Saturday, August 21, 2010

Dell and virtualization

I recently posted three long pieces on computing and virtualization.  There were two trends that mattered.

(a).  Enterprise computing was moving to the cloud - sometimes an external cloud - but more often-than-not an enterprise cloud.  Computers would henceforth be virtual machines sitting on some mondo-powerful server and thin clients.  This setup was superior for a range of reasons that I went into.  But more than superior - it was cheaper to set up and MUCH cheaper to run.

(b).  Personal computing was going towards ever-more-powerful handheld devices on which I thought PC as-an-app would be an end game.  These devices will plug into screens and keyboards at home if you wanted - and to a large extent they would replace or displace the personal computer.  

Microsoft would be a winner if it remained the cloud desktop (which it has at every firm I know who uses an enterprise cloud).  It would be a big winner if it also captured a worthwhile place in the handheld device market.  I thought that unlikely - but the reviews of new Windows 7 phones are convincing me that a Windows 10 phone could be a big winner.  

Although I did not discuss this, one implication was a bleak future for "beige box makers".  Computers disappeared into a cloud or into a shiny and extremely well made personal computing device.  A logistics firm which assembled beige boxes (as Dell was in its glory days) had no future.  You either needed to be in the enterprise server (especially blade server*) market or a consumer good maker.  Dell I posited was an extremely poor consumer good maker - as my old XPS1330 laptop would attest.  [That laptop had the notorious Nvidia graphics chip problem...  Hewlett Packard also dealt with it badly.]  

Now I know of no evidence that Dell is getting its device business together (I will never trust Dell with another consumer device).  However some think the new Dell phone (The Thunder) is cool.  

But Dell is approaching the enterprise cloud product in a purposeful and seemingly effective way.  Sure mostly they are doing it via acquisitions - see the latest purchase of 3Par.  3Par make integrated hardware-software solutions for storage in enterprise clouds - in other words sophisticated jbods**.  And this is after a few other acquisitions in that area.  

And surprisingly it is working.  Dell server growth was rapid even as their beige-box business continued to falter.  As they they stated "storage and server revenue increased 35 percent with rapid growth in blades".  They talk about integrated solutions to enterprise cloud - presumably so relatively unsophisticated IT guys can install it.  

Hewlett Packard is also getting its growth in blades - but - frankly - not at anything like this rate.  Partly they are incumbent - but mostly I think they are just ceding share to a company - that at least on that area has its eyes firmly focussed on the prize.  Here - and I quote verbatim - is the quarterly highlights.

Dell’s commercial business continues to benefit from improved demand across all products and services, and in all geographies as Dell expands its enterprise solutions portfolio. Recently, Dell acquired Scalent, developer of virtual infrastructure management technology, and Ocarina Networks, a leading developer of storage optimization technology. The company also announced an agreement to acquire 3PAR, the leading global provider of utility storage for cloud computing. These moves illustrate Dell’s commitment to build its capabilities for open and affordable enterprise solutions.

This has me more-than-passingly surprised.  In a world where the beige-box disappears Dell, once the biggest maker of beige-boxes, is finding itself a seat at the table.   It may not be as good a seat as they held in 1998 - but it is a seat nonetheless.  And they are doing it by going where the puck is going to be.   And that has me surprised because - personally - I hate Dell like I hate few other companies.  They sold a defective computer, replaced the defective part with another defective part and kept me waiting in endless phone-center loops.  

But their agility here had me covering my short last night. 

I am - I confess - truly stunned.  If you ask me whether Michael Dell should keep his job.  On the evidence of this I have no doubt...

 

 

John

*blades - for the non-geek - are the most effective way of getting really dense and cheap mondo-computing capacity.  Blades sit in boxes which have integrated power supply and cooling and are effectively the ready-built motherboard of a server ready to plug in.  You can get an enormous amount of computing capacity into a small room filled with blade servers.  [The room will also need serious cooling.]

**jbod = just a bunch of discs.

7 comments:

Steve said...

Microsoft stands a very good chance at being able to dominate the cloud computing world. They have the enterprise relationships and the depth of experience in running large scale services that is required. From my vantage point at a large technology company I don't think that companies that are running just a few data centers will be able to be cost competitive with companies like Microsoft and Google in terms of $/unit of work. When a company has a large data center footprint they can simply invest more in data center technologies. Power efficiency has become much more important than it was 10 years ago.

Because of this you can bet that ARM and Intel will be winners as the suppliers of chips to the next generation of cloud infrastructure. Look for Dell and HP to increasingly offer rack level solutions or container level solutions instead of individual servers.

nick gogerty said...

net connectivity increases mean that most devices will be interfaces, API's and drivers. This will continue. Centralized services allow for economies of maintenance, innovation and security. Bandwidth kills heavy processing at the edges. the interface is everywhere and the brains are nowhere. Nick Gogerty Former Chief Analyst Starlab (europe's answer to the MIT medialab.)

strowger said...

Thank you for a fascinating series of articles, John. As an IT guy, it's really reassuring to read your stuff on IT - as I can't fault what you've said, it gives great confidence that your finance writing is of equal quality.

On virtualisation - centralisation is an idea that pops up in the IT industry every few years and then fades away. In the 1990s it was Windows Terminal Server, Citrix Winframe and Sun Javastation. Corporate IT staff like centralisation - it protects their jobs, makes them seem and feel important, and gives them power over users. Never underestimate the average corporate IT guy's dislike of his customers.

Virtualisation is clearly of enormous use in the world of servers. I'm pretty sceptical, however, about it replacing desktop/laptop computers for large numbers of corporate users. As IT departments gain tighter control over users' desktops, they become less useful to the users and this saps productivity. This (IMV) is why PCs defeated mainframes 25 years ago.

Anonymous said...

HP bidding for 3Par this morning - $24 per share. Seems like an outrageous valuation to me - 3Par had good technology but not a very good business.

Mike said...

3par: Decent technology, however a lot of what was considered low end hardware previously can now perform equivalently. The main reason for purchasing 3par in the past was the massive iops you could achieve. However, given the cost of maintenance we're migrating to more special purpose hardware (ie, dell 510's for database, standalone sata jbods for mass storage). Performance on this lower end hardware is comparable if not better then we saw on 3par.

cloud: Good for processing on demand, but for daily operations it's not a viable replacement if you look at costs over 3 years. The storage services, however are very reasonable especially for data distribution.

I'll have to agree with strowger on virtualization and just add that while we haven't seen it as a viable replacement for a desktop, it has enabled our developers to operate much more efficiently with less hardware on their desktops.

Anonymous said...

You do realize that today you can buy a ConnectLand mobile SATA bay for US $16 on MonoPrice. The bay slides comfortably underneath your DVD 5.25" bay in any mid or tower computer. You can build a Hackintosh load on a hard drive, slide it into the bay underneath your desktop mid-tower's DVD reader, change one setting in the BIOS and re-boot into SnowLeopard. Or, you could slide your Win 7 drive into that bay and re-boot, or Linux, etc., etc. You can build your own 'virtual' hardware box that can run all three operating systems because they all run on Intel/AMD and by and large have support for the same types of external devices and protocols. Check out this mobile SATA bay. It is really slick.

server virtualization said...

Being a part of Dell, I found your blog on virtualization very informative. To add more to it, I think virtualisation is a really beneficial technology in increasing the data security and workforce flexibility.

General disclaimer

The content contained in this blog represents the opinions of Mr. Hempton. Mr. Hempton may hold either long or short positions in securities of various companies discussed in the blog based upon Mr. Hempton's recommendations. The commentary in this blog in no way constitutes a solicitation of business or investment advice. In fact, it should not be relied upon in making investment decisions, ever. It is intended solely for the entertainment of the reader, and the author.  In particular this blog is not directed for investment purposes at US Persons.