Tuesday, August 29, 2006

Novell disappoints - what next?

Tonight Novell posted results below market expectations (watch Wednesday share action for proof), but listening to the conference call really convinced me that Ron (the new CEO) really "gets it" about the importance of SUSE Linux Enterprise 10. I'm even more convinced now that Oracle or IBM will need to buy Novell to own SUSE before summer 2007. During the conference call, Novell talked about appreciating the subscription and consultancy business model - showing it's headed in the exact same direction as Sun Micro. Let's see if Ron can pull out the same kinds of positive turn-around results that Jon is enjoying at Sun.

Tuesday, August 08, 2006

Novell - copying Sun too?

Reading Ron Hovsepian talk on the boston.com business section about Novell reminds me about Sun's turnaround under Jon Schwartz. Here are the similarities:

  • Ron wants to simplify Novell and move to profitability
  • Ron's focus will be open source, especially his operating system
  • Ron understands the biggest asset of Novell is SuSE Linux

Compare this with Jon Schwartz at Sun and there are obvious parallels with Sun's OpenSolaris project, and recent job cuts.


So how does the Novell company situation compare with Sun? Well, just like Sun, Novell has a healthy balance sheet ($1.5bn cash) and good revenues ($1.2bn per year). At today's valuation of $2.2bn, just $60m a year profit would give Novell a 35x P/E multiple - and that's at a margin of only 5% of revenues. Ron Hovsepian shouldn't have to pull so many levers so hard to get to this place.


Going forwards how's the climate? IBM has made strategic investments in Novell, and needs a healthy Novell to avoid a Red Hat enterprise Linux monopoly. Novell's SuSE Linux is of huge importance to IBM's Power processor. And guess what - Ron spent 17 years working as an executive at IBM. Hmmm. Does it make you wonder if some big blue company might buy Novell soon?


And then there's Larry Ellison at Oracle, who wasn't happy to see Red Hat join forces with JBoss to edge into his middleware market. In an interview at Forbes.com, he said Oracle might consider taking a Linux distro and supporting it via Oracle. This had a predictable negative impact on Linux shares, which is good if you're a buyer. So maybe Larry is thinking about buying a Linux distro company?


In summary, one of 3 things will happen to Novell:


  1. Ron executes well and turns a profit - shares trade at $10 or more
  2. IBM decides that it needs to own Novell - and buys at $10 or more
  3. Larry decides to own Novell and run it against Red Hat - again $10 or more

...so IMHO it's a great time to buy Novell, assuming Ron is like Jon.

Wednesday, August 02, 2006

IBM - copying Sun with AMD servers

Yesterday, IBM announced a range of new AMD servers aimed directly at Sun's X4100, X4200 and X4600 servers. Coupled with this announcement was news that IBM is releasing 4-way AMD blades to compete with the Sun Blade 8000 Modular System. Well, they say that imitation is the sincerest form of flattery!


The downside for Sun is that IBM has managed to squeeze an extra 15% of speed out of the server RAM with patented Xcelerated Memory Technology™. Let's see how the benchmarks compare - and hope that Sun can stay a step ahead.

Monday, July 17, 2006

Niagara 2 chip - up and running already

CNET's news.com is reporting that Sun's Niagara 2 chip boots Solaris. This is at least 3 months ahead of schedule, and signals that Sun's throughput computing initiative is working well.


Niagara 2 will have 8 chip "cores" each running 8 threads (tasks). Niagara 1 runs the same number of cores, but only 4 threads per core. Competitors like HP and IBM have criticized Niagara 1 for its slow performance at floating point calculations. But Niagara 2 will have much faster floating point capabilities - with a floating point unit for each of its 8 cores.


The bigger news from a sales point of view is that a server with two Niagara 2 chips is planned. This would run a huge 128 threads (i.e. tasks) simultaneously. When set to work on web tier workloads, this specification simply blows away the competition - especially when you take the power consumption (70 watts per chip) and small size into account.


Sun's R&D is certainly shining!

Thursday, July 13, 2006

"Blades done right" - AMD CEO

At Tuesday's launch event, AMD CEO Hector Ruiz said of the new Sun Blade 8000 modular system it's "blades done right". By this he meant he was witnessing a blade chassis that supported every kind of IO (input/output), 8 core blade CPU modules, and enough cooling and power for the next 5 years of upgrades. Finally it looks like Sun might make a dent into IBM and HP's blade marketplace.


Similarly the exciting new X4500 "Thumper" storage server, and the new X4600 16-way AMD Opteron server are proving Sun's strategy to be wise. Although the X4500 could theoretically run Windows or Linux, the huge benefit of running Solaris is that is makes use of Sun's revolutionary file system technology called ZFS. For many years companies have used RAID to prevent data loss, but ZFS goes an extra mile to ensure data integrity and performance. Similarly, the X4600 server runs Windows and Linux, but by using Solaris the customer benefits from an operating system that was always designed for SMP (symmetric multiprocessing) workloads.


Sun is proving that innovation at both the hardware and software levels can dovetail into truly unique products without competition. As shareholders, let's hope the market gets the message!

Tuesday, May 23, 2006

Sun stock price - technical analysis

SUNW has been experiencing a lot of technical trading recently. The share price has broken down through resistance levels of $5.00, $4.50, and is likely to touch $4.20 before rebounding quickly to $5.00. Taking a look at the price over the past year, it appears to be leaping upwards in ever-greater bounds. Here are some factors that are likely to produce the next leap:

  • At a price below $4.50, analysts are more likely to upgrade the stock
  • Very soon, Sun and Ubuntu will announce that Linux runs fast on Sparc chips
  • I get the feeling that Sun, Red Hat and JBoss are planning something good
  • Mike Lehman, CFO, is about to announce head count reductions
  • Revolutionary new Sun storage products will be announced very shortly

So while the guys who bought at $5.00 are being "shaken out" at a price of $4.25 or so, maybe it's time to be contrarian and start to pick up SUNW at what will appear to be a cheap price in a month or two?

Tuesday, May 09, 2006

Sun and Oracle - the romance ends?

Recent news from Sun makes it appear the Oracle romance might be at an end. Sun has annouced N1 support for SAP to give it a leg up in the SAP marketplace. It is also including PostgreSQL in the next release of Solaris 10 in June. It sure looks like Larry Ellison is being given the cold shoulder.


This is good for Sun. Since Oracle never got with the open source thing the romance has been on the rocks. The old days of Sun E25k servers running huge Oracle databases have gone. The zeitgeist is with FOSS and that's where Sun is headed.

Friday, April 28, 2006

Sun and Google - an opportunity

Sun sells services, servers and storage. Google sells search, direct to the consumer. But Google is keen to get into the Enterprise Search business. So here's a great idea for Sun and Google - work together to deliver enterprise search tightly integrated with resilient fast storage. Sun's new StorageTek 5800 Content Addressable Storage system is a fine candidate for Google Enterprise Search integration. As Sun investors, let's hope it competes well against HP's RISS and EMC's Centera products.

Tuesday, April 11, 2006

J2EE and FOSS

Yesterday's news of JBoss being bought by Red Hat for about $400m sent Sun shares skidding badly. It seems the world knows what Sun can't admit - Sun needs to provide a popular Jave 2 Enterprise Edition (J2EE) server, to grow mindshare and market.


JBoss is popular for a number of reasons:


  1. It's free and open source software (FOSS)
  2. It includes a number of excellent technologies (e.g. Hibernate)
  3. It is backed by the JBoss service organization


When Jonathan Schwartz talks about building communities, this is what he must be aiming at. JBoss is a fine example of a successful open source community. As Sun invetors, let us hope that Sun can emulate this in the very near term.

Sunday, April 09, 2006

Sun vision

The Economist recently ran a story about Sun's history through the boom and bust. I believe it's a fair critique of Scott McNealy's direction of Sun in the days before Jonathan Schwartz revolutionized the company with per-employee support pricing, FOSS and true utility computing. The article highlights how Sun lost ground when it only offered a "my way or the highway" solution to customers. But things are looking very different today with Sun's AMD Galaxy servers and new throughput computing machines.


Last week, Sun announced its new CT900 telecoms blade server. This was a smart move. IBM and HP currently dominate the general blade market, so Sun is going specifically after its best vertical market - the telcos. By the end of the year, Sun's telco blades will also feature the T1 processor which is perfectly suited to telco applications. I have a good feeling about these new telco blades.


The new Sun Microsystems is all about smaller, faster, eco-friendly servers. For example, their recent aquisition of Aduva makes it far easier to maintain large numbers of small servers.
So it's no surprise that their Scalable Systems Group recently laid off 7% of staff. This is a sign of clear direction and cost control. Mike Lehman is delivering profitability as he takes control of company finances.


When the Economist says "Unfortunately, none of this visionary stuff seems to be helping Sun", it appears to be absolutely wrong in so many ways. I believe that investors can expect good news when Sun reports 06 Q3 results on April 24th.

Wednesday, March 29, 2006

Commodity pricing

Sun's Grid is priced at $1 per CPU hour. So you might expect that you could run 60 jobs each taking a minute for $1 - right? Unfortunately, as Clint Combs discoverd, this would actually cost $60! Instead of adding up the aggregate CPU duration for the day/week/month, the system currently rounds each job up to the nearest hour (i.e. dollar).


This might sound unimportant, but it fundamentally changes the market of potential Sun Grid customers. For example, my Site Stats business generates between 500 and 1000 web traffic reports per day for customers. If I could submit them all to the Sun Grid for processing, and it took say 5 hours of CPU time then I would like to pay $5, rather than $1 per report - a whopping $500 to $1000! I believe my scenario is typical for thousands of other small businesses.


Sun are looking to change their grid pricing, so those Web 2.0 start-ups can run a much finer granularity of grid job over the network. Imagine if your electricity provider charged you 10c per KwH, with a minimum charge of 10c for every light you switched on. Now that wouldn't be commodity pricing would it? To really follow Jon Schwartz' electricity analogy, it's time to start billing for aggregate usage over a period of time.

Monday, March 20, 2006

www.network.com

I've always been a fan of good domain names, my favourite being purple.com. I believe a well-chosen domain name says more than what it does - it somehow embodies the spirit of what it delivers. According to Jonathan Schwartz, Sun will imminently launch www.network.com as its publically available Sun Grid utility. Wow! In 10 years, I expect to look back on this event as an epoch in the world of network computing.

Tuesday, March 14, 2006

Novell

Having seen Novell's SuSE Linux Enterprise Desktop (SLED) version 10, I must say it looks much better than Sun's Java Desktop System. It comes with VBScript compatible Open Office, plug & play, bluetooth and wi-fi compatibility, and a Microsoft Windows Vista look-and-feel.


For me, this just adds yet another compelling reason to buy Novell. The company's market valuation is below $3BN, but since it has $1BN in the bank a purchase would really cost just $2BN. And what are the assets?


  1. SuSE Linux - the big prize
  2. Mono project - Microsoft integration
  3. IBM customers - a large market


Maybe $2BN nett is a large price to pay for SuSE linux, but Novell has done a poor job of promoting the popular Linux distribution against Red Hat - surely Sun could do better? And imagine running SuSE on all Sparc hardware? Now there's a market!

Monday, February 27, 2006

2006 - Sun's tipping point?

I've been reading around the blogosphere to see how data center customers perceive Sun's product and service line-up. I'm coming to believe there's a good chance that 2006 could be Sun's tipping point. Finally, Scott might enjoy his iPod moment!


Here are a couple of comments I've read recently:


  • "People fail to realize that when you want a 10tb database there are few vendors who can compete with Sun. I find our company which is globally aligned with Dell on the PC platform is trying to find a loophole to get Sun to replace their pc servers. Sun is closer then ever to having a platform at the right performance/price level for every level of computing," from Alec.
  • "Without doubt, the X4200 is the first X86 system on the market that I've seen that really lives up to the Sun logo on the box. Enterprise class, from end-to-end, without compromise and all with an amazing price point," from Ben Rockwood.

But there's a fly in the ointment. If you read Jon Schwartz' blog comments you'll see that the Sun sales force isn't keen to engage SMEs. Many people complain they never heard back from Sun sales people about trying a T2000 server. I get the impression that Sun sales needs to make a culture shift from Fortune 500 to "the rest of the computing world".


And finally, it appears public Sun Grid is about to go live according to InfoWorld. Jon and Scott blame US munitions controls, but I can't see how that could delay this thing for a whole year. I am convinced it will turn out to be the biggest and best thing Sun has ever delivered.


Maybe we've arrived at Sun's tipping point?

Wednesday, February 22, 2006

Sun and Linux

Sun's relationship with Linux has been tepid at best. But now it looks like the water is getting warmer.

The Register features an article about Sun using Xen to run Linux on Sparc and News.com has similar news about Xen efforts.

Couple this fevered software development with the OpenSparc GPL'd T1 chip and recent acquisition of a Linux/Solaris patch management firm and it really looks like Sun is warming to Linux.

It makes perfect sense. If Jon and Scott are really serious about increasing customer choice, then providing Linux compatibility on their Sparc chips is absolutely necessary. Look what it has done for IBM's Power chips!

Check out this boot log of Ubuntu Linux running on a cool new T2000 server. Very nice work Dave Miller, programmer extraordinaire!

Saturday, February 04, 2006

Sun isn't pushing for a Sun Grid future

Reading Scott McNeally at News.com, he is quoted as saying of the Sun Grid "Ultimately, our preferred answer is that we become a wholesaler. We don't want to be in the subscriber management business or constomer care business. We're not structured to answer the calls. We don't have big billing engines to do microbilling. We don't have a consumer brand." WRONG!


Sun's Grid needs to be exposed as a web service, just like the Google API. Only then can Sun proclaim "The network is the computer". Microbilling is not hard. Nor are customer services if you've got a good system up and running reliably. And it's not about being a consumer brand - it's about being a truly next-generation internet brand for the developer community so cherished by Jonathan Schwartz.


If, say, 20 or so large Sun partners used the Sun Grid as a web service to deliver their own services (e.g. medicine, groceries, banking), then Sun would effectively be a wholesaler, but via a secure network API. It seems crazy to force these customers to continue to buy the hardware, when they would rather buy the computing?


And since Greg Papadopoulos says "It's a lot harder than we thought" how can Sun expect service providers to achieve commodity grid computing provision to the masses? Selling kit to service providers isn't going to make a grid, but it would improve the company's profits: short term win, but a mid/long term loss.


Here's a 3 point project plan to take Sun to "network is the computer":


  1. Deliver your public grid (with user accounts and billing)
  2. Expose it as a web service (a la Google API, Amazon API, etc)
  3. Promote it via sales channels

The result would be that Sun would become the new Lego for every kind of internet development on the planet. Now there's a market opportunity!

Thursday, January 26, 2006

The Big Picture

When you consider Sun has annual revenues of $13bn, a quarterly loss of $200m isn't so bad. And when you consider Sun spends $2bn on research annually, you can see where the loss comes from.

So is the research worth it? Given that the T1 Niagara processor is world-beating at web workloads, and that its successors are on track for 2007 delivery, things are looking good for Sparc. The new Galaxy servers are gaining momentum. Sun is 6th in shipments of X86 servers, pushing for 4th by end of year. And Sun is talking about a return to Blade servers before the summer. Now that's a market Sun would like to shine in!

So maybe the research is beginning to pay back? It looks like the market is reading things this way, since the share price is holding firm despite the loss. Personally, I'm on the lookout for ZFS and new storage technology, but that's just a rumor...

Tuesday, January 24, 2006

Disappointment

Now I understand why I'm not a paid analyst. They said Sun would post a loss and they were right. Sad. Sad. Sad. Scott - wake up and smell the Java! Profit is not a new fashion - it's your raison d'etre.

Monday, January 23, 2006

Earnings

Today Sun posts its quarterly results. Most analysts predict a 1c loss per share. I am predicting a 2c profit per share based on new product growth, Sparc stability, and Storagetek integration. Let's see who is closer to the actual posted result. If anyone is interested, I'm happy to consider consultancy to any major investment bank. (I have history with UBS).

Thursday, January 19, 2006

AMD revenue begets Sun profits?

AMD has posted a 45% increase in revenue to $1.84 billion (an increase of $570 million). How much of this can we attribute to Sun's Galaxy servers (X2100, X4100 and X4200 SunFire servers)?
Let's take a low, medium and high estimate to guess the impact to Sun's results due out on January 24th.


Optimistic


Assuming 50% of the AMD revenue increase was due to Sun servers, that means Sun spent $280 million MORE on AMD Opteron chips. Now assume that this accounts for half the unit cost of a typical server, which means Sun shipped an extra $1 billion in Galaxy servers, of which half is profit (if Sun maintains its margins). Tasty!

Realistic


Let's say 25% of the AMD revenue increase was due to Sun servers, meaning Sun spent $140 million extra on AMD Opteron silicon. Assuming it accounts for half the unit cost of a typical server, Sun shipped $500 million or so in its Galaxy line (assuming 50% margins).

Pessimistic


If only 10% of the AMD revenue jump came from Sun purchases, then we're looking at just $50 million of extra chip spend, equalling say $200 million in Sun revenue incease.


Whichever way you look at it, Galaxy is driving profits at Sun unless Sparc revenue is losing ground to Power and Itanium. Today, news.com say that Itanium is losing backwards compatibility with X86, except via emulation software which takes it down to current Xeon speeds. And with the new T2000 Sparc servers looking so tasty, Sun may finally be IN THE BLACK. Buy SUNW before 24th!

Friday, January 06, 2006

2006 is looking good for Sun!

In the past week, Sun has received some excellent feedback from all quarters:

The Register


Their Reader Study (free PDF) ranks Sun above IBM, HP, Microsoft and Dell. They even quote a Sun customer as saying "Just reinvented my world with Niagara - I'm having to wipe the drool from my keyboard as I type this."


It's not all good news though - the C-level decision makers are not as keen on Sun as the rest of their organisation staff, so there is a communication issue for Sun to look at.

eWeek


In their Top Products of 2005 report, eWeek include Sun's Solaris 10 operating system, and Java Studio Creator 2 Interactive Development Environment (IDE). That's a really big achievement against some excellent Linux distros (e.g. Red Hat) and the popular IDE Eclipse.

InfoWorld


In their 2006 Technology Of The Year Awards, InfoWorld award the new SunFire T2000 Server as the most innovative server of the year. (Unfortunately for Sun, IBM and HP pick up the other 2 server awards).

And the market?


Well, all this good news and a sprinkling of speculation pushed the price of SUNW stock to a close of 4.71 on Friday - not bad at all! Expect further rises with good news about Open Document Format (ODF), ZFS and other technologies as Sun begins to find its stride and post some profits.

Wednesday, December 21, 2005

Red Hat makes tons more profit than Sun

The Register's story Subscriber love drives Red Hat in Q3 details Red Hat's surging revenue and profit - respectively to $73m and $23m for the quarter. Red Hat's excellent execution of the Red Hat Network (Linux update/distribution service) and its related subscriptions have returned excellent numbers.

Now it's Sun's turn. Solaris really needs the ability to run Red Hat Enterprise Linux, and at a lower price than a Red Hat subscription. It sounds crazy, but it's both possible and legal. First, Solaris needs "BrandZ" containers which can host any compatible operating environment (e.g. Linux). Second, you need a free copy of Red Hat - and here's where CentOS comes in handy. From information at InternetNews.com, it appears Sun is hosting CentOS via BrandZ containers.

To summarize, the ABC for competing with Red Hat is:
A: AMD Opteron server (this won't run on Sparc)
B: BrandZ containers in Solaris
C: CentOS, the free Red Hat Linux clone

...then the revenue (profit?) comes from the service contracts with Sun.

IBM - Sun's biggest threat

In 2006, Sun's biggest competition will come from IBM. For example today's CNET news.com article IBM to acquire Micromuse for $865 million shows how IBM will compete against Sun's N1 Software for network service management. It might be time for Sun to give N1 a "shot in the arm" and potentially a rebrand since the N1 name is not hot.

Other IBM threats on the horizon include:


In these key areas of Throughput Computing, Office Software and Middleware, IBM has a strong pedigree. I believe that Sun can beat IBM by leveraging its much touted but under-utilized Sun Grid. Example applications for the Sun Grid:

  • Manage my servers (healthcheck, patch, load-balance, etc)
  • Host my office documents, so I don't need to
  • Provide a J2EE container environment for my business

2006 sure will be fun to watch!

Thursday, December 15, 2005

OpenOffice and StarOffice

ZDNet reports that Sun is going to give OpenOffice and StarOffice the ability to open and save documents over the internet. Jon is quoted as saying "OpenOffice is simply another network client. Being able to just save to the cloud, using WebDav, is the direction that we're heading. If it's not in there now, I can take you back to Sun and show it to you."

This is VERY GOOD NEWS and provides yet another perfect application for the Sun Grid. It also helps to build momentum behind OpenOffice and the Open Document Format (ODF). If Massachusetts makes a move to open formats, expect to see many other companies and agencies follow suit.

Tuesday, December 13, 2005

Today I installed Solaris 10 on my Sony Vaio Laptop. It reminded me of having a blood test. Not particularly painful, but unpleasant. I felt relief when it was over.

Why did I do this? Well, many sources tell me there are millions of downloaded copies of Solaris, but are they getting used? Did they morph from download to install, with a potential to become a support subscription? From my experience this "path to profit" could be easier. For example, I would have liked more options to change my installation choices, and a clearer explanation of what the installer was waiting for when things appeared to stall. To summarize: "good work but could do better".

On the upside, OpenSolaris is looking very tasty indeed, with its hugely important ZFS filesystem and BrandZ containers for RedHat Linux. The sooner this becomes production-ready, the sooner the Linux wave moves over to Solaris.

Thursday, December 08, 2005

An idea for a "Service Oriented Architecture" world

I was reading the Sun Grid overview and the thought struck me:

How come I can't use the Sun Grid as a regular web service?



What I mean is, if I can make Google searches via web services (SOAP) and check Amazon book prices (REST), then how come I can't use a computer program to do this:

  1. Sign in with a token
  2. Upload my data
  3. Upload a job to run
  4. Run the job
  5. Check the status until ready
  6. Download my results


Come on Sun - remember "the network is the computer" right?

Great execution

In my last post, I questioned Sun's execution of their grid service - since it's still impossible for anyone to sign up for an instant-access account via the web. But Scott and Jon's unveiling of the new T1000 and T2000 servers was literally spectacular.

I went to see Jon launch the "pizza boxes" at the Hilton Met in London. You couldn't help but come away with a deep feeling of satisfaction (as a shareholder), much akin to eating a perfectly cooked large rib-eye with a good bottle of red. This was definitely a "turn-around moment", and possibly one of Scott's iPod moments.

Here's what needs to happen now:

  1. Industry role models (e.g. eBay) need to give glowing case studies
  2. The "try and buy" offer needs to be simple and genuinely taken up
  3. The PR hype needs to be sustained, and amplified (e.g. Red Hat on Sparc?)


As other industry analysts have noted, if the "T-machines" can stabilize or even grow Sparc revenues, then the Galaxy X64 servers can drive big profit numbers. A FY2006 number of $500m or more is now a possibility. When the analysts improve their "hold", "neutral" and "underperform" opinions, expect to see SUNW climb to $5 and beyond very soon.

I just wish I had been able to buy more when they were cheaper - a nice problem to have.

Thursday, December 01, 2005

Free and Open Software (FOS)

Jon Schwartz has long been a fan of Free and Open Software (FOS) and yesterday announced that Sun will bundle Java Enterprise System and Developer Tools with Solaris, to make the Solaris Enterprise System.

This was poorly received by the markets, as Sun stock fell hard on the news. But today things are rebounding - so either it was a technical price movement, or the analysts are waking up to the fact that it really does make sense.

The challenge for Sun is to persuade business customers to subscribe to support services. If they don't, they will get high adoption with no/low revenue growth. This is where Red Hat excels. Red Hat only provide download access to Red Hat Enterprise Linux (RHEL) via paid subscription per server. How will Sun's "per employee" model compete with this?

Monday, November 21, 2005

The Sun strategy in a nutshell

Here's my super-simplified four point Sun hardware strategy analysis:

  1. Get Solaris (and OpenSolaris) adopted by loads of developers and businesses
  2. Make computers that run Solaris faster than anyone else
  3. Get those same developers and businesses to go buy the fast servers
  4. Deliver next-generation storage solutions, to complement the fast servers


Here's my super-simplified four point Sun service strategy analysis:

  1. Make the Sun Grid available to all
  2. Leverage Tarantella to offer hosted desktops
  3. Sell more Sun Ray "thin clients"
  4. Sell more JES subscriptions, with See Beyond web services technology

Sunday, November 20, 2005

Where we're going

Since Jon Schwartz became President and COO of Sun, they've gained a clear vision, strategy and direction. The upside is that they know where they are going - the downside is the long slow wait to get there. Jonathan appears to have a keen grasp of economics and industrial history - and these attributes bode well for Sun's future. He clearly believes that computing will become a utility resource, like electricity and water in our offices and homes. He also realises as the price of software trends to zero, the value (profit) will come from services and hardware.

But here's my first problem. In February 2005, Jon proudly switched on the Sun Grid but I still can't use it. Now that's poor execution. Similarly, in a Q1 earnings call, analysts said customers could not get hold of new "Galaxy" servers. It will probably be the same with "Niagara" servers in Q2.

My second issue with Sun is usability. Although Jonathan is pushing the value of simplicity through the organization, much of Sun's wares and web sites are HARD TO USE. Let me give you an example - try downloading Sun Studio 11 for Linux. You'll have to jump through a few hoops. Even if you just want a faster C++ compiler, you'll probably need to install Java, NetBeans, etc. Why the unnecessary complexity? If Google can give anything to Sun might I suggest interface design would be a good starting point?

How we got where we are

Let's start out with the Big Picture. Around the year 2000, something big happened. Just as the internet bubble was bursting, Google showed the world that you could run your business on a number of standard PC computers connected to the internet. Open Source projects like Linux, Apache, Sendmail and MySQL showed that you could get hold of high quality software for free. This was bad news for most people who sold big computers and expensive software. It was especially bad news for people who sold web servers. Sun collapsed from a peak of over $60 a share to today's $4 level.

Most companies were quick to adapt to this disruption, but Sun decided to "tough out" the storm. It didn't work. Sun gradually down-sized, but too slowly to "get underneath" its large cost overhang. Gradually Sun reduced its head count to 30,000 and this year recovered to break-even while maintaining a 2 billion R&D spend. Quite impressive, but profit is what counts right?

Has Sun reached a turning point? Can it deliver profit to shareholders? Well, the signs are good. Stay tuned for a detailed analysis of new Sun technologies and services, such as OpenSolaris, ZFS, Niagara (AKA T1 chip), Chip MultiThreading (CMT), grid/utility services, and most importantly storage.

Welcome to Sun Visor!

I'm an investor and I choose to invest in Sun Microsystems (NASDAQ: SUNW). Right now I hold 27,500 SUNW shares at a price of $3.75 per share. Over the coming months (maybe years) I'll be providing a Technical Analysis of Sun's products, services, communications, competitors and employee compensation. I'll be providing objective criticism with the intention of improving Sun's products and services - hence improving the value of my shares. Of course, if I feel that Sun is "going down" I'll be sure to close out my position and reflect my bear-market opinion in due course - but let's all hope for a bright future for Andy, Jon, Scott and all the guys and girls at Sun.