So Oracle ends up owning Sun Microsystems.  I couldn’t believe it at first, thinking somehow that it was all just a ploy to get IBM to pull out the Big Checkbook.  And while the deal may have begun with that thought glowing in the mind of Jonathan Schwartz, it ends with the heart of Sun moving a few miles up 101 to where it will certainly die.

IBM doesn’t want Sun and is gleeful with the idea of Oracle taking over, as you’ll learn if you read the internal IBM memo copied below.  Big Blue does a very good job here of explaining its thinking and most of it makes sense.  No white knight.

But what will Oracle DO with Sun?  Make a lot of trouble for IBM, or try to, I think, but even doing that will be a challenge.  Java is deliberately unprotected by patents and subject to enough industry oversight that Larry Ellison can’t just kill it or somehow make it proprietary overnight.  MySQL could be killed, but for Open Source that just means it would branch and be reborn a day or a week later mostly intact and protected by nerds who would by then be very, very angry.  On a positive side Oracle will undoubtedly make some very useful database appliances and may well come to dominate that as yet non-existent product space.

But for the most part what Oracle will do with Sun is show a quick and dirty profit by slashing and burning at a produgious rate, cutting the plenty of fat (and a fair amount of muscle) still at Sun.  If you read the Oracle press release, the company is quite confident it is going to make a lot of money on this deal starting right away.  How can they be so sure?

It’s easy.  First drop all the bits of Sun that don’t make money.  Then drop all the bits that don’t fit in Oracle’s strategic vision.  Bring the back office entirely into Redwood Shores.  The cut what overhead is left to match the restructured business.  Sell SPARQ to some Asian OEM.  Cut R&D by 80 percent, saving $2.4 billion per year.  I’m guessing sell StorageTek, maybe even to IBM.  And on and on.  Gut Sun and milk what remains.

The plan has to have been on the table since last Fall when Andy Bechtolsheim, the mine canary of Sun’s executive suite, left the company for the second time.  Even then it was clear that the options were a good sale or murder-suicide.

I blame Schwartz, of course, but I don’t blame him, too, because I think he had little choice.  He just wasn’t a lucky guy, it turned out.

So what’s next for Sun?  Nothing, I think.

Here’s the internal IBM memo on the deal:

PPublished on 21 April 2009
News home > Top stories >
Oracle enters a twilight zone

The acquistion of Sun creates opportunities for IBM.
The surprise announcement of the Oracle deal to buy Sun Microsystems creates
some new opportunities for IBM. Since its days as a bright star in the dot-com
era, Sun gradually lost its place in the UNIX server market it once dominated.
How does IBM stand to gain, if and when this transaction closes?
Momentum
First, momentum. According to IDC’s latest report, IBM’s share of the $17 billion
UNIX server market grew to more than 37% in 2008 while Sun’s share fell to 28
percent; since 2000 IBM has gained 19 points of share, while Sun has lost 7
points.
Since 2006, the number of clients that have migrated from Sun to IBM Power
Systems has grown 10% annually to more than 750 clients as of 1Q 2009. IBM’s
Migration Factory has eased the transition for these Sun clients to the Power
platform with its leadership performance and virtualization technologies. In
addition, IBM technologies such as its Infosphere Information Server have enabled
a steady stream of Oracle clients to migrate from Oracle’s high-maintenenace-fee
database to IBM DB2 and Informix data servers. The fact that Oracle and Sun will
share the same address does nothing to change these trends.
Openness
Second, openness. IBM offers every client the greatest choice and best value in
both hardware and software to meet their business needs. IBM will continue to
support Power Systems clients that have chosen Oracle’s middleware or database,
just as we will continue to support the IBM middleware and data server needs of
Sun server clients.
Oracle, after acquiring many software vendors partial to IBM server platforms, has
long promised to protect the compatibility of IBM servers, notably Power; Oracle
clients will continue to demand this compatibility moving forward.
Oracle’s self-serving interpretation of “open” sharply contrasts with IBM’s
championing of Linux and the broad open source community. Despite this, clients
committed to IBM middleware have forced Oracle to maintain long-term
compatibility with IBM software through previous Oracle acquisitions of IBM
Business Partners such as PeopleSoft and Siebel, and this bodes well for Java

News
4/21/2009

http://w3.ibm.com/news/w3news/top_stories/2009/04/stgswg_sunoracle.html

technology. Oracle is unlikely to make sweeping changes – it’s the subtle changes
we’ll watch for. MySQL, an open-source competitor to Oracle’s database that was
acquired by Sun last year, should pose an interesting test of Oracle’s openness.
Sun’s billion-dollar acquisition was hurting Oracle. If they kill MySQL they could
alienate the open-source community, which loved Sun. If they keep it, they may
not have the ability to capitalize on it.
Client confidence
Third, client confidence. IBM’s consistent roadmaps and disciplined delivery enable
clients to effectively gauge the long-term value of their investments in systems,
middleware and services. Sun’s much-publicized business problems will not be
erased in the minds of clients by the Oracle acquisition. If anything, significant
questions are raised. For instance, the omission of any mention of SPARC in
yesterday’s statements from Sun and Oracle is certain to make Sun hardware
loyalists very anxious about a future where Oracle is calling the shots.
Earlier this month the latest of a long line of executive departures from Sun was
its lead processor designer who headed development of Sun’s long-promised and
much-delayed next-gen RISC processor, codenamed “Rock.” The future direction
of Solaris in the hands of Oracle is also unknown, while IBM’s substantial
investments in the future of AIX and Linux are ongoing and well known.
Cost advantage
Finally, cost advantage. In today’s economy, clients are looking to reduce the
heavy maintenance costs associated with Oracle database use. IBM hardware and
software technologies together provide a significantly lower total cost of
ownership.
Several Wall Street analyst reports yesterday saw Oracle’s move as defensive in
response to a dwindling ecosystem. Other observers see the Sun deal as an
attempt to emulate IBM’s successful solutions strategy. However, Oracle’s ability
to manage this type of integration is unproven. Oracle’s remains an application-
led play while IBM has a thriving software ecosystem of application developers
and a much different acquisition style.
Whatever changes take place within Oracle and Sun, one thing that remains
unchanged is IBM’s position of strength and our proven ability to win against both
these competitors.
More details on the Oracle-Sun announcement are posted on the Market Insights
Web site .
For more information concerning this article, please contact Smith, Bruce P.
(brucesmi@us.ibm.com).

Podcast: Play in new window | Download

This week more than 20 million people watched on YouTube and other video sharing sites a single performance from the ITV show Britain’s Got Talent in which a frumpy spinster from Scotland sang like an angel. You can see her astonishing performance here.

It’s not the singing that makes me write this, though the singing was good. I lived as a boy in the north of England and knew ladies like this Susan Boyle. What makes me write about it is the effect she and her singing had on the Internet and the Internet in turn had on the performance and its aftermath.

The video file as presented on YouTube is just over seven minutes and 26 megabytes long. Twenty million (and counting!) times 26 megabytes is 520 terabytes or approximately half the size of the Internet Archive. That’s 520,000 gigabytes or the equivalent of maxing-out in a single week the monthly bandwidth allotment of 260 co-lo servers at Rackspace.com. Running at top speed for a week would require 1040 such servers to do the job and we haven’t even made it to a week yet. That’s 520 million-million bytes.

Okay, so it was a nice lady singing a nice song, but what’s astounding is the performance had been round the earth twice or three times before the broadcast in the UK was even over. It was one of those seminal moments of mass-communication that showed the world was different than it used to be and thank God it didn’t require a wardrobe malfunction to do so.

What resonated with audiences about this performance was that it hit everyone – everyone – the same, as a long-coming reward for a life of good cheer and choir practice. I make documentary films from time to time and this performance is one of those emotional moments that every documentary director dreams of. It’s not the facts, you see, or even the stories that matter, it’s the emotional state of the people on-screen and how the viewer relates to them that matters. Real feelings count.

And thanks to the Internet in this instance such feelings count everywhere, it seems. For one happy moment we’re drawn together as a single audience to share a single emotional high that involves, for a change, no losers at all.

Think how rare that is, which explains its power.

Marshall McLuhan, who seems smarter every day, called it The Global Village. He said communication technology would link us together in ways we couldn’t imagine and those ways would lead to common experiences and shared values. McLuhan didn’t know about the Internet when he wrote that and he sure as Hell didn’t know about Twitter. But his prediction came true.

This Susan Boyle experience doesn’t come along very often, but with the growth of broadband technology it can’t help but happen more and more. It’s not the Super Bowl or the World Cup — it’s better. That’s because it is personal – a moment we all can share, well so far 20 million of us, one at a time.

Now the folks at Google are no doubt scratching their heads, as are the TV producers back in the UK, trying to figure how to put this effect in a bottle and make a living from it. But it can’t be done.

This is an event that was created for TV but not really anticipated by its creators, I’m guessing. They couldn’t reliably repeat it if they tried.

If they did try, it wouldn’t work.

That’s the beauty, because every time this happens, every time our Global Village comes together in this way, it’s because of a shared delight that makes us feel more alike and less apart.

We could all use more of that.

And the next time it happens, now we all know what to do.

Podcast: Play in new window | Download

When Barack Obama was running for President one of his favorite sound bites was that any financial bailout should not just involve Wall Street, but Main Street, too – that the government’s responsibility was to help both bankers and homeowners. But now that the election is won and Obama is in office, the two streets are still being treated very differently, with Main Street getting a lot less help from Washington.

This is a HOUSING crisis, not a BANKING crisis, yet $700+ billion has gone to help bankers and only $75 billion to “help” homeowners. The banker’s money has mainly been spent and the homeowner money has hardly been touched. If this is a HOUSING crisis, why aren’t more resources being devoted to housing?

It comes down to an issue of morality, believe it or not, with homeowners expected to be moral and bankers not. Everybody blew it, but the homeowners are being disproportionately punished for their actions.

There is no morality issue in the bank bailout. Banks are having their capital boosted based not on whether they are well run or in some way “deserving,” but purely on the basis of whether they are viewed as being in three groups: 1) doomed; 2) capable of being saved through injecting government funds, or; 3) too big to be allowed to fail no matter how poorly run. This means the least-deserving banks tend to get the most help.

But the Obama Administration’s attempt to help mortgage holders is different. If you hope for government help in restructuring your mortgage you’d better not be behind in your payments. If you missed a mortgage payment months into this crisis, you are out of luck. If your mortgage isn’t guaranteed by Fannie Mae or Freddie Mac, you are out of luck. If your mortgage is jumbo you are out of luck. And if you owe more than 105 percent of the value of your home you are out of luck.

That’s a lot of homeowners out of luck. No wonder the Obama Administration thinks it needs only $75 billion to do the job, it is excluding so many people.

Let’s try applying the homeowner rules to the banks. If both played by the same rules, then banks with mortgage portfolios that have dropped by more than about 15 percent (are five percent or more underwater) would be ineligible for government assistance. Banks that MADE jumbo loans would be ineligible for assistance. Banks that made loans with private insurance or no insurance would be ineligible for assistance. Banks that had shown themselves unable to meet capital requirements (had effectively missed a payment) would be ineligible for assistance. In each case, these criteria define EVERY bank that has received assistance. They ALL have mortgage portfolios down in value by 15 percent or more, ALL made jumbo loans, ALL made uninsured loans, and ALL are under capitalized.

So if we apply to banks the same rules that are being applied to homeowners, then no banks deserve support and there should be no bank bailout. Well that can’t be, can it? So screw the rules, screw the idea of there being a moral issue with bankers, just start handing out cash without even requiring that they use any of it to make or restructure loans.

So that’s what the Treasury and the Fed have done – bailed out the bankers without regard to their past OR FUTURE behavior. And $700+ billion later do we really truly feel better as a result?

Hell no we don’t, because we still can’t pay our mortgages!

This bailout is broken, it is unfair, and it is incredibly inefficient as a result. The bank bailout is based entirely on providing INCENTIVES to the banks – bribing them to THINK ABOUT doing the right thing.The government won’t MAKE the banks do anything. They just ENCOURAGE the banks by giving money.

Where are the incentives in the much smaller housing bailout? There are incentives. THEY ARE ALL BEING GIVEN TO THE BANKS. It is very difficult to find in the new Federal mortgage modification rules much of anything that truly helps homeowners. Banks aren’t REQUIRED to do anything; they can reject any mortgage holder for any financial reason. The banks are PAID to restructure the mortgages and the way those mortgages are being restructured (primarily through increasing term and adding balloon payments) not only costs the banks nothing, it tends to make them MORE money over the life of the loan.

So that $75 billion allocated to modifying mortgages and keeping people in their homes, how much of that $75 billion will actually go to homeowners? About 25 percent, or $18 billion almost entirely in first-time buyer tax credits. This means the bank bailout isn’t $700+ billion, it is $758+ billion or FORTY-TWO TIMES the size of the housing bailout.

And why only first-time buyers? What makes them more deserving of help? The theory is that these are new homeowners so they’ll be buying-up excess inventory and helping to firm prices. They aren’t people selling one house to buy another. In another view they are virginal and uncorrupted by the housing bubble.It wasn’t their fault, so they are being rewarded. More morality, inequitably applied.

Main Street isn’t doing very well under this policy. Main Street is being cheated.

This is a bad plan, unfair and poorly executed. It places a moral burden on individuals and not on banks, yet there is no good explanation for why it has to be so.

What is it about banks that make them deserving of 42 times as much support as your Mom?

Nothing.

Like the Bush Administration before it, the Obama Administration has a bias for helping Wall Street. They couch this as a claimed inability to come up with any better ideas. Yet better ideas – ideas NOT couched in moral argument (or more appropriately couched in EQUAL moral justification) were presented right in this spot in the post titled The Not So Bad Bank. That’s a plan that helps banks and homeowners equally, doesn’t require incentives to work, acts faster, and costs a tenth as much.

What’s wrong with doing the job better, faster, and cheaper?

I recently participated in a conference for financial bloggers at the Kauffman Foundation in Kansas City, MO.  Kauffman, if you haven’t heard of it, is dedicated to the promotion of entrepreneurism and supports more economic research than any other foundation.  It is a fabulous place and I really enjoyed the conference.  For some reason they felt inclined to interview me, too.  There are plenty more interviews to be seen at the Kauffman web site (interviews with really smart people, too, not just folks like me): Kauffman Conversations

Podcast: Play in new window | Download

I’m not the biggest fan of Sun Microsystems CEO Jonathan Schwartz. Okay, I am not a fan at all. But I have to give the guy credit for keeping up company morale, because when I polled my Sun contacts recently on why they thought IBM might be interested in buying the company, each thought it was because of his or her division.  What charming — if misguided — loyalty.  These people still feel good about their company.

Of course they are wrong to do so.

I know that Sun walked away from the deal, by the way.  I’m not THAT out of touch with the world. I just think that understanding the Sun mindset might help explain that bonehead move.

So here’s the result of my unscientific sampling of a too small group with the question being “Why would IBM want Sun?”  Remember these answers came from inside Sun, itself.  Notice nobody said “servers”:

1) StorageTek. There’s a big market in z/OS storage still and Sun has
been stepping on their nether regions with it ever since they bought
STK. As depressed as Sun’s price has been, I can imagine (with no
numbers to back it) that getting that track back into competition with
the big storage guys could be appealing.

2) Solaris. There are several ways that Solaris really is the best
UNIX derivative around, and AIX just doesn’t have all that big a share
of the market. Open source seems to be winning in general in that
world. AIX shares code with some pieces of AS/400 though, so IBM
really doesn’t want to open the source to AIX. AIX doesn’t do intel
well. Solaris in a nevada-plus version could replace AIX, with Andrew
support added.

3) Sun Labs and Java. Sun is better at looking *cool* than IBM, IBM
wants to buy that. Even further, IBM has made a huge investment in Java that has paid-off in enterprise application success.  Then Big Blue aoke and realized its future in this market was totally cpongtrolled by Sun – a company IBM viewed as being, at best, adolescent.  IBM wants to control its Java destiny.

4) R&D in general.  For all its failings, Sun has continued to spend about $3 billion per year on R&D, which is nearly the adjusted price IBM would be paying.  There has to be some cool stuff in there.

5) Fishworks/Amber Road, Mike Shapiro’s storage appliance. This system is clever but not THAT clever.

6) Services. Hardly. Not even close.

Schwartz and Sun have a pretty high opinion of themselves despite years of losses.  So they walked away from IBM’s last offer.  But don’t expect them to stay away for long.  There simply isn’t another buyer for the whole thing and Sun doesn’t have the patience to cut itself into pieces.

And the real price IBM will be paying is at least $2 billion less than you read in the newspaper, because Sun – like a lot of tech companies – has profits stashed overseas that it has been unable to bring back.

IBM knows how to play that game, too.

Podcast: Play in new window | Download

Last week Google revealed to the world its shipping container modular data centers that I was the first to write about almost four years ago. I was invited to the event and expected to be there until the parking brake on my 34-foot Winnebago motor home let loose on a slight slope during my trip to California. Though I jumped in the way trying to stop with superhuman strength the half-mile-per-hour collision with a 2000 Ford Excursion SUV, the crunch happened anyway and I was a day late and about $4000 short getting to Google.

There’s not much about the container strategy that I didn’t already write years ago. They use standard shipping containers fill them with 1000+ homebuilt servers cooled with both water and air. As countless stories pointed out last week the servers are standardized with two CPUs, two hard drives, and a backup battery that negates the need for a UPS or even power conditioning as far as I can tell.

It’s nice to be right about this stuff and nicer still if my first report is acknowledged, as the Register did and most other web sites didn’t, but what the heck. The far more interesting part of this story for me is how much Google is into server- and power-efficiency.

This efficiency measuring and calculating is based in part on maximizing Google profitability, but it goes far beyond that to an almost cult status at the search giant. Where the rest of the world just says “throw more hardware at it,” Google works to optimize every level of hardware and software to the point of even making code run better and in less memory specifically to save machines.

The power of this optimization ethos at Google can’t be over-estimated. Code is optimized not just for performance but also to use less CPU to reduce power consumption and heat generation.  Every data center that has a plan and a clue optimizes for code speed or size, but only Google (as far as I know) optimizes also for power and heat. Google server code might run a bit slower but used less average CPU (keeping the CPU idle more and reducing power consumption).

At Google EVERY optimization matters.  A one percent performance improvement could save the company thousands of servers and hundreds of thousands of dollars.  This really matters not just in company expenses, but in environmental impact, too.  Google is that rare place where writing “green code” is really encouraged.

But this is a delicate balancing act. Writing faster code saves servers. Writing code that uses less memory saves servers. The bottom line is these things really matter at Google where “throwing more hardware at it,” is not a common technique.

EVERYTHING matters. There was even an infamous Google white paper comparing the reliability and efficiency of Western Digital and Seagate hard drives. Think about it: if you are buying half a million drives even a 1-2 percent efficiency difference comes down to a lot of money.

It is surprising, then, that I’ve never heard from Google an expression of interest in the Antek Metal Foil Drive (MFD) I’ve been helping to develop for the last couple years. Why worry about comparing WD and Seagate with their 1-2 percent efficiency differential when an Antek MFD of the same capacity relies on a much thinner disk and a much smaller motor to use 85 PERCENT less energy for identical performance?

Good question.

And there are good – or at least interesting – answers, too, that say plenty more about Google culture.

Google is very Do It Yourself.  I’m surprised they haven’t built their own hard drives already. They have built other parts.  But exactly for this reason Google would be skeptical of the Antek MFD. Even if the drive performance were verified they’d still have doubts about reliability.

Then there is the emotional component of a decision to even evaluate an Antek MFD. Like every other geeky company, Google thinks of itself as ruthlessly analytical yet is actually very emotional. Nerds hold positions often as not on what they BELIEVE as any other reason. They aren’t opposed to the scientific method, but aren’t above choosing not to use it if doing so makes them feel better.

That’s the way I believe Google is about the Antek MFD. Google is big and arrogant. Even more importantly, the gatekeeper in this area is Google VP Urs Hoelzle who doesn’t like me very much, even though I might just be the nicest guy in the world.  Since I am associated with the Antek MFD and Urs doesn’t like me, Google has no interest in a clearly superior technology.

How geeky is that?