Heading Back to Cloud Computing

Sorry, got a little distracted by the megadeal last week. I see I wrote four entries about the deal, which was probably at least one too many.

Yet I spent too many years writing, publishing, and producing for Sun to let this deal go uncommented upon. The relationship started when we shook hands and signed a deal with Ed Zander to acquire an in-house magazine on the day the Loma Prieta earthquake hit the Bay Area (I was with IDG at the time).

Zander was rude, scathing, really funny, and quite straightforward and honest in what was a negotiation lasting several months. Ed signed at about 1pm, and four hours later, we just about got rocked right out of our sales office in Palo Alto.

A lot of exciting things happened in the ensuing decade or so, including the broadcast of a satellite TV event from Brazil featuring John Gage and a squadron of carpenters pounding away, building booths at a major trade show that was to start the next day.

Gage was in Japan at an event we produced a few years later, along with McNealy, waiting patiently to deliver a keynote in a big hall with many TV cameras and no people. About 15 minutes before their speech was to start, my colleague Jim Povec went running down to registration, screaming (in English) "McNealy-san and Gage-san in room upstairs! 10 minutes!"

There were thousands of people milling around, each waiting for another to be the first to move toward a hall. They responded to this crazy American's pop-eyed performance, and "Jim's Stampede" succeeded in filling the place faster than you can say "Ohayo-gazaimas."

Getting involved at the beginning of JavaOne was another great time. Told by Sun mid-level management that the company wanted something "way outside the box," I suggested a JavaRave to reflect Java's essential coolness, or Java-Ichi (One) to reflect its truly international nature, and to hold the thing in one of those then-emerging hip warehouses South of Market in San Francisco.

In the end, JavaOne was launched at the most traditional of venues, the Moscone Center in San Francisco (mostly because it had superior electrical outlets). But it was a smash hit, drawing 20,000 the first year and scaling up to twice that before the meltdown.

Oh wait, was this post supposed to be about something other than Sun? Sorry again.

The merger does make Sun very relevant again, as its hardware will no doubt play a major role in the coming Cloud Computing deployment battle.

The future of Java is tied to cloud as well, as that of Solaris and its uneasy relationship with Linux. In fact, the entire enterprise software space is newly up for grabs. Companies must now ponder how much they will trust with SaaS and IaaS, how much glue they need, and really, how much is social networking going to play in their customer-facing futures. This pondering will take at least five years to give us a true read on Cloud and what it all means.

And then there's storage. This may be the most important area of all. Storage may be the least efficiently-used resource in enterprise IT today, its price per X-byte keeps falling, by nature it's a conservatively managed resource, yet the demands for scalability and performance will be unyielding.

Storage is the Golden Fleece for doers of misdeeds, too. To paraphrase that weathered old quote from Willie Sutton, vandals and criminals attack storage because that's where the data is. Whether you have a good grip on storage or not, here's a tip: There's a clan of storage experts who share their opinions across Twitter (start with www.twitter.com/3parfarley and go from there).

Now that I think I have Sun-Oracle out of my system completely, I'll start to break down these areas vis-a-vis Cloud, interview people who know far more about each aspect than I do, and report to you what I find.

Sun Builds Computers - Duh

I had a rather terse exchange with Scott McNealy in the mid-90s, just trying to make conversation while we waited for the CEO of my employer to show up to a meeting.

The appointment's time, location, and topic matter had been painstakingly scheduled over a period of months, and my CEO's lateness that day induced some sweaty palms on my behalf and some scowls on behalf of Scott and his handlers.

The meeting (which ended up being severely truncated) was located behind the scenes at a tradeshow called SunWorld that we had developed to complement a magazine by the same name that we published.

I decided not to try to talk golf or hockey with him, as he didn't seem in the mood. Instead, given we were at a trade show dedicated to Sun's business, I mentioned the business that Sun was in, namely, computers.

The gist of my rather simple point-of-view was that "Sun builds computers."

At that point in time, Sun was the clear market leader in its space and was threatening to drive major players such as HP and IBM right out of the arenas in which they mutually competed.

So why, I wondered, fuss with all this chip-building, software-producing, network-inducing, and other paraphernalia, which was obscuring the message?

My unsolicited opinion did nothing to lighten Scott's mood. But yet, here in 2009, it seems these three words remain the gist of the company and the gist of the Oracle-Sun deal.















Has Anything Really Changed?

How Did Ellison-McNealy Affect Oracle-Sun?

One question about the Oracle-Sun deal, not scandalous except perhaps in the eyes of IBM executives, concerns the personal relationship between Larry Ellison and Scott McNealy.

The deal, to me, shows how little power shareholders (ie, owners) of public companies really have in the face of personal relationships. I believe IBM was ready to pay a max of 9.40 per share for Sun, an offer that was vehemently castigated by Sun Chairman Scott McNealy, according to many reports.

Yet Oracle came swooping in to pay $9.50, after only four days of discussion. Of course, Sun shareholders are no doubt ecstatic about the deal, seeing their holdings quadruple from recent lows. Yet Sun's stock was hovering at about $10 before last fall's marketwide meltdown (after a 4-1 reverse split in late 2007).

What if IBM had run amok and offered an insane price like $12-13 a share for Sun? Would the market have killed IBM for this? Would Sun have been forced to accept it?

Assume that IBM had done this, with no ensuing bidding war. How does the few billion dollars premium compare to the revenue (and valuation) that IBM now risks losing against a fortified, fire-breathing Oracle? What if Oracle and HP now partner more closely? How much additional damage does this mean for IBM?

Back to my lead, back to the personal angle. McNealy and Ellison forged their apparent friendship from years of battling Microsoft. I remember self-described Libertarian McNealy testifying in front of Congress when Microsoft was in the cross-hairs of the US Justice Department, all but begging the Feds to dismantle The Borg.

The obvious business reason for the two companies to cooperate--most of Sun's server hardware is running Oracle and a large percentage of Oracle customers are on Sun--no doubt forms the basis of the relationship.

Yet, the personal angle still nags. As I noted in my previous post, John Dvorak was all over this story more than a year ago, with a column that at the time sounded paranoid but now looks to be spot-on:

http://bit.ly/15lYnG

McNealy was famous for his funny (if immature) jibes at Microsoft and IBM; I heard him once say that not only was Sun's approach superior to Redmond's but that "my kid is better-looking than his, too." Ouch.

Additionally, we're all aware of the seeming personal animus that drives Ellison in some of his deals, whether against competitor SAP or former competitor/subsumed Peoplesoft.

On a lighter note, one blogger, Marc Farley from 3Par, offered this hilarious custom-crafted video take on the process:

http://bit.ly/EkwHY

You know, satire is funny because it hits so close to the mark. Maybe Marc's video is all we need to know about this deal.

The Oracle-Sun Repercussion Discussion Begins

I just started following sfearthquakes on twitter, a useful service that provides the scoop whenever we get a temblor in norcal.

If it followed biz earthquakes in Silicon Valley, it'd be reporting on a magnitude 7.0 event, maybe higher, today. For the repercussions from the Oracle Sun deal have just begun, and there are several fault lines along which major aftershocks could no doubt occur.

The major threads in the repercussion discussion are:

* What is the future of MySql? John Dvorak was all over this angle more than a year ago, with a column that at the time sounded, frankly, paranoid, but which now looks incredibly brilliant and prescient:

http://bit.ly/15lYnG.

(As he said in a twitter post today, "not to brag, but I nailed something here.")

* Will Oracle now get real close to HP, and provide a blockbuster competitor to IBM? ZDNet's Dana Gardner has thought this and other aspects through in is recent blogpost:

http://blogs.zdnet.com/Gardner/?p=2903

* How will the deal affect Java and all of the community-centric initiatives it's spawned over the years? (Related to this topic, JavaOne in June should be a must-see event.)

* Will this deal help consign Microsoft to irrelevancy? Or does it strengthen Microsoft's hand (through Java FUD)? TIBCO CEO had some thoughts on Java and Microsoft in this interview, conducted by John Seely for Dana Gardner's column:

http://blogs.zdnet.com/Gardner/?p=2907

* How does the deal affect Cisco's nascent, unformed efforts to be a big Cloud player?

* Is the deal a good thing for Cloud (by offering a new great alternative) or a bad thing for Cloud (by squeezing Google et al through both potential vendor lock-in and as a competitor)?

Of less concern to most is the fate of thousands of Sun employees--early reports indicate as many as 10,000 will be gone--and the demolition of Sun's unique, feisty, personality-driven by sometimes dysfunctional culture.

In my opinion, the glory days of Sun's culture are far in the rear-view mirror anyway. The dot-com boom was probably the worst thing that ever happened to Sun, because it ballooned the company up to an unwieldy size, and it made everyone there think they were as smart as the few truly smart people who actually built the company.

Oracle-Sun is Great for Cloud Computing...or Not

It's an obvious point, already commented upon elsewhere, that the respective cultures of Sun Microsystems and Oracle Corp. couldn't be more different.

Sun, although brash and aggressive in the mode of its co-founder and chairman Scott McNealy, has produced numerous "characters of the game" other than McNealy over the years.

Folks such as co-founders Bill Joy and Andy Bechtolsheim, Ed Zander, James Gosling, John Gage, Radia Perlman, Whitfield Diffie, Bill Raduchal, Greg Papadopoulos, and Bernie Lacroute to name just a precious few.

The company was also a famous breeding ground for CEOs, including Zander, Eric Schmidt, Bill Larson, Carol Bartz, current Sun CEO Jonathan Schwartz, and in the glory of the dot-com days, Kim Polese.

McNealy was always clearly the boss, but having grown up around powerful people, always seemed utterly at ease in working with high-wattage personalities and letting them shine on their own.

Not so at Oracle. Larry Ellison's company has also produced CEOs such as Marc Benioff, Craig Conway, and Tom Siebel, but there has never been an indication that Ellison is comfortable sharing either power or the limelight. This history is well-documented and needs no further comment here.

Ellison has also seemed to get personal in some of his acquisitions, notably the extended, nasty PeopleSoft takeover.

So the grand meeting of idealistic Sun and hard-boiled Oracle will no doubt be difficult, with an expectation that much of Sun's idealism will ultimately be squeezed out of the combined organization.

If you go to the very southern tip of Ellison's and Siebel's home state of Illinois, you will see the pretty blue Ohio River merge with the dirty brown Mississippi. This is one of the greater river confluences in the world.

The two colors run along side by side for awhile, but all is brown soon enough. After the rivers merge, the river is known simply as the Mississippi. The bright blue Ohio simply disappears.

The technology merger aspect of this deal--presumably why Oracle is paying $7.4 billion--seems more harmonious. Many of Oracle's database customers run on Sun, and most of any hardware company's vendors run on Oracle.

The two companies have been long-time marketing partners, there are oodles of ways the two companies are touting their Java synergy, and this merger seems to create a nice overall platform that should not raise anti-trust hackles in the way an IBM/Sun might have.

Oracle will presumably not croak open-source competitor MySql (now owned by Sun), but use it to hook and reel in new generations of smaller businesses. Sun also is a significant player in the storage arena through its StorageTek acquisition of a few years back.

But, getting to the point the headline of this article makes, is this merger too harmonious? Will it simply represent more vendor lock-in?

Sun has been aggressive in promoting its Cloud Computing strategy, yes...but is vendor lock-in to a major Cloud provider fundamentally different from vendor lock-in to today's IT customers?

Capitalism does not reward that nice spirit of sharing that we all teach our kids when they're fighting for toys in the sandbox. And Ellison seems hell-bent on turning his observation--that Silicon Valley would start to look like Detroit in a few years by dint of massive consolidation--into a self-fulfilling prophecy.

Cloud Computing promises new waves of innovation by IT visionaries who will be less encumbered by the same old same old, having delivered enormous cost savings by outsourcing the drudgery to Google or Yahoo or whomever.

Will this promise be broken when, encumbered by precious little vendor choice, cloud provisioners establish pricepoints that simply return IT departments back to the budget jams of old?

Cloud Is Bigger Than the Internet

There, I said it. I wasn't the first--a couple of guys funded by Google speculated in March 2009 that cloud computing "could be" bigger than the Intertubes.

And I'm sure there are hundreds of you out there, if not more, who have voiced this opinion to colleagues, in a blog, or maybe in a corporate memo that's been ignored completely.

Someone else said that cloud will be bigger than we can imagine. This is not true and not possible. Nothing created by humans is bigger than we can imagine, because we imagined it.

The universe is bigger than we can imagine, to be sure. I'm not even sure if the universe is finite, infinite, or one of those mathematical infinities in which some infinities are much larger than others (think irrational numbers vs. rational numbers).

Even the manner in which we humans misuse new technology shouldn't be bigger than we imagine anymore, given what we've learned from The Manhattan Project.

I certainly don't want to sound glib along the lines of "The Internet changes everything." I was never sure what that was supposed to mean. The Internet didn't change the laws of physics, didn't change past historical facts, and certainly hasn't changed fundamental human nature just yet.

Plus, sounding glib in a blog would make a travesty of all the deep, considered thinking that goes on in the blogosphere.

So, back to the statement: Cloud is Bigger Than the Internet and Web.

Why?

Because as the much-smarter-than-me Nicholas Carr (among others) have pointed out, Cloud will turn IT power into a measured, utilitarian commodity. "Yes, IT does matter" was, of course, the rhetorical answer to Carr's provocative article "Does IT Matter?" a few years back. It matters in the way electricity matters.

IT does not matter in the water matters, because we can live without electricity but we can't live without water. So we'll limit the utility analogy to electricity.

Electricity, once it was standardized on a national basis, captured, and distributed more-or-less safely, has indeed transformed society for the better.

Electricity spawned innovation. For the good--safe and reliable lighting, furnaces, numerous labor-saving devices, and radio. For the bad, television. For the in-between, air-conditioning and cellphone/blackberry/iPod chargers.

Cloud will spawn similar innovation. Companies will no longer have to devote such a large percentage of their effort to keeping HAL in line.

Hand those hassles over to someone who can make money providing reliable computing power, and now you're freer, ie you have more resources, to think of all the good things you should think about.

Penicillin, Post-It Notes, and Viagra were all semi-accidental discoveriers. The researchers weren't necessarily looking for these market-busting products when they were monkeying around in their labs.

The Web was introduced in similar fashion. But additionally, legions of very smart people had been discussing hypertext and hyperlinks for years before the Web came into being. Its realization certainly kicked off a new era of dot-com innovation. But the dot-com crash illustrated that the Web itself didn't really give us a new paradigm--it just made communications a lot more convenient, while also spawning endless amounts of trivial chattering.

And its inventor will be the first to admit he stands on the shoulders of many, and he didn't expect his clever invention to crystallize with the Internet and provide the greatest IT story of the past two decades.

The Internet itself was a government operation. No more need to be said about whether it was one of history's great inventions.

Ironically, it was designed in a hugely decentralized fashion to protect against something that never happened (a large-scale nuclear attack), yet which has proven limited in its ability to handle what has happened (magnitudes upon magnitudes of increased network traffic).

Do we have a "father" (or mother) of Cloud Computing? Probably not. The topic has been discussed for awhile, and seems to represent a convergence of recent thought about software provisioning, server consolidation, frustration over long deployment times, and the usual drone of budgetary concerns.

But as Cloud is adopted, it will be historically clear that we have finally caught up to where electricity was in the late 19th century. To be sure, there will be standardization issues and disputes, failures, mass confusion (never underestimate the power of FUD), and buyer's remorse along the way.

But just as "the current wars" in Edison's time over whether to adopt AC or DC seem quaint and amusing to us today, the initial cloud-v-cloud skirmishes over the next several years will seem quaint a century from now.

It's a done deal. Cloud is the way of the future. For the first time since computing resources started to be managed by customers in the 1950s, Cloud places this responsibility back on the shoulders of utility providers.

Sure, pesky things such as application development, deployment strategy, capacity planning (which will become much more sophisticated once people realize they can micromanage this), and knowing exactly what you are trying to do with your IT will remain in the hands of the business.

And it is here where Cloud becomes most profound. Electricity doesn't enable a company to do anything. Computing resources do.

Farming out your IT functionality means you can now focus on making all the great brains in your company--and brains are not computers, never have been, never will be--spend their time inventing the future, rather than hassling with problems of the present.

Will Cloud Providers Become Too Big to Fail?

It's a beautiful Sunday afternoon in the SF Bay Area, and I can think of nothing better to do than sit on my tail and write up some recent thoughts.

Let others hit the beaches, walk in the redwoods, cruise one of the three wine countries he have here, enjoy The City, or maybe sneak up to Tahoe for the last of this year's skiing. Today, I'm all about Cloud Computing and its place in history.

I'm not a technical person, as anyone who has ever read my stuff can attest. So I won't tell you how cloud will affect EJB or you PHP/Perl/Python folks, or whether open-source will boon in a cloud-obsessed world.

Truth is, this last sentence was a straw man, as we all know that technical discussions about cloud--at least as they relate to application development--miss the point. Cloud Computing requires a very high-level, organizational dialog.

Therein lies Cloud's vast potential. Because it is through this dialog that decades of mutual misunderstanding and mistrust between IT and the business side may finally come to an end.

One does not have to know a whit about IT to know that letting another company deliver computing power as if it's electric power is a world-shaking idea that on the surface seems to exude risk.

One does not need to be sympatico to business a whit to know that this idea, with enforceable SLAs, is no more risky than the current situation of in-house IT fraught with a nightmarish complexity among servers that leads to innumerable 3am phone calls.

It occurs to me that if Cloud Computing takes hold in a big way and Amazon, Google, and Yahoo become successful in gaining high-majority market share, we may face the dreaded "too big to fail" argument should something go wrong.

Will it be a smart move to move a significant percentage of the world's IT resources to just three companies? Since the companies don't tell us much about their operations, where they are, and how they work, can we be sure that they are failsafe enough to withstand HAL episodes, the ironic electricity blackout, or dare I sound rabid, terrorist attack?

Business leaders live in an amoral world, one in which stonewalling and obfuscating is perfectly acceptable behavior. Google has already been accused of being "moral pygmies" by one influentail politician, a man (the late Tom Lantos) who was prone to bluster as much as anyone but who had a moral compass inside of him that knew where true north was.

But don't think for a minute that there is any other company on the planet that wouldn't sell its users down the river. Google's biggest crime was being clueless about what they were doing and in getting caught.

My point here is, can we trust any mega-provider to provide us the truth about the reliability of their service and how the risk their other business activities incur may affect the service they are providing to us?

If the stock market goes south on them, if they run into a cash crunch, or should they "be struck by a bolt of lightning" (apologies to Francis Ford Coppola) , will they make the argument that they are too big to fail and hold the IT operations of untold numbers of companies hostage?

"Too big to fail" is the most weasely argument I've ever heard. Guys, the Roman Empire was not too big to fail. Nor were countless other empires over the past several thousand years. The Soviet Union was not too big to fail.

And this just in, the United States is not too big to fail. There is no such thing as "American exceptionalism" unless its people and its leaders continue to be exceptional in meeting exceptional challenges. Look around today, what are the odds?

So, it seems as if one of the following scenarios will unfold in the next five to 10 years (from least likely to most likely, imo):

* Cloud proves technically infeasible on a grand scale, fizzles, joins Teletex, ISDN, and "chiclets"

* Conservative IT grudgingly deploys resources to cloud providers, but stealthily builds in 100% local redundancy, thereby negating savings and increasing overall IT costs

* Cloud wins a few high-profile accounts, more than 50% of IT runs on cloud provided by The Big Three providers, things look great all the way to the horizon

* A welter of new providers jumps into the competition, driving utility costs down, increasing the deployment of cloud dramatically, lessening the big vendor's market share to about 30%. The truth no one tells is that companies use multiple vendors for their cloud deployments, just to be safe.

* The above scenario, followed by a big consolidation among a few vendors, led by either a late-to-the-party Borg or by a twin-axis enterprise run by Larry Ellison and Rupert Murdoch. IT rebels and starts moving resources back in-house, the industry is in a huge muddle. An isolated guerilla attack in Moldova or Indonesia precipitates a massive house-of-cards destruction in global IT provisioning, once again it looks like the world is going to end, bailouts ensue, we are told to stop being reliant on oil and change our consuming ways, and a couple of young hackers go to jail for a few months. In other words, the same old same old.

So yes, Cloud Computing will take hold. It holds the promise of reducing costs, increasing corporate agility, and freeing up vast new resources for innovation (more on that in a later post).

But back to the beginning of this column, the most important thing about it is that it will force IT and business to talk--clearly, unambiguously, and maybe some day, honestly.

Cloud's day-to-day benefits will accrue slowly and incrementally, but not fundamentally change the fact that we humans are still in charge. And when humans are in charge, things never turn out as planned.

Cloud Computing Will Drive Innovation

Flip 80/20 and Go.

What does that mean?

Flip 80/20 means to stand the conventional wisdom that IT spends 80 percent of its resources on management and 20 percent on innovation. Actually, while researching an article last year on this topic, I located some academic work done in Europe that operations and maintenance taking up 90 percent of budget. A Gartner report on the topic put the ratio at 91/9.

Industry CEOs such as Mark Hurd and Michael Dell have put the maintenance percentage at 70.

So we all know the ratio tilts very unfavorably toward pesky operations and maintenance.

By sending all this stuff to the cloud--ie, a service provider's cloud--and saving money in the process, the modern-day hepcat CIO should now, in theory, be able to focus the IT department's energies on innovation.

If there is even a 10-percent savings, for example, one would think enlightened companies would plow all or most of that back into innovation. (The company has also removed a lot of capital expenditures over to operational expenditures, but I won't write much about that until I've talked to a few expert beancounters.)

Innovation is not a precise word. Rather than debate incremental vs. quantum, updates vs. breakthroughs, etc., let's look at innovation the way Bobby Kennedy looked at the world, ie "...I dream things that never were, and say 'Why not?'"

Think of what you would like your organization to be able to deliver. Break through The Innovator's Dillemma by dreaming of things your customers aren't (yet) demanding. Understand that once you have outlined your dream, you can order up the IT power you need to fulfill it. Similar to adding a new building, then ordering up the water and electricity. Yes?

This could lead to a modern-day CEO with a background in philosophy, history, literature. Yes, I know many high-level execs (including CIOs) are already broadly educated and do need to see the forest for the trees.

But, once you have outsourced the task of providing the gas to power the car, you can just step on the accelerator and go. No year-long, 18-month, or longer lag between conception and deployment.

Of course, it's not that simple. But we can always dream, can't we?

Cloud Computing - Flip 80/20 and Go

I am not an early adopter. As a semi-slow native Midwesterner, I need to think new things over, often for a long time, before deciding they're OK.

Thus I am starting my Cloud Computing conversation now, rather than a year or more ago.

In my job as Editor-in-Chief of NOW Magazine, I've been covering SOA, BPM, Ajax, BI, and Virtualization for awhile. I remember when Web Services turned into SOA. (Heck, I remember when 45s turned into LPs.)

Now all this and more is being hoisted into the sky and into the cloud. Blue Sky has become a legitimate way of thinking about IT. Wow. The physical reality of the original bug (a moth that Grace Hopper found) has now metamorphosized into the complete abstraction of all IT resources (to the user).

So what does this mean? Is this the latest paradigm shift, the latest techno-marketing babble, a revolution, a revolutionary evolution, what?

I'm up to speed on my SaaS/IaaS/PaaS trilogy, and happy to see that PaaS (Platform as a Service) has only been listed in wikipedia since January of this year. So I'm not wholly late to the cloud conversation.

I (and you) know all about new RAS concerns within the cloud, the stringency and effectiveness of what will be a new generation of SLAs, and how much interaction the local IT team really will spend hassling with server issues, now that these problem have, in theory, been delegated to the cloud provider (Amazon, Google, whomever).

But I think the key insight here involves what the Cloud-empowered CIO will do, and what type of people will be hired in the Cloud-empowered organization? I can summarize the conversation into two short sentences:

Flip 80/20 on its head. Hit the gas and go.

In other words, you will stop devoting 80 percent of your resources to maintenance and only 20 percent to innovation. And when you want to achieve something, you don't have to wait for your new architecture to take shape in your data center.

Cloud is a utility. You just put in the wires and pipes, turn on the faucets and flip the switches, and you've got water and electricity.

This is the thought that I'll develop in a new, extended article I'm writing at the moment. I can't wait to see the final draft!