The real question for Dell is: Quo Vadis?

For me, the acquisition of Perot Systems provides no answer real to that question.

Dell has gone from an irresistible force to illogical object in record time. The IT market can be so cruel.

The Dramatic Rise: Dell created a manufacturing culture that almost rivaled Toyota. It squeezed cost out of the manufacturing process in a dramatic way and pushed down the price of PCs and servers accordingly. It was perhaps the first company to use the web as its primary sales channel and in doing so it captured a very large slice of the corporate PC and server market – and also won a healthy slice of  the consumer PC market. For a while it controlled the price of PCs and low-end servers and was a truly tough competitor.

The Hiatus: Odd though it may seem, Dell saturated its market. Just as only a certain percentage of the market for books migrated to the web, only so much of the market for computers migrated there. Indeed from about 2002 onwards Apple demonstrated (despite the scoffing of skeptics) that the consumer market for PCs is best addressed by having a retail chain. In recent times we have begun to observe another phenomenon. The commodity Intel server market is itself under pressure. Here’s why:

• As regards servers, the SMB market is now drifting towards the cloud.
• It is beginning to become apparent that larger servers deliver bettwer value. So there is now a downward pressure on the commodity server market.

If you’re wondering why Dell’s stock price has headed so far south, that’s why. HP’s stock price looks a bit sick for the same reason, but note that the purchase of EDS (a much larger operation than Perot Systems) has done much to bolster HP’s health.

Ultimately Dell had to decide whether it was a consumer company or a corporate provider and it has come down on the side of the corporate market. That’s the only thing I can deduce from its surprise purchase of Perot.

Why the server market will only get harder.

I read the tea leaves for the server market in Say Goodbye To The Server Market but as I glance through it I realise that I left one company out of the picture; Cisco. The server market has collapsed by 30 percent and, imho, it’s gone forever. That 30 percent isn’t coming back when the economy revives. It’s all about virtualization and in that game I see IBM as a strong player and I see Cisco (with its impressive Nexus virtualize-your-network-in-one-fell-swoop switch) as a possible contender (starting from nowhere). I see HP as reasonably strong except in the Unix market (given recent performance). I see Dell, Sun and Fujitsu as armies in retreat.

There are too many server companies. The industry could lose 2 and it still might have too many. It’s rough out there – high water everywhere.

In simple terms, the low end of the market will gradually evaporate into the cloud and the upper end of the market will not be a place where the impressive manufacturing skill of Dell is going to be important.

The Services Market

The acquisition of Perot Systems is the biggest acquisition Dell has done. It is paying a premium at 1.6 x revenue (HP only paid .6 x revenue for the moribund EDS) and it has no real services operation (whereas HP had actually been growing its services arm quite respectably for several years before it swallowed EDS).

None of that bodes well for Dell, but that’s no reason to dismiss it. It might pull a rabbit out of the hat. However I think Dell is now drinking in “The Last Chance Saloon”, if this doesn’t work, my guess is that it will follow Wang, Digital and Compaq into the setting Sun.

The war is over. It is time to go round and bayonet the wounded.

What am I talking about?

I’m talking about the legitimacy of Open Source. There was a war in the IT industry between the Open Source fan boys and those who believed that Open Source was a plague on the legitimate business model of selling licensed software. And it is – if by “plague” you mean “challenge.” The war is over. The Open Source guys won.

The Politics of Open Source

The problem with politics is that it gets in the way of rational discussion and rational conclusions. Politicians cheat. I’m not just talking about the politicians you happen to dislike, I’m talking about them all. There may be a few honest politicians, but the political honesty gene is rare. Most politicians are cheer-leaders for a specific political posture, idea or movement and they’ll happily tie the facts in knots  and invent bizarre slogans in order to support it.

Some of the supporters of Open Source are just raucous politicians with a very definite agenda. Open Source is good because it’s right. Power to the GPL. Software wants to be free, like music and movies. Infinite goods. Blah, blah, blah.

Back To First Principles: The Problems of Closed Source

There is a problem with closed source and it’s simply this. You have no idea what it’s doing or how it is doing it. That’s fine if you trust it, accept its limitations and don’t want to interface with it, but otherwise it’s not so fine.

Let’s take this piece by piece:

1. Trust. With Open Source nothing is hiding. If the code was stolen then the rightful owner can identify the fact. If the code is closed there is a risk. So there’s a trust issue. You might trust the vendor of closed code, but you don’t have to trust the provider of Open Source as much. Oh and you can also see what it’s doing so you know can find out whether its messing anything up.
2. Limitations. Some software does exactly what you want and you don’t care to change it. Some does not and you’d like to change it, even if only in a small way. I use a fair amount of Wordpress-based code, including plug-ins and themes and I make changes. I would never use the software if I couldn’t do that.
3. Interfaces. Finally we need to consider interfacing. In truth, this is the whole point of this posting.  In the end it’s all about the ecosystem that can be generated around a given Open Source component or set of components. If you make the interface to the software available it also makes sense for the source code to be available.

Open Source resolves all these issues and as long as it fulfills your application need, it is much preferable to closed source. This is not about cost. Open Source products are not all free and we shouldn’t expect them to be. Software doesn’t write itself. However having the source open is an important feature in many instances and sometimes it’s a must-have feature.

For example, I wouldn’t accept any Wordpress plugin if I didn’t get the source code. It starts to become really important when you encounter shoddy programming. If you have the source code you can support the software yourself – rewrite it even and you may well prefer to. And, of course, you can pick up support if the original developer is hit by the proverbial bus.

Open Source Today

Having given the rational argument, lets now simply examine the evidence that Open Source has acquired true respectability.

There are many successful Open Source products; too many to list; so many that just about every software application you can think of has an Open Source alternative. There are over 300,000 Open Source projects listed on SourceForge, some of which are defunct and some of which are trivial, but many are active and support well established and widely used products.

There are many important companies that devote man years of time to writing Open Source code. There is, of course, Mozilla (with FireFox) and Red Hat both of whom base their whole business on Open Source. But there are many others. Sun Microsystems (soon to be Oracle) has given massive support to Open Source. It owns the most popular database in the world; the open source MySQL. It has Open Sourced the whole of Solaris and much of Java and it also is the custodian of Open Office the popular Open Source office application. Add it all up and it amounts to tens of millions of lines of Open Source.

IBM’s contribution to Open Source is almost as great with 12.5 million lines of code contributed to Eclipse and a great deal of code contributed to Linux, Apache, Geronimo and other important Open Source efforts. Intel has written its own version of Linux called Moblin, which it recently donated to the Linux Foundation. And then there’s Google with Android (its own version of Linux), the Chrome browser and other incidental Open Source products. Google even provides a Open Source project hosting capability that rivals Sourceforge in size.

Even Apple, which majors in proprietary software, has learned to coexist with Open Source. Beneath Apple’s excellent OS X sits Darwin, the core operating system which is fully Open Sourced. Darwin integrates the Mach 3.0 kernel based on version 4.4 of BSD (Berkeley Software Distribution) Unix.

Nowadays the only real bitter-end opponent of Open Source is Microsoft. Pretty much everyone else has found a way to accommodate it within its business model. Microsoft occasionally makes gestures, such as its recent plan to fund The CodePlex Foundation – a not-for-profit extension of its CodePlex web site. But it has too much to lose to really bite the bullet.

But so what? The war is over. Open Source won. The whole IT industry now depends on a vast amount of Open Source code and many key Open Source projects.

Vista was a sad and costly failure from day one. It was a mistake and Microsoft is wiser now. Vista might have fared a little better had Microsoft not had to negotiate its way through a recession, but that’s incidental. It wasn’t the cause of Vista’s failure.

The main reason Vista came crashing to the ground was that Microsoft believed it could just continue to repeat its modus operandum, which ran something like this:

1. The chip market delivers faster chips.
2. Manufacturers produce bigger and faster PCs.
3. Microsoft produces a new OS to chew up as much of the newly bestowed resources as possible.
4. Older PCs now seem obsolescent.
5. Users get excited about the new PC, which is just like the old PC only the look and feel has moved one inch south and half an inch to the left.
6. A flash flood of new revenue pours into Redmond and the sound of champagne corks popping is heard all over Seattle.

But a funny thing happened on the way to the Vista: this business model collapsed. Several disruptive factors, some that were completely out of Microsoft’s control, blew it apart. Here’s what happened:

1. The Longhorn/Vista project grew late, then very late, then exceedingly late then abysmally late.
2. Microsoft acquired a very powerful competitor in Apple. Apple took the intelligent step of moving to Intel chips – a move that was not made in order to have a more powerful processor (in reality, the processor was less powerful), but in order to make the Mac directly comparable and therefore competitive with the PC.
3. With OS X, the Apple OS was not only better than Windows, but it had three new releases in the time the world was waiting for Vista. The OS X development team must have worn a permanent smile for years. The killer capabilities of OS X (from the user’s perspective): no blue screens, no viruses, a GUI that worked really well.
4. Economy of resource utilization became an issue. People had begun to get tired of long boot up times and then suddenly the small footprint laptop – or Netbook as it is called – appeared. Vista could not even run on a Netbook.
5. The Netbook gave Linux a toehold in the PC market, which Microsoft did its best to stamp on. But it was too late – another if less powerful competitor was now gathering at the gates of Redmond.
6. Corporate PC buyers began to get sick of the cost of PCs when they no longer deliver significant new benefits to the business. Suddenly business benefit in the corporate world was all about mobile devices – a space in which Microsoft did not play well. they didn’t want or care about Vista.

Vista flopped because it was bloated and compared poorly – not just to OS X and Ubuntu – but also to Windows XP. Many companies stayed with XP and cut their PC costs accordingly.

The Prospects for Windows 7

The prospects for Windows 7 are very good. I doubt if the level of uptake is going to be dramatic, but it will be dramatically better than the uptake of Vista. Microsoft will be able to declare a success of a kind. Windows 7 is the first Windows that is not deliberate bloatware. It may not have the economy of OS X (especially the Snow Leopard) or the extreme economy of Linux, and it is still trailing in usability. But it is streets better than Vista and many businesses will make the upgrade as a means of skipping past the Vista wreckage.

Nevertheless, Microsoft faces a huge challenge with Windows 7 and the outcome will not be clear for quite a while. Microsoft once had a monopoly and that’s gone. As time passes there is going to be competition from Google’s Android as well as OS X and desktop Linux. For Microsoft winning will not mean gaining market share, it will mean not losing so much market share that it appears mortally wounded.

The enemy is at the gates and Microsoft has everything to lose.

Yes, pretty much everything.

If Windows falls, a great deal more comes tumbling down with it.

OK I exaggerate, but say goodbye to the server market as we know it. It’s over. And hey, don’t be sad. This is the IT industry after all, nothing lasts very long. Do you remember when Sun was a computer company?

Examining The Entrails

IDC estimated the server market as worth $55.1 billion in 2007, but a mere $53.3 billion in 2008 – a decline of 3.5%. That didn’t look too bad, after all a recession had begun. However, the last quarter of 2008 was abysmal. Revenue fell by 14% over the previous year. Things had never been that bad since the dot com bust, but they were about to get worse.

Q1 2009 saw revenues collapse by 24.5 percent to $9.9bn and server unit shipments were even lower (by 26.5 percent). Then the second quarter made the first seem quite respectable. Revenues were down 30 percent—to $9.8 billion, with unit shipments down by 30.4 percent. The quarterly revenue was the lowest since IDC began server counting in 1996 – yes, that’s not a typo, it’s 1996, 13 years ago!

That marked the fourth consecutive quarter of revenue decline. Since it was little different in revenue terms to the previous quarter, we can now think of the server market as being worth about $40 billion.

So in the space of a year it lost the equivalent of a whole server company about the size of, well, Sun Microsystems.

But Hey, Don’t You Know There’s A Recession In Progress

And that’s the whole point of this posting. Yes I do. So let’s look at the PC market as a counterpoint.

In the wake of years of robust growth, the PC market came to a grinding halt in Q4 2008 with PC shipments falling by 0.4 percent. In Q1 2009 the decline in the PC market was 7.1 percent in terms of shipments, according to IDC, but in Q2 2009 it was only down 3.1 percent.

Now let’s not pretend that the PC market is all hale and hearty, because it is not. Apple had a negative first quarter and Microsoft is reporting serious declines in Windows revenue. The size of the PC market was about $136bn in 2008 and the forecast overall for 2009 (from IT Candor – August 2009) is a revenue decline of 14 percent to $117bn.

So there’s a distinct difference here between the decline in PC units and the decline in PC revenues – a gap of at least 9 percent. This can be explained simply by the phenomenal success of the Netbook. A Netbook is a half price PC. You can quibble about the quality of the product, but not with the fact that it is selling well. The problem for the PC vendors is that the margins are razor thin.

Figures on corporate PC buying intentions (from Changewave) show a very recent uptick, indicating that PC volumes may rise in the final quarter of 2009 and the availability of Windows 7 may stimulate the market a little. It’s unlikely that the market will recover fast. The Netbook trend has set low price expectations for all laptop PCs and teh market will have to live with that. Indeed the market may never return to its previous glory, but nothing in any of these PC figures suggests that the pain is anywhere close to the pain in the server market.

It’s early days in the desktop virtualization market. This is  why:

1. There’s not much activity in this market. The recession didn’t prompt a whole series of companies to start virtualizing their desktops in a mad frenzy of attempted cost-saving, although it’s possible that they may do so in time.
2. VMware isn’t really playing in this market in an aggressive way, for two reasons. First it has better fish to fry and second, it’s VDI approach is currently quite expensive on a per seat basis.
3. Citrix did well to put its mark on this market with the acquisition of Ardence and Xen Source, and also a strong partnership with Microsoft. I thus expected Citrix to be able to defend its position against attempts by VMware to impinge on its primary market – client enablement – and it seems to be doing so.
4. However its recent robust financial health stems from its on-line services, not from client virtualization.
5. In the main, desktop virtualization still includes a fair amount of D-I-Y, requiring the integration of several products.

That final point is why I was surprised with what I saw when I was briefed recently by Virtual Bridges.

Virtual Bridges

Virtual Bridges is an Austin start-up with a desktop virtualization product that is an all-in-one package – not a broker that connects to some other virtualization capability (such as VMware’s VDI) or a VDI capability that needs a broker. It’s a full solution.

Technically there are just two parts: VERDE and SMART.

• VERDE, Virtual Enterprise Remote Desktop Environment, is the virtual desktop management software that creates new instances of desktops (from a standard image or images) and deploys them on the server.
• SMART, which stands for Self Managing Auto Replicating Technology, provides local instantiation of the virtual desktop, creating a local cache of the virtual desktop’s state and thus enabling the desktop to run in disconnected mode, for example on a laptop. SMART itself runs on the server but installs a hypervisor on the client.

So what do these two components deliver? Well, of course they’ll deliver a Windows PC (2000 or XP or Windows 7 when available) with either local storage (on the client device) or server side storage as desired. The PC is delivered as an image and, if disconnected, it will synchronize with its master image when reconnected so it remains in step with its master image.

The VERDE/SMART combination can work on any client, including an employee’s own PC and it will even work on a Netbook, because it has a low resource overhead. It can work in VDI mode where the server image is full mirror of the PC session, so if the PC fails, the session can be pulled down onto another device.

Virtual Bridges already has some large customers, including McDonalds and Radio Shack and it also has some large deployments including a Netherlands School District where 5000 desktops are deployed. 5000 is a large enough number to demonstrate scalability of client population.

The Defining Points

This is all reasonably impressive, but on it’s own it wouldn’t warrant me thinking of Virtual Bridges as a potential Citrix competitor. There are several other details that convince me of this possibility. They are as follows:

1. The Price Per User. Because of the way its technology works Virtual Bridges can sell at a very low price point – it claims to cost about a fifth of any other equivalent solutions, including those from Citrix and VMware. That’s a huge cost differential. The actual price per user varies according to the other licenses involved (particularly Windows licenses) but it is certainly low enough to make the competition groan.
2. Cloud-Based Virtual Desktops. Because of its architecture, VERDE/SMART is ideal for cloud deployment similar to the way that Desktone deploys (read this for further details), with clients on desktops and servers in the cloud.
3. The IBM Partnership. Other software companies are partnering with IBM in this area, but Virtual Bridges’ partnership with IBM appears to be very active, with IBM directly introducing Virtual Bridges to some customers.
4. The Linux Option. Virtual Bridges also partners with Canonical and can deliver Linux Desktops. Now you may be thinking that there’s not a lot of demand for those at the moment and in most cases you’d be right – but there is a demand in some specific environments, particularly in retail for retail kiosks which need to do little more than run a browser to access the Internet. There are also similar small groups of desktops in some organizations that have a similar requirement – just a browser and email, say. Virtual Bridges has a neat solution for that.

Taking these things together, I can imagine Virtual Bridges making an impact if it keeps its sales pipeline humming, and that it turn makes me suspect that Citrix may soon have a genuine competitor in desktop virtualization.