For You Believers in DELL

[txlaw]

Dell mix of revenue - ~80% products and ~20% service.

 

Year over year, Dell's product revenue mix is showing a decline ( corporate, public government and retail customer ) and service revenue is growing at low single digit rate.

 

Dell is hamstrung by a change in business model.

 

For dell thesis to work:

  - the business shouldnt decline quickly (ala RIMM)

  - market should be willing to pay a higher price - ala Ericopoly's volatility trade

 

Ah, but what percentage of the product revenues are PC/mobility revenues?  And what percentage of income comes from PC/mobility revenues?

 

DELL is actually doing what needs to be done in light of changes in the supply chain.

 

This has nothing to do with a "volatility trade."

[thomcapital]

Ah, but what percentage of the product revenues are PC/mobility revenues?  And what percentage of income comes from PC/mobility revenues?

 

That is the million dollar question. I've seen estimates all over the place, with the most PC-weighted coming from Bernstein at ~72% of revenues and ~66% of EBIT. I'll attach the table from the note when I get home.

[Aberhound]

I bought today. Demand for desktops, laptops and servers will not end because of the IPhone and IPad. We decided to stay with a central server at our small office. Desktops will be replaced with thin clients as they get upgraded. We looked at Wyse which Dell recently bought last upgrade round 2 years ago but we felt it was not simple enough yet. Eventually I am sure we will all access your virtual computer in the cloud from any thin client but we will wait until it is simple. For now that choice seems insecure. If anything computers are so cheap and useful now we have ended up buying more computers as laptops have become so inexpensive.

 

The cash flow you get buying Dell stock for the price you pay looks compelling. Computers are so powerful now that buying Dell is easier and they do the hard work for us of negotiating cheaper licensing deals from the software providers then loading the programs on the computer. We hate Microsoft products but we are forced to buy them every round to allow us to share documents with other firms and customers. Dell indirectly enjoys some of Microsoft moat by making it cheaper and easier to get the licenses. For us the Dell moat is that purchase is fast, trusted and easy. It no longer pays to shop around.

[alwaysinvert]

Dell is way, way different from what I usually buy,  but I started nibbling at it yesterday too. Now I just need to raise some more cash from more illiquid investments.

 

Seems like they are in the perfect storm on the consumer side and if that weakness only is lessened in the midterm you can make out like a bandit. No credit is being given to their transformation and stickiness on the corporate and public business is probably greater than recognized. In the somewhat probable worst case scenario the consumer side keeps getting revenues slashed quickly for the forseeable future, but from what I can see that is already priced in like it's a fact. It won't be a homerun if that happens, but the fiasco risk seems very limited at these levels.

 

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Remains to be seen if I am dumb money competing in this space, but if that is the case, at least I'm in good company with Dell, Watsa, Simpson and some outstanding investors from this board. 

[claphands22]

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Dejavu... I read almost the same thing about RIMM not too long ago. That is, people saying the IT department would never switch their Blackberriers to iPhones.

 

*shrug* I have no opinion on Dell, so may the best analysis win.

[racemize]

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Dejavu... I read almost the same thing about RIMM not too long ago. That is, people saying the IT department would never switch their Blackberriers to iPhones.

 

*shrug* I have no opinion on Dell, so may the best analysis win.

 

Sure, but it seems like phones that people use for business and personal is quite different than the infrastructure used by the company.  E.g., IT has complete control over its own servers and presumably all desktops that remain on company premisis.  Laptops->tablets maybe a different story though.

[compoundinglife]

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Remains to be seen if I am dumb money competing in this space, but if that is the case, at least I'm in good company with Dell, Watsa, Simpson and some outstanding investors from this board.

 

I have worked in some form of IT/Tech for 17 years. Initially Fortune 500 large enterprise and the past 7 years in start ups. My experience is that large enterprises that are not software innovators (not amzn, not google, not facebook) generally buy Dell, HP, or IBM/Lenovo (tier 1) on the server side and on the client side. I don't see that changing anytime soon (5 years). However there are some potential game changers in the pipe that could change this longer term.

 

The first is private cloud, which is basically off the shelf cloud controller products that allow a large company to buy commodity (tier 2) servers such as ZT systems or supermicro and the cloud controllers take these cheap systems and make them into cloud storage and cloud compute, basically amazon web services behind your firewall. All the major IT players have a private cloud strategy at this point including Dell.

 

The companies that are making private cloud products if successful will likely by buy out targets of the big boys in 5 years. But there is some potential for these products to really commodify servers as the standard redundancy features such as data replication, failure tolerance can be handled in the cloud/software layer and not with machines that have redundant disks, power supplies etc... Redundancy is lots of cheap machines and you can get them for a lot less than the standard SKUs at Dell or HP. This also plays into the OpenCompute open source hardware that folks have mentioned on the board before.

 

The second is the trend in higher density, lower power platforms. Basically ARM based servers with tons of cores. You get more core density per rack at lower power consumption. The super efficient data center operators like amazon, google, facebook, msft are already working on this. This is evidenced by increased Linux support for ARM platforms and Windows 8 on ARM. If you are interested in where data centers are going I suggest you check out James Hamilton who works at AMZN formerly of MSFT. He is considered one of the large scale data center Guru's, his blog is: http://perspectives.mvdirona.com/

 

Both of these things are very interesting IMO and can potentially change server side computing and how it is done. I don't think these things change the Dell thesis dramatically in the next 5 years. I also think its likely that companies like Dell and HP will adapt to the new paradigm as it starts to unfold. If they don't then they are probably screwed. If they do it also gives them opportunities for great services business. They could start offer hardware that is priced near tier 2 levels but sell cloud setup/support services with it. Their customers would rather not do business with tier2 providers IMO.

 

Once large enterprises start dealing with a services company they are likely a customer for the long haul. They are an approved vendor, they have relationships based on trust, past experiences etc... At that point the customer is yours to loose, its all execution and providing the customer what they want (adapting).

 

So in conclusion I would agree that corporate IT is reluctant to switch vendors or to buy servers from the cheapest provider solely because of cost. However if the vendors they are currently working with fail to adapt to the new paradigm and provide the product/service mix enterprise customers will want 5 years from now they will suffer.

 

IMO Dell recognizes these things and is adapting accordingly.

[compoundinglife]

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Dejavu... I read almost the same thing about RIMM not too long ago. That is, people saying the IT department would never switch their Blackberriers to iPhones.

 

*shrug* I have no opinion on Dell, so may the best analysis win.

 

Not exactly the same IMO. The last BB only company I worked at refused to turn on Active Sync for Iphones on their exchange email server. Until the CEO bought an Iphone and he demanded it. Once that happened they made a special exemption for Execs, then they started to work out a plan on how to support the new platform and eventually everyone was allowed to hook their iphone up to corp. email server using their approved process.

 

The same thing could and does happen on the desktop/laptop side with Macs as has been talked about in this and other threads. But on the server side these changes are not going to be triggered by end user demand. As we have seen the future for these companies is not selling desktops and laptops if they think it is then they are already toast :)

[smathew]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

[DCG]

One thing I think (haven't really heard other people mentioning this anywhere) can be a big catalyst for computer (especilly laptop) sales is 4G-compatible laptops. As 4g/LTE ges rolled out on a wider scale, there is the potential for people to ditch their current ISP, modems and routers, and connec primarily through mobile broadband.

[txlaw]

Michael Dell interview at http://www.pcworld.com/article/2010574/michael-dell-on-surface-tablet-the-impact-will-be-limited.html

 

We’ve undergone a significant transformation, but we’re not as different as you might think. Dell has always been about creating customer value and solving customer problems. For a long time we did that by advancing personal productivity through devices and the adoption of the PC.

 

Today we’re still very focused on helping our customers get more value and better results from technology—but customer needs have changed, and we now offer a much broader set of solutions. It’s really an exciting time to be in IT. Innovations in areas like cloud, mobile, and big-data analytics are changing the way the world works, and we’re aligning our business with these new opportunities to better serve our customers. In five years, I expect we’ll be leading the way as an end-to-end IT solutions provider, but in some ways, we’ll also be the same—meaning very attuned to the needs of our customers.

 

-----

 

We’re focused on the entire IT ecosystem. The devices our customers use to generate and consume information are a critical starting point, and that remains very important to us. We have some of the best products in the marketplace. For example, Dell Precision workstations and our XPS ultrabooks. But we recognize that PCs are just part of the picture. We have leading capabilities to manage customer information seamlessly and securely in multiple-device and BYOD environments, including virtualized desktop that you can access from any device.

 

But beyond devices is an ecosystem of networking, storage, security, servers, virtualization, and cloud. This is where a lot of the opportunity in IT resides and where we believe we can lead. The total solution is world-class devices backed by world-class infrastructure and services that support, connect, manage, and secure customer information. If the only tool in your box is a hammer, then every problem looks like a nail. We’ve built a toolbox of customizable, scalable, flexible end-to-end solutions that put the customer first, not the technology or service we’re trying to sell.

 

-----

 

More than 80 percent of our business is what we call commercial—a combination of SMB, Enterprise, and Public Sector—so that’s a strength and a priority for us. We are particularly focused on the midmarket, which is underserved and also a segment where we are positioned to lead with our open, scalable solutions. However, with the ongoing consumerization of IT, we are also fully aware of the blurring of lines between what is consumer and what is commercial, and the products and services we’re delivering today and into the future are designed to bridge that gap.

 

[Guest rimm_never_sleeps]

not yet. maybe in a few years. and you may end up buying your LTE from the cable guy when you do. truly unlimited LTE will be expensive. You pay more to tether. Plus it just doesn't have the bandwidth to serve an entire household right now. Most people bundle their Internet service with TV. The cable companies are very clever about how they tie these services together so that it doesn't make much sense to drop internet. iow, the price of your other services will go up if you do.

 

However, this is why you are seeing comcast and others try to make deals with mobile operators. they see the writing on the wall. Dish sees mobile broadband as his way "in" to offer his customers internet access. But it's a long range plan.

[Guest valueInv]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

This I think is a big threat to companies like Dell. The big cloud companies get cheap, custom designed gear for themselves and don't buy from Dell. They have volumes high enough to do this and don't buy from the likes of Dell. Enterprises are shifting their data and computing to the cloud which means they don't buy as much gear as they used to. The end result - server and networking vendors get squeezed. Overall IT spending may increase, but the dollars may not go to current enterprise equipment vendors.

 

This could mean that Dell is skating to where the puck has been.

[compoundinglife]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

That article does not really dive into SDN (software define networking), but it mentions Nicira a SDN company that was just bought by VMWare for around 1billion. SDN is another thing driving the changes in switching and routing. You buy non-cisco, non-juniper switches that move bits at very high rates and then you virtualize the intelligence of the network using something like Nicira's solution. Again this is something that you are only seeing at companies that manage lots of data centers but something to keep a pulse on as it will change enterprise networking over the long haul.

[txlaw]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

This I think is a big threat to companies like Dell. The big cloud companies get cheap, custom designed gear for themselves and don't buy from Dell. They have volumes high enough to do this and don't buy from the likes of Dell. Enterprises are shifting their data and computing to the cloud which means they don't buy as much gear as they used to. The end result - server and networking vendors get squeezed. Overall IT spending may increase, but the dollars may not go to current enterprise equipment vendors.

 

This could mean that Dell is skating to where the puck has been.

 

Nope, they're well aware where the puck is going.

 

--------

 

Now, talking about the data center, the data center solutions I talked about are clearly quite different than the database solutions that you're bringing up, which is some very large service -- I'm just explaining your question -- some very large service providers are sophisticated enough that they want to bring the components down to the very base level buy the most inexpensive components, and they'll do all the integration, in essence, themselves. And there are some ODMs that are competing for various components in those, and quite successfully.

 

Dell also competes in that business. We have actually a very large business in providing a very integrated set of solutions. So, for the companies that don't want to invest in all their own integration and engineers to do that, we have the ability to integrate not only the very low cost server elements, but also the networking, power, cooling, storage, into, in essence, a very large infrastructure that we can deliver, in essence, as a data center in a carrier, and we can drop that on people's roofs, and have, or we can put it in a field. And in a very short period of time it, in essence, is a data center.

 

So, our offering is very different from what some of the ODMs are offering, which is very low cost piece parts, but you must integrate it and build the data center on your own to we have a very low cost and efficient integrated data center offering that we can drop and really shorten the time to deployment. Again, our strategy is all about ease of use and time to deployment. And so we are very consistent on those themes. So, our implementation dramatically cuts down on all of the effort that the customer needs to do in order to implement our solutions. So, we have a very different strategy there.

 

Now, I also would like to explain that many companies must decide which applications they want on and off premise for the reasons I articulated before, whether it's security, regulation, performance, whatever the reason is, many customers, if not most, are going to live in a hybrid world where they run some applications on premise, and some applications off premise in some form of a public cloud, let's say.

 

But once you've made that determination that you've got to have some applications in your environment. There are all sorts of advantages for a company like Dell that is a developer and distributor of those core components, because we'll be able to burst to a comparable environment. So, that will allow customers to operate their data center at average workloads versus peak workloads as long as they know the infrastructure on the public cloud side is comparable to the private cloud side. Then that provides all kinds of flexibility for them, which a traditional public cloud can never provide, because they can't mimic the exact environment they'll have on premise.

 

So, again, a different strategy, but one that clearly in that space also advantages our unique capabilities. So, whether or not it's in the purveyor of data centers that lends itself to our capability to fully integrate, or whether it's a public-private cloud hybrid, it plays to our ability to provide a comprehensive set of software, hardware and solutions that represent the next generation of converged infrastructure.

 

[compoundinglife]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

This I think is a big threat to companies like Dell. The big cloud companies get cheap, custom designed gear for themselves and don't buy from Dell. They have volumes high enough to do this and don't buy from the likes of Dell. Enterprises are shifting their data and computing to the cloud which means they don't buy as much gear as they used to. The end result - server and networking vendors get squeezed. Overall IT spending may increase, but the dollars may not go to current enterprise equipment vendors.

 

This could mean that Dell is skating to where the puck has been.

 

It is a threat if they don't adapt.  There is potentially a day when individual companies do not buy any server hardware and they only rent it from Amazon or Google or someone else. However these companies will still need services and software to build and support their enterprises and products.

 

That being said if it does happen, the shift will not happen overnight, enterprise IT moves slow which gives companies like Dell time to formulate a strategy and execute.  Also the huge interest in private cloud solutions indicates that enterprises are not ready to give up all their infrastructure to cloud providers but they do want the ability to have computing and storage as a utility. That is why companies like Dell and HP and currently out their offer solutions and service to help companies build their private clouds.

 

I think what will likely end up happening is that in the long term most enterprises will adopt a hybrid strategy where they have stuff in their own data centers and stuff in amazon or google's data centers, allowing them to shift load to where it makes the most sense whether it be because of cost or because of maintenance/outages.

 

[txlaw]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

That article does not really dive into SDN (software define networking), but it mentions Nicira a SDN company that was just bought by VMWare for around 1billion. SDN is another thing driving the changes in switching and routing. You buy non-cisco, non-juniper switches that move bits at very high rates and then you virtualize the intelligence of the network using something like Nicira's solution. Again this is something that you are only seeing at companies that manage lots of data centers but something to keep a pulse on as it will change enterprise networking over the long haul.

 

QUESTION: Thanks, Chris. Vijay Bhagwat, Deutsche Bank. My question for you is, we all read the same news. You saw the VMware acquisition from Nicira, and at VMWorld recently, you know, kind of had a sense of their strategy in terms of software-defined networking, software-defined storage, cloud orchestration. I would like to get kind of a Dell point of view on all of this, which is network virtualization, SDN, virtualized storage, what's Dell's point of view?

 

DARREN THOMAS: Yes, that's actually a great question. There are so many moving parts going on here, I'm sure that the storage industry must look like a hornet's nest to you guys trying to track a single bee. So, in that particular issue, here's what's going on. You have companies, to some great degree application companies, saying, hey I can just put storage directly on the server, direct attach. That's pretty fast. Direct attach storage is pretty fast. And I'll manage all the reliability and stuff like that up above it.

 

And now with virtualization you say, okay, so now I want to do all that virtually. So, I just want to snap my fingers, install a server, all that storage is under there, and I want to have it virtually managed with a software tool. So, this concept of software, just software only, so it's like storage is out of thin air. It's just whatever is on the disk.

 

First of all, all our storage, all almost everybody's storage is about 98 percent software, only about 2 -- I have 1,300 engineers, I think like five of them are hardware engineers. And they help me build my hardware in some offshore design team, to be honest. All my IP is in software.

 

Also, if you notice, Compellent runs on a Dell Server. Let's back what that means. Compellent runs on an Intel standard architecture. Compellent will run in a VM. EqualLogic, while it runs on a different architecture, will run in a VM. We can run it in a VM. And that's exactly that software.

 

So, what you have is all these companies saying, hey, I can make a piece of storage out of thin air, and I'll do it like either at the Microsoft level or I'll do it at the VMware level. I'll do in one of those. But what you get is, you basically get reads and writes, and if it fails we'll copy it somewhere else and do it. Well, what about snapshots, replication, what about all of this stuff you got used to using? What about that stuff that protects you? What about thin provisioning? What about all that? Well, we haven't gotten there yet. We haven't even talked about it yet.

 

Well, our virtual solutions, the ones I'm talking about running on VMs, come with snapshots, replication, thin provisioning, all that feature set. So, right now I'm saying bring it on, I would love to see that world where I can eliminate the hardware. I would like to see, if I do that, will you guys represent me as being a software-only company?

 

Yes, that would be pretty cool, because at that point in time I've extrapolated myself from the hardware. I have no hardware at all. My value is pure software, and we all know the software companies have great multiples. So, I'm looking forward to that.

 

But, to be quite honest, we can run our software on a server with no assist from any kind of storage hardware. And that's coming, that's actually coming. What I don't think is every customer out there is going to go to that, because there's a few little minor flaws with that. Number one is direct attach, if the server goes down, all those disks go down. Yes, you can copy that data somewhere else, but now you've got to pay twice the cost if you're copying the entire content to another location. So, RAID 5,RAID 6, that's out the window. All the cost savings of using six plus one or seven plus one are gone. Now every piece of data is mirrored to another piece of data.

 

Those things have their value when you want to do it quick, set it up, run it in a virtual world, and then shut it back down again. But if you're going to run it and set it for a long period of time, you're going to pay about twice as much for the disk, roughly. So, they had their limitations.

 

So, I think we're headed to a world where that just is one more bee in the beehive, to be quite honest. Does that make sense?

[Guest valueInv]

So, I think we're headed to a world where that just is one more bee in the beehive, to be quite honest. Does that make sense?

 

Actually it makes no sense. Compounding life posted about networkings (SDN), you responded by cutting and posting about storage. Two different things.

[Guest valueInv]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

This I think is a big threat to companies like Dell. The big cloud companies get cheap, custom designed gear for themselves and don't buy from Dell. They have volumes high enough to do this and don't buy from the likes of Dell. Enterprises are shifting their data and computing to the cloud which means they don't buy as much gear as they used to. The end result - server and networking vendors get squeezed. Overall IT spending may increase, but the dollars may not go to current enterprise equipment vendors.

 

This could mean that Dell is skating to where the puck has been.

 

It is a threat if they don't adapt.  There is potentially a day when individual companies do not buy any server hardware and they only rent it from Amazon or Google or someone else. However these companies will still need services and software to build and support their enterprises and products.

 

That being said if it does happen, the shift will not happen overnight, enterprise IT moves slow which gives companies like Dell time to formulate a strategy and execute.  Also the huge interest in private cloud solutions indicates that enterprises are not ready to give up all their infrastructure to cloud providers but they do want the ability to have computing and storage as a utility. That is why companies like Dell and HP and currently out their offer solutions and service to help companies build their private clouds.

 

I think what will likely end up happening is that in the long term most enterprises will adopt a hybrid strategy where they have stuff in their own data centers and stuff in amazon or google's data centers, allowing them to shift load to where it makes the most sense whether it be because of cost or because of maintenance/outages.

The threat is commoditization and buying power. With SDN, open compute, etc, they're moving to stringing cheaper gear together with high redundancy. This means price wars and lower margins.

 

The move may be slow but the effects could be felt quickly. If enterprises decide to implement new projects in the cloud, they may not need to buy new gear as much as they did last year, they'll make do with existing gear. This could lead to very lesser new demand created, maybe even overcapacity, price wars, etc.

 

Like I said, these are plausible threats to keep to watch, not predictions.

[Guest valueInv]

My dad used to work with IT purchases in the public sector and he says their technicians refuse to work with different brands of computers to avoid hassle. They buy only Dell. Anecdotal as hell, but I think it speaks to a stickiness which maybe is not clearly understood. If they want to change they are going to redirect ALL future purchases and it will raise problems for the organization. Not worth doing if the savings aren't substantial.

 

Remains to be seen if I am dumb money competing in this space, but if that is the case, at least I'm in good company with Dell, Watsa, Simpson and some outstanding investors from this board.

 

I have worked in some form of IT/Tech for 17 years. Initially Fortune 500 large enterprise and the past 7 years in start ups. My experience is that large enterprises that are not software innovators (not amzn, not google, not facebook) generally buy Dell, HP, or IBM/Lenovo (tier 1) on the server side and on the client side. I don't see that changing anytime soon (5 years). However there are some potential game changers in the pipe that could change this longer term.

 

The first is private cloud, which is basically off the shelf cloud controller products that allow a large company to buy commodity (tier 2) servers such as ZT systems or supermicro and the cloud controllers take these cheap systems and make them into cloud storage and cloud compute, basically amazon web services behind your firewall. All the major IT players have a private cloud strategy at this point including Dell.

 

The companies that are making private cloud products if successful will likely by buy out targets of the big boys in 5 years. But there is some potential for these products to really commodify servers as the standard redundancy features such as data replication, failure tolerance can be handled in the cloud/software layer and not with machines that have redundant disks, power supplies etc... Redundancy is lots of cheap machines and you can get them for a lot less than the standard SKUs at Dell or HP. This also plays into the OpenCompute open source hardware that folks have mentioned on the board before.

 

The second is the trend in higher density, lower power platforms. Basically ARM based servers with tons of cores. You get more core density per rack at lower power consumption. The super efficient data center operators like amazon, google, facebook, msft are already working on this. This is evidenced by increased Linux support for ARM platforms and Windows 8 on ARM. If you are interested in where data centers are going I suggest you check out James Hamilton who works at AMZN formerly of MSFT. He is considered one of the large scale data center Guru's, his blog is: http://perspectives.mvdirona.com/

 

Both of these things are very interesting IMO and can potentially change server side computing and how it is done. I don't think these things change the Dell thesis dramatically in the next 5 years. I also think its likely that companies like Dell and HP will adapt to the new paradigm as it starts to unfold. If they don't then they are probably screwed. If they do it also gives them opportunities for great services business. They could start offer hardware that is priced near tier 2 levels but sell cloud setup/support services with it. Their customers would rather not do business with tier2 providers IMO.

 

Once large enterprises start dealing with a services company they are likely a customer for the long haul. They are an approved vendor, they have relationships based on trust, past experiences etc... At that point the customer is yours to loose, its all execution and providing the customer what they want (adapting).

 

So in conclusion I would agree that corporate IT is reluctant to switch vendors or to buy servers from the cheapest provider solely because of cost. However if the vendors they are currently working with fail to adapt to the new paradigm and provide the product/service mix enterprise customers will want 5 years from now they will suffer.

 

IMO Dell recognizes these things and is adapting accordingly.

 

Great post!

 

I saw the trend on ARM-based servers. Seamicro was doing very well, but got bought by AMD for really cheap. So makes we wonder about the traction. AMD turned around and converted Seamicro's architecture back to x86. There were some other startups doing similar things with low power processors but haven't heard much about them recently. Do you know anything about their traction? What do the data center people think about it? Are they excited about this kind of architectures?

 

Another architecture that is interesting is the Cisco UCS approach (server + networking + other components in a box). I think IBM also released something similar. I don't know how successful this has been. I remember Cisco touting some big numbers but don't know how true they were. Any idea how that is doing?

[txlaw]

So, I think we're headed to a world where that just is one more bee in the beehive, to be quite honest. Does that make sense?

 

Actually it makes no sense. Compounding life posted about networkings (SDN), you responded by cutting and posting about storage. Two different things.

 

I thought that Compounding Life's post about software defined networking (SDN) was very useful, and I wasn't trying to refute anything he said.  Instead, I wanted to point out that DELL is thinking about this very interesting shift, which is occurring not just in networking but also in storage, which is a big part of DELL's IP portfolio.

 

Clearly, you didn't read the excerpt that closely.  The DELL guy's answer to the question he got -- which was about software-defined networking, software-defined storage, "cloud orchestration", etc. --  was interesting to me because it indicates that DELL has been thinking about the importance of software for provisioning cloud infrastructure.  True, the guy focused on storage because DELL-owned IP is very storage heavy.  But I would not be surprised if DELL had actually been looking at Nicira when VMware snapped it up. 

 

Just another sign that DELL sees where the puck is going.

[txlaw]

See this interesting article about what google/amazon and facebook are doing in their

data center

 

http://www.wired.com/wiredenterprise/2012/03/google-microsoft-network-gear/all/1

 

This I think is a big threat to companies like Dell. The big cloud companies get cheap, custom designed gear for themselves and don't buy from Dell. They have volumes high enough to do this and don't buy from the likes of Dell. Enterprises are shifting their data and computing to the cloud which means they don't buy as much gear as they used to. The end result - server and networking vendors get squeezed. Overall IT spending may increase, but the dollars may not go to current enterprise equipment vendors.

 

This could mean that Dell is skating to where the puck has been.

 

It is a threat if they don't adapt.  There is potentially a day when individual companies do not buy any server hardware and they only rent it from Amazon or Google or someone else. However these companies will still need services and software to build and support their enterprises and products.

 

That being said if it does happen, the shift will not happen overnight, enterprise IT moves slow which gives companies like Dell time to formulate a strategy and execute.  Also the huge interest in private cloud solutions indicates that enterprises are not ready to give up all their infrastructure to cloud providers but they do want the ability to have computing and storage as a utility. That is why companies like Dell and HP and currently out their offer solutions and service to help companies build their private clouds.

 

I think what will likely end up happening is that in the long term most enterprises will adopt a hybrid strategy where they have stuff in their own data centers and stuff in amazon or google's data centers, allowing them to shift load to where it makes the most sense whether it be because of cost or because of maintenance/outages.

The threat is commoditization and buying power. With SDN, open compute, etc, they're moving to stringing cheaper gear together with high redundancy. This means price wars and lower margins.

 

The move may be slow but the effects could be felt quickly. If enterprises decide to implement new projects in the cloud, they may not need to buy new gear as much as they did last year, they'll make do with existing gear. This could lead to very lesser new demand created, maybe even overcapacity, price wars, etc.

 

Like I said, these are plausible threats to keep to watch, not predictions.

 

No, Compounding Life is exactly right.  The threat that DELL really must watch out for is failing to adapt to hardware commoditization.

 

You're basically falling prey to the common misconception that DELL is simply a hardware seller.  They're not.  They are a productivity solutions provider that offers both software and services.

 

Moreover, as Compounding Life points out, there are both public clouds, private clouds, and hybrid clouds.  DELL will focus on private clouds, hybrid clouds, and turnkey solutions that include more than just the hardware.

 

Your point about watching out for these plausible threats is a good one, though.  Now if only you applied that rationality to AAPL and the potential threats they face to their margins.

[smathew]

Another related article

 

http://www.wired.com/wiredenterprise/2012/05/quanta-facebook-us/

 

I think it will take some time to see the impact this shift in networking switches .  But in server market it is already happening . 

 

 

[compoundinglife]

Great post!

 

I saw the trend on ARM-based servers. Seamicro was doing very well, but got bought by AMD for really cheap. So makes we wonder about the traction. AMD turned around and converted Seamicro's architecture back to x86. There were some other startups doing similar things with low power processors but haven't heard much about them recently. Do you know anything about their traction? What do the data center people think about it? Are they excited about this kind of architectures?

 

Folks I know in the large scale data center area are excited about ARM. I think its a long way out before it will be replacing all their PC based systems. James Hamilton talks about it a little bit on his blog: http://perspectives.mvdirona.com/2012/01/02/ARMV8Architecture.aspx But in summary he sees advances in the server designs coming from the mobile space where power is already a huge constraint.

 

I think right now there is not a huge space for companies making off the shelf ARM servers, its all custom stuff right now IMO. Might be why Seamicro went back to X86. Considering that the early adopters like Google, Facebook and Amazon will likely be getting their ARM stuff from ODMs I don't see anyone making a ton of money right now selling ARM servers.

 

Another architecture that is interesting is the Cisco UCS approach (server + networking + other components in a box). I think IBM also released something similar. I don't know how successful this has been. I remember Cisco touting some big numbers but don't know how true they were. Any idea how that is doing?

 

To be honest I am not deeply familiar with the UCS offering but my impression is that its just a package of existing components (network, server etc..) with some central management software in condensed form factor w/backplane. It also AFAIK is not any cheaper than most other hardware it selling point is that it's supposed to lessen operational burden.

[Guest valueInv]

So, I think we're headed to a world where that just is one more bee in the beehive, to be quite honest. Does that make sense?

 

Actually it makes no sense. Compounding life posted about networkings (SDN), you responded by cutting and posting about storage. Two different things.

 

I thought that Compounding Life's post about software defined networking (SDN) was very useful, and I wasn't trying to refute anything he said.  Instead, I wanted to point out that DELL is thinking about this very interesting shift, which is occurring not just in networking but also in storage, which is a big part of DELL's IP portfolio.

 

Clearly, you didn't read the excerpt that closely.  The DELL guy's answer to the question he got -- which was about software-defined networking, software-defined storage, "cloud orchestration", etc. --  was interesting to me because it indicates that DELL has been thinking about the importance of software for provisioning cloud infrastructure.  True, the guy focused on storage because DELL-owned IP is very storage heavy.  But I would not be surprised if DELL had actually been looking at Nicira when VMware snapped it up. 

 

Just another sign that DELL sees where the puck is going.

 

Explain to me what happens to their Force10 acquisition in the era of SDN and what happens to their server business in the era of open compute. Why did they buy Force10? What market segment (not vertical) does Force10 target?