I read with interest today Steve Wozniak’s skepticism towards “All Things Cloud”. He worries about security and a lack of user data control. Even though he raises important points, they seem a little strange coming from the Chief Scientist of Fusion.io – a company suppling servers to the Cloud. Where’s the vision?
Some History
US Patent n° 391.968, Tesla’s AC Motor
Wozniak’s skepticism is akin to Thomas Edison’s towards Nikola Tesla and his work on Alternating Current (AC) back in the late 19th Century. Edison dismissed Tesla’s work as “splendid, but impractical ideas” and it delayed the introduction of AC electricity distribution for several years. It wasn’t until Tesla left Edison and sold his patents to join Westinghouse did AC power thrive and revolutionize industry as we know it.
With Edison’s Direct Current (DC), power generation was limited through voltage drop to within very short distances of the load. Tesla’s AC however, delivered electricity over far greater distances using transformers to connect high voltage trunk circuits to lower voltage devices. It enabled the centralization of electricity generation and the creation of the power grid.
Edison fought for what he considered more practical local power generation (his patents), whereas Tesla was for a more economical, efficient and centralized approach (his patents). Tesla won what became known as the War of the Currents (yet still managed to die penniless).
What’s this got to do with the Cloud?
Economics drove the adoption of “AC power” and the software virtualization that underpins the Cloud is today’s “AC power”. Software virtualization enables ecomonies of scale that will drive Cloud adoption. And just as Tesla overcame every technical challenge that nature could throw at him, so will the architects and engineers developing the Cloud. Problems usually get solved (quickly) when there is money to be made and the Cloud is making a lot of people a lot of money. Just as AC made Westinghouse the company it is today.
What’s the big deal with the Cloud?
Think about it. Tens of millions of computer servers have been installed over the last fifty years, powering hundreds of millions of desktops. Each one of those servers has its own memory, CPU, disk storage and applications that go largely underused for most of the day. Customers pay for huge amounts of redundancy and inefficiency. This was part of the problem with Edison’s DC power distribution – it needed a generator at the end of every street. And while most computer servers today go unused outside business hours, they can never be used by other companies. Yet those systems continue to draw energy from the power grid.
The Cloud proposes to fix that ineffciency.
The promise is to use every last drop of redundant computer power by centralizing it, virtualizing it (ie. sharing it) and publishing it to the Internet so that customers have access wherever they are in the world. Computers become virtual machines that can be swapped in and out on a demand driven basis. So when users of one application have all logged out, that same “bare metal” is available to other company’s users logging in to probably a different application. It’s a compelling business case that will eventually drive down energy consumption, drive up operating profits and enable companies to get out of the computer science business and back to their knitting.
By virtualizing IT resources, companies no longer need to own expensive computer systems that need refreshing every 3 years, no longer need to invest in expensive software licenses, no longer need specialized IT facilities with cooling plants. Most of these responsibilities will be farmed out to Software-as-a-Service providers that will charge on a “pay-as-you-go” or “utility” basis. The role of the IT department will evolve towards greater “added value”. This is the vision and, granted it is not going to happen over night. It took fifty years for AC to be widely used. The Cloud won’t take that long, but 5-10 years from now, the IT landscape will look very different.
It will happen; it is inevitable, and just like AC power, it will take a little time. Elementary, my dear Watson.
Why didn’t we think of this earlier?
We did, but in short, without multimedia. In those days (60s, 70s, 80s), we had mainframes, 1200 baud modems and dumb terminals running centralized text based applications from service bureaus. They were gradually replaced by distributed computing, open systems and PCs with more intuitive graphical user interfaces. The network though didn’t permit centralization on the kind of scale that mass virtualization would have demanded.
So the first enabler was vastly improved Internet network infrastructure, then disk capacities, and more recently and crucially multi-core microprocessors that support virtualization on a scale that makes the Cloud economically viable.
The phenomenal demonstration of a 10 Gigabit per second file transfer between US and Chinese research bodies shows where we are heading over the next ten years. Imagine transfering 24GB in under 30 seconds halfway across the world and compare it to the 26 hours needed under the current network. Wide Area Network speeds will fast approach those of today’s disc drives as more secure fibre optics connect “that last mile” to the desktop. Local storage will no longer be really necessary; this trend has already started with video-on-demand devices such as Apple TV, Roku and Google’s Nexus Q.
Where’s it all going?
The question is, in a world where businesses are increasingly outsourcing internal services such as computing, where will it all end? Will we witness the gradual fragmentation of the enterprise with accounting, marketing, human resources and God forbid, sales “-as-a-service” providers? What does that mean for you? 🙂 Start planning now!
No one knows the ultimate impact of virtualization, but to remain competitive, it’s vital that every business, be it small medium or large, explores ways to make use of Cloud Computing… for at least some elements of their enterprise.
My vision
At the turn of the 19th century, cars looked like horseless carriages because of “designer inertia”. The same is true today of electric cars – they look like combustion cars, but as designers integrate the benefits of electric engines into their designs (cf. BMW i3), their form will radically change.
The same is true for software.
As software designers absorb the potential of the Cloud, so their software will evolve. Five years from now, I predict successful applications will look radically different; more intuitive, more “social”, more connected through mobile devices and more seamlessly integrated with systems that surround us in the “outside world”.
So I suppose today, Wozniak is right, there are still some wrinkles to iron out with this “Cloud thing” as demonstrated by the high profile outages of Amazon, Ebay and Microsoft, but like Edison, could he just be lacking a little bit of vision?
The arguments are so compelling and the rewards for all parties so great that the problems of security, availability and data control will all get resolved, and sooner rather than later. This is why, over time, I firmly believe the Cloud will proliferate and engulf every aspect of consumer and business software.
The Fourth Column 
Leave a comment