A week and a half ago, Larry Dignan shared some key takeaways from a Bernstein report titled “The Long View: Netbooks, Wireless and Cloud Computing — Client Software’s Imperfect Storm.†I do not have access to the entire report but based on what Larry summed up in his blog post, it sounds like the report painted a dim view on cloud computing; hard stats were cited such as the lack of a noteworthy impact on Amazon’s bottom line from their foray into cloud computing.
Since then, a few others have briefly voiced their opinion on the report, largely disagreeing with the “move along, there’s nothing to see here†outlook associated with cloud computing. I think the report is rather short sighted and this post is a collection of my thoughts (that others may not have mentioned) based on clear examples of what’s in the near future.
I’m reminded of Steve Job’s commencement address at Stanford where he admitted to the audience that when he was young, he couldn’t see how his future would unfold; however, looking back after he had succeeded, he could see the path by “connecting the dotsâ€. In other words, one can not connect the dots without hindsight. The Bernstein report captures the current state of the landscape in cloud computing, which today is in its infancy, but provides no visibility into the opportunities made possible by cloud computing. The Berstein report thus takes on a proverbial “the glass is half empty†approach.
I’m bullish on cloud computing as I have previously written on SYS-CON, so it’s only fair if a contrarian perspective is provided for readers to decide for themselves.
Augmenting and Extending Our Humanly Abilities
We advance our civilization—or at least our own best interest—by extending the number of operations we can perform without thinking about them.
— Dr. Roy F. Baumeister, Social Psychology Area Director and Francis Eppes Eminent Scholar (Princeton University)
What mainly separates humans from all other living beings is that we’re great at creating and using tools to augment our own physical abilities. I’d like to feature two products which, without the “cloudâ€, would not work.
Shazam (white-labeled to carriers and rebranded as other names like “MusicIDâ€) is a mobile application that captures a >60 second clip of a song playing through any audio output device (e.g., your car stereo) and then proceeds to inform you of the artist and title.
Pattern recognition is not a “solved†problem in computer science. Even today, recognizing audio and video is still very difficult, as the A.I. experts in that field will tell you. Such operations are fundamentally resource-intensive. That being said, your average, low-power mobile phone processor is probably not cut out for such operations.
Where is Shazam’s solution delivered from? That’s right, the cloud.
The other product I wanted to highlight is somewhat of a hack, yet it highlights the power of the cloud. In this YouTube clip, Stephen Myers demonstrates how he is able to feed his dog a treat from wherever he may be, using his iPhone (view his full blog post here). In case the larger implication isn’t obvious because you’re still drooling over how cool this is (as I would be), Stephen has essentially figured out how to be in two different places at the same time.
Again, impossible without connectivity from the cloud.
The Technology S-Curve
Technology S-curve
(Image from AVG Aerospace)
The traditional on-premise desktop application platform has matured. Its capabilities and limitations are well understood, and non-experts can use it with ease. When was the last time you heard of a new breakthrough in desktop applications? That’s right, as of late, innovation on the desktop has been minor and incremental.
Cloud computing and SaaS may very well be the next disruptive technology. I say “may†because I can’t prove that statement until it actually happens, but that’s how it always is: incumbents have trouble identifying disruptors until it is too late. We’ve been seeing a lot of inventions made possible in this area, and many of them are early proof-of-concepts (like Stephen Myer’s iPhone-dog-feeder mentioned in this post). In result, these inventions are only attractive to early adopters. When the cloud matures and crosses the chasm, it will displace many desktop applications.
** Case in point: Salesforce.com, the poster child of SaaS has been around for a long time, but one can argue that SaaS isn’t as mature as it has the potential to be, because we’re still seeing lots of innovation from them. Salesforce recently launched its Platform-as-a-Service (PaaS), opening up the possibility of a symbiotic, co-creation relationship with otherwise would-be competitors. McKinsey identified “co-creation†as one of eight business trends to watch and more recently, Salesforce has announced partnerships with Facebook and Google Apps.
In the S-curve graph above, you’ll note that what we’re discussing here fits the disruptive technology S-curve. We may be investing more on the desktop (how much did MSFT spend on Vista?), but we’re not getting much in return (Vista’s P&L, anyone?)
We’re spending on cloud computing and not seeing too much return yet (this fits perfectly with what Bernstein analyst Jeffrey Lindsay is saying). However, the tide will turn, and when it does, the displaced incumbents will feel like they have been blindsided — as history of disk-drive manufacturers have taught us (see Clayton M. Christensen’s book titled “The Innovator’s Dilemma“).
Or, maybe if everyone agreed with this view and adjusted engineering resources accordingly, then this would just be a self-fulfilling prophecy — which if that was the case… aren’t you glad you picked the right horse? 
Amazon Web Services
I have to disagree that AWS is little more than a pet project to keep their engineers busy. If that were so, Jeff Bezos himself certainly need not bother waking up early on a Saturday morning, getting on a plane just to meet a bunch of scrappy college/startup kids whose meals consist of ramen noodles, and explaining to them how revolutionary cloud computing is.
** Now of course, these bunch of kids are from the Y Combinator/Stanford crowd, so they already know how revolutionary this was (it’s pretty obvious).
Above: Picture of me and Jeff Bezos at Stanford University – Y Combinator’s Startup School earlier this year (his posse of bodyguards are excluded from picture).
So what if AWS doesn’t make up a large percentage of AMZN’s revenue right now? AMZN already has a cash-cow book business. All difficult problems theoretically take 5-10 years to solve. If you waited until then, it would be too late to gain market share. This is the long term action plan, and most Wall Street analysts are too hung up over this quarter and the next, unable to effectively see past the 3-6 month horizon.
HP CTO Shane Robison said (on the recent Web 2.0 Summit cloud computing panel) that HP is like an arms supplier to cloud computing companies. In a way, Amazon Web Services is not only an arms supplier for technology startups but also for large corporations. Remember NYT’s one-time large scale project that was otherwise prohibitive without EC2? AWS augments a corporations IT department well for those one-off occasions where you just need that extra bandwidth/processing power, etc. They’re like guns for hire.
Even if this is a low-margin business, as more people lean on EC2 and S3, AWS as a whole would eventually solidify as a “safety net†revenue stream for AMZN, allowing them in turn to take bigger risks on other projects. All in all, I really don’t see how this is a bad thing for Amazon. In fact, it’s a great thing for AMZN.
Google Docs vs. MS Office
It’s not fair to shine a light on G-Docs failure against MS Office and use that as an example of cloud computing/SaaS’s failure to deliver. Google Docs hasn’t succeeded for the same reason other non-SaaS solutions haven’t. Take Open Office for example, a traditional on-premise MS Office direct competitor, with a 98% accuracy in opening an MS Word doc (according to Bernstein). It’s not a SaaS offering and it failed too.
What all of MS Office’s competitor’s are up against is a classic vendor lock-in strategy. In particular, if all of your peer co-workers at the office use MS Word, are you going to be the jerk who says “No, ALL of you should switch to what I’m usingâ€? Unless you’re the CEO with the power to mandate that, it’s not going to happen. On the other hand, if all your peers used Google Docs, are you going to be the only one using MS Word? Probably not.
To all the existing MS Office users on the planet today who insist on using MS Office, Microsoft thanks you for unknowingly strengthening their grip on the market. Bob Warfield shares his experience in this file-format battle — and I agree with the assessment that this is not a “shades of gray†grade. It’s black or white; you either convert the file 100% successfully or you just don’t.
Lacking of any breakthroughs in the file format war, MSFT’s powerful competitive position would eventually diminish, although that may take a very long time.
When all the old mainframe “nobody got fired for buying IBM” dinosaurs in the corporate work place who insist on the familiar MS Office file format gets end-of-life’d (and shelved), and as Gen Y‘s dominate the corporate workforce, they will bring their own tools that they grew up on. Facebook, Twitter, and yes, SaaS office productivity suites. The tipping point is a question of when, not if.
** Interesting off-topic: this form of vendor lock-in cuts both ways; Microsoft is having serious trouble getting people to let go of WinXP and adopt Vista, their latest and greatest. You know how much people think your baby is fugly when people will actually shell out $50 to downgrade from Vista to WinXP. Calm down, Mr. Ballmer! No need to throw chairs at your already stressed out underlings. Looked the other way, that $50 is the profit from Vista 
Business Trends
McKinsey Quarterly’s article on business trends discussed the benefits of:
- unbundling production from delivery
- breaking down tasks into discrete work units and then reaggregating the results
On unbundling production from delivery:
“Unbundling is attractive from the supply side because it lets asset-intensive businesses—factories, warehouses, truck fleets, office buildings, data centers, networks, and so on—raise their utilization rates and therefore their returns on invested capital. On the demand side, unbundling offers access to resources and assets that might otherwise require a large fixed investment or significant scale to achieve competitive marginal costsâ€
That sounds a lot like Amazon Web Services (supply side) and its AWS customers (demand side). This concept is not exclusive to “soft†information products only but also to physical goods such as the purchasing of fractional time on a private jet, a high-end sports car, and even mobile phones. That serves as an enabler for other amazing things, like using mobile phones to bust poverty through microfinance. Just ask Nobel Peace Prize winner Muhammad Yunus.
On breaking down task and reaggregating the results:
Last week, I was talking to LiveOps CTO Vidur Apparao about their SaaS strategy. LiveOps is a SaaS call-center platform, whose customers range from corporations to individual soccer moms working from home. The call-agent soccer mom would not otherwise be able to work from home, as the costs incurred in supporting a traditional, on-premise app tailored for a geographically dispersed work force would be highly prohibitive (think of the million desktop configurations and remote software upgrade challenges). LiveOps’ SaaS strategy enabled the creation of more jobs by making jobs fit the demanding lifestyle of a particular job-seeker market segment: the busy soccer mom.
With their SaaS platform, LiveOps is able to break down a project from their large clients into discrete tasks, farm them out to the soccer moms at home, and then reaggregate the results for final delivery. This is basically like Amazon’s Mechanical Turk but specialized for a vertical, with added value (LiveOps’ expertise in call center management).
From the (job) supply side, large corporations don’t have the time to parcel out individual customer calls to a dispersed and fragmented, ad-hoc work force. From the (job) demand side, no individual mom can be there 24/7 to meet the needs of a corporation. But by spreading out the load over large numbers, they can collectively meet that business requirement.
Padmasree Warrior (Cisco CTO), on a cloud computing panel during the recent Web 2.0 Summit, said “We don’t go to work anymore, we just work.†Cisco’s acquisition of Webex enables a virtual meeting over the cloud, resulting in reduced expenses…which in our current state of economy, is very much welcomed. Here in the US, we just witnessed a very cool telepresence proof-of-concept on CNN during the ’08 elections.
Is Your Data Working For You?
As we continue to observe more clear benefits of using knowledge technologies to produce economic benefit (see knowledge-based economy), we’ll soon realize that there are certain types of data within the confines of our pry-from-my-dead-fingers firewall that are worth more to us when it is outside of the firewall. Such data that when overlayed with data shared by others, will in turn enable us to either cut cost or increase revenues.
The best example of this is the Goldcorp Challenge. CEO Rob McEwen was on a sinking ship and made a bold decision to provide some of its trade secrets (gold-mining data) to the public, as a part of a contest (akin to the Netflix 1 Mil prize contest). It was an open call to anybody on the internet who could produce better results than the company itself with the data they had just shared. By sharing that data, they profited and turned their fate around.
This class of information behaves like money — when placed in a bank, gathers interest. As of today, we’re still trying to figure out what’s safe to let go and what isn’t. A change in mindset is required. Is your industry a zero-sum game, or can the pie be expanded for everyone?
Netbooks and Cloud Computing
With Netbooks flying off the shelves in stores, it’s not difficult to see that these simple, low-power mobile computing devices are not as beefy as the desktop and would need to tap into the cloud to extend their abilities for anything beyond what the device can handle. Dropbox is my favourite online cloud storage that is positioned to well capitalize on this opportunity (Dropbox is also a Y Combinator hatchling).
Market Opportunity
IDC (October 2008) estimates that the cloud computing infrastructure market will be worth $42Bil in 2012, up from $16Bil this year. They also predict a 6X growth in spending on cloud computing vs. traditional IT. From a presentation at Stanford by Andy Bechtolsheim (Sun Microsystems co-founder), now representing Arista:
- Decoupling users from datacenter locations by moving apps to “the cloud”
- Cloud computing is the fastest growing part of IT
- Cloud services are simpler to acquire and scale up and down
- Tremendous benefits to customers of all sizes
- Current economic environment is accelerating adoption of cloud solutions
Conclusion
The summary shows a current snapshot of the landscape purely from the lens of a quarterlyWall Street financial analyst. It does NOT paint the infinite opportunities made possible. Ray Kurzweil, an inventor/futurist and AI expert explained at a TED talk that while many things in life are unpredictable (like stock price picking), technology growth is quite the opposite. Tech growth is not linear, it is exponential! Google was yet another search startup in a crowded space when people thought the search problem was “solvedâ€.
That’s why I think the glass is half-full.
Disclosure: My portfolio includes some stock in CRM and AKAM. (Did I mention that I’m “long†on cloud computing?)
