Large Nib Icon bliterati. [sign in] or Register

hank

Why I Trust Google

a discourse on the future of computing

In the near future we will all do the vast majority of our computing on the cloud. An increasing share of our processing and information storage will happen on distributed networks of computers over which we have no direct personal control. Our personal devices will be dumb terminals with just enough power to render the content served up by the smart cloud.

We will get there because the cloud will be cheap. Our PCs are redundant. We leave them sitting around doing pretty much nothing the majority of the time. Even when we do sit at our computers, we're usually only using a fraction of the resources available. The cloud lets us share resources and only use what we need. By eliminating the redundancy of localized computing, cloud providers can do away with a large proportion of the currently wasted computing potential.

If you use your computer an average of three hours a day, then that means that seven eighths of the typical day your computer is serving almost no purpose. If you could somehow share your computer with eight other people (who conveniently happen to want to use the computer at the times that you don't want to use it) you could split the cost with them and you'd all suddenly be able to pay 87% less for your computing. If you paid a thousand dollars for your current computer, wouldn't you be willing to listen to someone who presented you with a sensible scheme for buying your next one for $125?

That does actually overstate the advantages somewhat. That thousand dollars you paid may have included your mouse, keyboard, speakers, monitor, printer, etc. You would still need all of those input and output devices, plus a basic local operating system to get connected to your cloud resources. Also, assuming there is some overlap in the hours people want to use computers, there will still be periods of peak utilization and periods of low utilization. The cloud needs to provide enough capacity to guarantee service at peak usage. So even it will have wasted cycles at the times when usage is low. Then maybe your next computer—really just a package of input and output devices—costs say four or five hundred dollars (so 50% of what it otherwise would). That's still a big improvement and we haven't even considered all of the advantages yet.

The cloud is computing as a utility. When your lights flicker you don't have to run down into the basement and throw more coal into you personal power generator. When your power goes out you don't sit in the dark parsing your generator manual by flashlight and listening to elevator music on Yamaha's tech support line. You just read a book or go to bed and in a few hours the power magically kicks back in. When you buy a new appliance you don't wonder if your generator can handle it, you just plug it in, pay for a few extra megawatt hours and let the electric company figure out how much capacity they need. People who grow up using the cloud will look back on stories of people having to maintain and update computers and losing their data in crashes as funny anecdotes from a quirky dark age before real computing began.

Separating computation from local hardware will free us from a wide range of headaches that we currently accept. For instance there will be little need to make choices about system specifications. If you want to do some high power gaming you use the same old system you've had for years, you just pay an extra dollar or two to your cloud provider for the extra resources you've consumed (indeed the creator could build a custom cloud dedicated to nothing but running their game). When you browse the web you use almost no resources and you pay almost nothing, when you want to edit videos or render graphics you pay a small amount more. How much hard drive space do I need? A nonsense question when you pay by the gig.

And that hard drive space is likely to be more stable as well. It's easy for a cloud storage system to save multiple copies of your work on different machines in different data centers, quite possibly on different continents. For the most part, storage will benefit from the same economies of sharing as computation. Pretty much any non-unique data can be efficiently shared. Netflix and YouTube for instance have the potential to radically reduce the number of bits world wide that are dedicated to video storage. You may keep a video file stored on your hard drive or on a laser disc in your home, but how often do you actually watch it? A couple times a year, if you really really like it. Imagine how many people you could effectively share that file with. Easily a few thousand. That is, a few copies of that file on Netflix's servers can make the entire world supply of recordings of that film redundant. Other than home videos, pictures, and documents, pretty much everything people store on computers can be scaled back like this on the cloud. Effectively all the redundant copies of non-unique media and programs can be eliminated making room to store all of the unique personal files we really care about in triplicate, at the same cost as before.

The cloud also makes software a much stronger product. Companies can run their software on hardware of their choice, thereby solving the problem of making their application perform well on ludicrously diverse consumer hardware. This also means that there is no hardware barrier restricting your target market. And when you want to make updates to your code, you can roll them out any time you like without interrupting your user's experience. Also, if people are consuming your software on the cloud, then you never need to give anyone direct access to your code. So piracy is solved in one easy step. I started by saying that consumers will have a strong incentive to move to the cloud, but it also looks like producer have just as much to gain and will be eager to push things along from their side as well.

So in the future you won't need to replace your computer every few years. You won't need to install new hardware or software. You won't need to make permanent choices about platform specs. You won't need to deal with maintenance and crashes. And all of this is likely to come at a reduced cost. Your desktop will literally serve as nothing more than a window onto your virtual cloud-based machine which will be infinitely flexible and extensible.

Trust

All of this does involve giving up a lot of control though. What if your internet connection goes down? Your computer would be pretty much just a dead lump and all of your files would be inaccessible. What if your cloud provider went down due to maintenance issues, or due to a malicious attack? What if your provider decides to share all of your private information with greedy corporations, or government snoops? What if your provider changes the terms of your contract and holds all of your precious data hostage while raising the price of your service? What if your provider goes bankrupt, or is shutdown by the feds, or is vaporized by an alien death-ray?  Could all of your pictures, home videos, documents, and other unique files be permanently lost?

Being too trusting is a form of foolishness. But, when the recipient of your trust has a strong interest in being trustworthy, trust is probably a safe bet. A lot of people currently feel that it would be foolish to rely too heavily on any computer that isn't their own. I think that people in the future will look on that mentality the way we look on people who keep their savings under their mattresses. 

In a sense, people are right to fear putting their precious data into someone else's hands. They could easily lose it. And then we'd be screwed. But that healthy fear is precisely why I trust Google. A company in Google's position has everything to gain by being seen as trustworthy. Their entire future depends on it. If they want to make as much money as possible, they will make every effort to be as conspicuously and scrupulously trustworthy as they possibly can. Google's unofficial motto is “Don't be Evil”. Their official motto should probably be “Be Trustworthy”.

So when I say I trust Google, I'm not making a statement about the inherent goodness of human nature. I'm saying that I trust well structured companies to try to make as much money as possible. I'm also saying that I trust that the people at Google are smart enough to realize that making money and being trustworthy are, for them, inextricably connected.

In the cloud driven future I dream about, people's concern over the security of their data is a central driving force. Between here and there, there will inevitably be some shake ups and wrong turns that result in epic horror stories. But just one or two such incidents will be enough to destroy the reputation of a company. And in a market where trust is everything, a bad reputation is a death sentence. In the cloud, survival of the trustworthy will reign supreme. Cutting corners to reduce price or racing ahead to add features will both be punished if, in the process, a company sacrifices security and stability. 

The cloud actually strengthens these sorts of market forces by weakening lock-in. In the past, if you wanted to exchange one piece of software for an alternative, it could mean anything from paying a large up-front fee and taking time to physically procure and install the new software, to laboriously converting file formats, to replacing hardware or even installing a new operating system or buying a new computer. In the old model, lock-in was inevitable. Software producers had no choice but to code for a fairly specific technology stack, thus limiting their market, but also limiting their competition. Using a new cloud product on the other hand is frequently as easy as clicking a link. When you have lots of competitors and they're only a click away for your customers, you need to treat your customers very well. Of course, companies could go out of their way to create lock-in by writing prohibitive contracts and making it difficult to move data off of their system, but this is precisely the sort of untrustworthy conduct that will scare away customers and, ultimately, will probably spell a company's doom.

Of all of the companies out there, I think Google sees this future most clearly. I think they have geared their entire business model toward being the first true cloud platform. They've even introduced the Chromebook as a prototype cloud terminal. They seem to understand that building a reputation is their core business and they have the people to get them where they want to go. I, for one, am content to put my eggs in Google's basket.

The Future is Now

The transition is happening insensibly. Though people may not realize what is happening, they're using webmail, and keeping their pictures on various web platforms, and using free webapps instead of paying for the alternative in a box. They are getting comfortable with small dumb mobile devices that get most of their power from the cloud. Ultimately no one really needs to consciously accept my arguments above, because the change is happening whether they realize it or not.

It is the job of smart companies to keep easing us down the path and to plan ahead so that their customers don't suddenly realize that they've given over control of their computing experience in some jarring and painful way. I think the few companies that violate this rule will suffer so severely and publicly that their example will tend to make such transgressions very rare.

Undoubtedly there will continue to be people who build and use their own computers. They will be hobbyists and eccentrics though, like the people who operate ham radios, or people who assemble kit cars, or survivalists with their own generator and water purification system.

The age of the personal computer is nearing an end. For the most part we'll all be much better off. Those companies who lead the way in this transition and avoid stumbling, may be the best off of all. But their gain will be the good kind of gain. It will be the wealth of those far sighted few who add massive value to society by leading us into a new phase of technology and growth. See you in the clouds.


Tags:


Views: 870
Favorites: 0
Rating: 0 Sign in to vote on documents.

Comments


Need to establish norms (Rating: 3)
[Fri Apr 6 20:04:29 2012] joe wrote:
Registered users can gain the ability to moderate comments.

Hank,

You've got a lot of good stuff in here. Cloud computing is a big deal and is shaping a lot of markets now. I agree that companies in this space have every reason to be careful to maintain their users' trust. However, I am not as sanguine as you about the prospects of them appropriately caring for all of our data for a couple of reasons.

First is the simple matter of security. It has proven very difficult for many companies to maintain the security of online data. Even things that very obviously need to be protected, like credit card account details, are lost or hacked with shocking regularity. The more of your personal data that is put in the cloud, particularly the more it is consolidated in one account, the greater the risk of exposure becomes.

So security is an issue. But I think the more interesting issue is one of perspective. I think in many cases the companies in this space get too absorbed in their technology and cool products and fail to see their implications for end users. Google makes a good example. Street View is an awesome feature, but it has raised a host of privacy issues, from people photographed in compromising locations/situations, to Street View cars recording the names and locations of all of WiFi hotspots, to actually capturing live data flowing across those networks (including emails and other personal data). Or the Google Buzz privacy debacle. Or Google's recent decision to unilaterally consolidate information across all of its services to build detailed user profiles. Or the location-tracking data Google collected from Android devices and iPhones. And take a look at the concept videos circulating for Google's upcoming augmented reality glasses -- imagine what sort of data they could accumulate once they get a few million of those in circulation.

In each of these cases, I don't believe the company meant to do anything wrong. And in each case there were legitimate business or technical reasons for them to do what they did. A lot of this data can be used to support really cool features or to drive advertising revenue that supports the myriad awesome free services Google offers to the public. But it appears that engineers and product teams are often too wrapped up in the cool things they're working on to see the bigger picture regarding privacy issues. Also, I think in some cases they simply have a different perspective on privacy issues given their comfort and familiarity with the systems in question and their insider's perspective regarding the inherent goodness of the company.

In short, I don't think motivation or the company's prudently considered best interests are really the key where privacy and personal data are concerned. They may want to do the right thing, but still fail to recognize what that is. If you asked them to draw the line across which they shouldn't venture, it is quite likely that they'd draw it in a significantly different place than the rest of us would.

For my part, I would feel better about the situation if there were more clearly established legal norms regarding the handling of private personal data. Europe has done a much better job on this front with the EU's Data Protection Directive. The U.S. has taken an ad hoc approach, with occasional investigations by the FTC, FCC, and Congress when a privacy issue receives a lot of press (including several of those I listed above). But it's not entirely clear who has jurisdiction over these issues or what the standards are. If we could take some of the difficult line-drawing questions out of the hands of people involved in the products, it would probably be a relief for the cloud service providers as well as their users. I think it is inevitable that this will happen at some point, and I'll feel better about entrusting my data to the cloud when it does.

Re: Why I Trust Google (Rating: 2)
[Mon Apr 2 20:01:19 2012] veritas wrote:
Registered users can gain the ability to moderate comments.

Your pitch for cloud computing reminds me of the arguments in favor of car sharing (e.g., Zipcar). I think the results may be the same too--I agree with you that cloud computing is the next step in network evolution, though I think your prediction of the death of the personal computer are greatly exaggerated (sorry Mark Twain).

It seems to me that some amount of processing power and storage will always be needed on the user's side. And you may be downplaying the inescapable power of consumerism to create and spur demand. ("I don't care . . . I want an iPhone.")

Don't get me wrong: I love the cloud (Well, I don't LOVE the cloud--that would be wholly unnatural and inconsistent with the principles on which our great nation was founded). In fact, I have been pushing my firm to embrace cloud storage and software options (with mixed results so far). That said, you will have to pry all of my shiny Apple products out of my cold dead hands.

Zipcar (Rating: 1)
[Tue Apr 3 21:09:54 2012] hank wrote:
Registered users can gain the ability to moderate comments.

While it's true that the average proportion of idle cars to moving cars is also pretty wasteful, I don't think that the arguments in favor of car sharing are very convincing (at least not for daily use).

There are some key differences between cars and cloud computation. When a smoker uses gmail, it doesn't make your gmail account smell funny. When someone across town uses gmail, it doesn't mean that you now have to go to their house in order to use your account. If other people use gmail carelessly, it won't stain your inbox or wear out your send button.When you log out of your gmail account you don't have to take all of your data with you.

That is, apps can be sandboxed so that users have no impact on one another's experience and the location of resources is largely inconsequential.

Using a shared car will always be limiting and inconveniencing, though perhaps more cost effective. Cloud computing on the other hand is exactly the opposite. It's the equivalent of a car that is magically always there when you need it, say right off the plane in Tahiti, and can get 50 mpg when you just need to get around, but can also do zero to sixty in three seconds when you're in the mood for fun, and with infinite trunk space to boot. If you could get one of those for less than you're spending on your current car, it wouldn't be a tough decision.

Re: Zipcar (Rating: 1)
[Thu Apr 5 16:07:26 2012] veritas wrote:
Registered users can gain the ability to moderate comments.

I don't take issue with the distinctions you highlight. My point in making the comparison, however, is that consumerism may be a stronger force than your essay implies. Notwithstanding relative efficiency, consumers may be reticent to abandon their powerful electronic devices for dumb terminals because of conspicuous consumption--they want other people to see the snazzy (and expensive) device they bought. This phenomenon may explain why we cling to product-driven markets rather than service-driven markets.