focusing on things like storage, networking equipment, software, and services like consulting and systems integration. They're not making enough money selling servers and there's no growth there.

Who still thinks that Apple should have bought Sun? There was a time when Sun was nearly $20 billion in annual revenues and had a market cap of over $120 billion during the dot com boom about a decade ago. After the dot com bubble burst, Sun went through the decade losing billions and got snatched up by Oracle for the bargain basement price of around $7 billion and now the revenues are below $4 billion and sinking fast.

Does anyone here truly think that Apple can compete in this space with razor thin margins just to gain a few market share points in the "Others" category in a shrinking market?

problem I see with that number at 20% is economy crashed happened. Budgets were cut big time and IT was often top of the list in terms of hardware. Instead of doing planned replacements they were put on hold. I know some servers from someone I know personally he had to put on hold several upgrades due to lack of funds. Server upgrades and replacement needed but he knew he could push it another year or 2. But it going to make it tight. When things kick back up I bet you see a huge growth in servers as they are going to rushing to replace the aging equipment that had to be put on hold.

 

[lilo777]

Let's look at the big picture of global server market. I used to follow the enterprise server market in the late-90's because that's where the hot growth used to be and I had a small number of shares of Sun stock (which are absolutely worthless now).

It had been a while since I checked on this industry and I knew the server market was bad but I didn't know it was quite this bad.

The hardware server market is a sinking ship that has been completely commoditized. The margins must be horrible. Apple's share as part of "Others" in this chart doesn't even figure to be 1% based on Apple's breakout of their revenues by product categories.

Apple's Mac business is growing at 25%+ rate and is now worth $22 billion (which would rank as #65 on Fortune 500 as Tim Cook recently pointed out), which is more than half of the entire server market which is declining at nearly 20% rate. Why should Apple invest and put their resources into this market? It's no wonder that the likes of IBM, HP and Dell are focusing on things like storage, networking equipment, software, and services like consulting and systems integration. They're not making enough money selling servers and there's no growth there.

Who still thinks that Apple should have bought Sun? There was a time when Sun was nearly $20 billion in annual revenues and had a market cap of over $120 billion during the dot com boom about a decade ago. After the dot com bubble burst, Sun went through the decade losing billions and got snatched up by Oracle for the bargain basement price of around $7 billion and now the revenues are below $4 billion and sinking fast.

Does anyone here truly think that Apple can compete in this space with razor thin margins just to gain a few market share points in the "Others" category in a shrinking market?

Nobody is denying these numbers. However, the important thing about enterprise market is to be able to provide comprehensive solutions. Admittedly, Apple could not do it even before but now it is possible that some companies that might otherwise buy Mac Pros would start looking elsewhere. I guess somewhat valid analogy could be iPods and iTunes. Not that Apple makes great money from iTunes but imagine what would happened if they stopped supporting it.

 

[alexkhan2000]

problem I see with that number at 20% is economy crashed happened. Budgets were cut big time and IT was often top of the list in terms of hardware. Instead of doing planned replacements they were put on hold. I know some servers from someone I know personally he had to put on hold several upgrades due to lack of funds. Server upgrades and replacement needed but he knew he could push it another year or 2. But it going to make it tight. When things kick back up I bet you see a huge growth in servers as they are going to rushing to replace the aging equipment that had to be put on hold.

This trend has been happening in the server market for years before the recession. The server market has become commoditized with increasingly thin margins due to low-end Wintel and Linux blade servers pushing up from the bottom. This is why IBM is focusing on services that have recurring revenues and why HP acquired EDS to compete with IBM in that arena. It's also why HP and Dell went in a bidding war for a little-known storage technology company.

Sure the economy will pick up someday but the likes of IBM and HP are entrenched on the high-end with multi-million dollar servers and mainframes and they also dominate the cheap blade server segment with Wintel and Linux along with Dell. Sun is virtually finished and I don't see how Oracle could right the Sun ship as they have absolutely no experience with hardware.

Sun was a vertically integrated server maker with their own SPARC CPU, their own OS with Solaris, Java which they developed but was forced to give away, their own server designs and various middleware offerings. Due to Sun's relatively small volume, they didn't have the economy of scale to compete with the the super cheap Wintel and Linux servers and they couldn't compete with the high-end servers from IBM and HP who had the resources to develop offerings that far outperformed what Sun could come up with.

Since this market has become commoditized and is shrinking, everyone in this sector is losing money or barely breaking even. What value proposition does Apple have to compete here? Apple is growing the phone business at 91% rate, the iPad is the fastest selling consumer electronics product in history, and the Mac is growing at 27% rate despite the crummy economy. What can Apple do with a server that the likes of IBM, HP and Dell can't? Why will IT departments pay the Apple premium when these entrenched and very experienced big boys of the enterprise market can do it much cheaper and much better?

I believe Apple has a very good shot at becoming big in the enterprise but as the premier provider of thin mobile clients - specifically the iPhone and the iPad. That's what the whole Unisys deal was about. Apple hired Unisys so that they can integrate these devices into the IT infrastructure of companies like Bank of America, Citicorp, Disney, Hyatt, and various other Fortune 500 companies. The money for enterprise computing is in software, consulting and systems integration, not the hardware.

Nobody is denying these numbers. However, the important thing about enterprise market is to be able to provide comprehensive solutions. Admittedly, Apple could not do it even before but now it is possible that some companies that might otherwise buy Mac Pros would start looking elsewhere. I guess somewhat valid analogy could be iPods and iTunes. Not that Apple makes great money from iTunes but imagine what would happened if they stopped supporting it.

If you look at Apple's financials, the server/enterprise segment isn't even mentioned nor does it seem like a blip on Apple's radar as far as revenues and profits are concerned. This segment just doesn't fit in with what Apple does - providing powerful and stylish mobile clients that access a rich ecosystem of content, info, apps, and various services on the cloud.

Apple can manage their own cloud (like the NC data center and others that I'm sure they'll build around the world), but they don't need to provide the gear that go into these data centers. For their own internal data center needs, Apple uses a hodge-podge of gear and software from IBM/AIX, Sun/Solaris, Linux, SAP, Oracle databases, and even Windows. The Xserve and the OS X Server were never meant for running multi-billion dollar operations.

I posted this on another thread and I'll post here again. Here are some typical IT job postings on Apple's site for both their HQ in Cupertino and the data center in NC:

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=57374&CurrentPage=8

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=62314&CurrentPage=8

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=62312&CurrentPage=8

For SAP specialists:

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=60099&CurrentPage=6

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=62489&CurrentPage=6

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=52567&CurrentPage=7

Even a Windows specialist at the new NC data center:

http://jobs.apple.com/index.ajs?BID=1&method=mExternal.showJob&RID=63053&CurrentPage=11

 

[AidenShaw]

problem I see with that number at 20% is economy crashed happened.

Exactly - the 2010 and latest numbers are very different:

Worldwide, server shipments grew 27.1% year-on-year in the second quarter of 2010, while revenue was up 14.3%.

http://www.computerweekly.com/Articles/2010/08/27/242535/Server-sales-rebound-after-dismal-2009.htm​

Quoting 2009 numbers to claim that servers are dying is suspect, at least.

 

[alexkhan2000]

Exactly - the 2010 and latest numbers are very different:

Worldwide, server shipments grew 27.1% year-on-year in the second quarter of 2010, while revenue was up 14.3%.

http://www.computerweekly.com/Articles/2010/08/27/242535/Server-sales-rebound-after-dismal-2009.htm​

Quoting 2009 numbers to claim that servers are dying is suspect, at least.

2010 numbers are up compared to the very depressed 2009 numbers, but what are the margins? Oracle/Sun continues to decline while things have picked up for HP and Dell. Based on these IDC numbers for 2010 1Q numbers, Oracle/Sun is in a free fall.

http://www.idc.com/about/viewpressrelease.jsp?containerId=prUS22360110

My point is why does Apple need to be in this business at all? Apple couldn't be even 1% of the "Others" category and that section continued to decline as well although it picked up in 3Q. Dell's overall gross margins are in the 15% area with net margins in the 2% range. Why would Apple want to compete with that?

 

[Rodimus Prime]

2010 numbers are up compared to the very depressed 2009 numbers, but what are the margins? Oracle/Sun continues to decline while things have picked up for HP and Dell. Based on these IDC numbers for 2010 1Q numbers, Oracle/Sun is in a free fall.

http://www.idc.com/about/viewpressrelease.jsp?containerId=prUS22360110

My point is why does Apple need to be in this business at all? Apple couldn't be even 1% of the "Others" category and that section continued to decline as well although it picked up in 3Q. Dell's overall gross margins are in the 15% area with net margins in the 2% range. Why would Apple want to compete with that?

In terms of Apple it more than likely should be a loss leader. The servers generally was funded by large sales of the Mac and mac Pros. Killing Xserver just means that more of those rendering houses will say screw it and go windows and leave FCP. FCP has been lagging for a while and now will go even farther be hind. Just anohter nail in the coffen.

 

[MegaSignal]

Even if they do announce it at this point, they failed. You don't leave enterprise customers hanging and don't EOL products on a Friday without anouncing a migration path.

Seriously, they really burned some bridges on this one.

After reading ALL of the posts with regard to this subject, this particular post is The One with regard to the relevancy of the subject at hand: once you've burned your bridge, you're never coming back, at least with regard to anything more than an iGadget.

Surely I'm reading way too much into this, but I somehow feel that my Mac Pro just became more closely related to an iGadget than to a professional workstation...just sayin'...

I'll leave the remaining poetry to a one John Osbourne:

http://www.lyricsfreak.com/b/black+sabbath/looking+for+today_20019440.html

Cheers and all the Best!

 

[alexkhan2000]

In terms of Apple it more than likely should be a loss leader. The servers generally was funded by large sales of the Mac and mac Pros. Killing Xserver just means that more of those rendering houses will say screw it and go windows and leave FCP. FCP has been lagging for a while and now will go even farther be hind. Just anohter nail in the coffen.

I feel bad for those who have a vested interest in the Xserve. I can certainly understand the frustrations of the Xserve users and them feeling betrayed by this announcement. I just get the sense that Apple has no interest in this business at all, which is understandable as well from a cold-and-hard business perspective.

Apple has always thrived by differentiating their products from the competition and charging a premium for that differentiation. The only way large enterprise players like IBM, HP and Oracle can charge a premium is through software and services, not the hardware. There's no way Apple can compete in this field.

I'm sure that Apple came to the conclusion that the Xserve simply isn't worth the development cost and that they couldn't compete on either price or features and performance looking forward. Why do something just to be an industry laggard? Where does Apple fit in this sort of a market landscape?

http://www.zdnet.com/blog/microsoft/idc-windows-server-still-rules-the-server-roost/6424

Apple's Xserve and OS X Server would just be part of the "Other" category that makes up 0.3%. Although OS X has Unix underpinnings, it wouldn't be categorized as Unix that has the likes of IBM's AIX, HP's HP-UX and Sun's Solaris, which is also used by Fujitsu in a modified form. Just like the PC desktop market, Microsoft and Intel have commoditized the server market where they make the lion's share of the profits and the hardware vendors fight over the scraps.

 

[crackpip]

Now you've just shown how much you don't know about this topic.

I can run 20,000 clients off of about 15 VMs for Exchange. However, a render farm for 10 simultanous projects of video editing is probably going to require about 50 times the processing capacity.

Apple and oranges really. Final Cut Server doesn't just serve small text files from a database like Exchange does (yes, even with attachments, a 8 MB file served off of Exchange's EDB backend is nothing).

I show my ignorance by saying that there isn't a deployment for 20,000 clients using XServes for FCS? You know of a production deployment this big?

Or do I show my ignorance by saying that for large deployments services like exchange are better than Apple's solutions? Do you know of anyone seriously considering using OSXS in place of more established server offerings?

The only real example so far without a decent alternative appears to be FCS and this most likely for live action video. The digital animation houses don't seem to depend on Apple hardware for rendering. Certainly the data won't reside on the XServes, the data will be on 3rd-party RAID arrays. All you did was quote some general and 'probable' resource numbers. So what specifics do you know about Apple's server offerings? What do you know about the bugs and idiosyncrasies of their software? Just because there is a single application that isn't cross platform doesn't mean Apple's offerings are good. At least Rodimous mentioned that he knew of a place with about 100 client machines actually using FCS.

So Rodimous can you share any details about their setup? How do they handle 100 clients with each possibly using more than 100 mb/s of bandwidth? How many Xserves do they have? How many large RAID arrays? What's the work flow like?

Crackpip