'Cloud Computing' in the newest valley buzz word. Quickly retiring 'web 2.0' as the wonder note to strike as hungry entrepreneurs pitch the venture capital set for money to create the commercial enterprise of their dreams. Cloud computing is a variant on 'grid computing' and 'utility computing'; past notes that have failed to grip in one way or the other, at least not up till now. The main knock on these buzz words is that no one has figured out how to make substantial profit from them except for the pundits. So the search for the next buzz word continues, so that the money can keep flowing, so that the entrepreneurial set can keep renewing their lease to try and crack the code.
There are more companies bleeding red from these generally related concepts than not. Basically if you can't manage to charge people a subscription or a stiff fee for your cloud computing offering, you're probably losing money. If you're storing more than text-based payloads in your cloud computing infrastructure and you're not charging a fee upfront to the majority of your customers; you're most definitely losing money – even if you have ads.
Let's review: Profitable 'cloud computing' companies – Salesforce.com (subscription), Digg (text based payloads), Google (text based payload). Unprofitable cloud computing companies – Gmail (text, pictures, video), YouTube (video), Flickr (photos), Hotmail (text, pictures, video), Plaxo (contacts). You get the picture. While always-available consumer and corporate data and content services are incredibly popular for a range of scenarios, they're incredibly expensive to host and maintain. In the end very few companies will be able to do this profitably, because it will be a tripartite equation: audience size, data payload size and egress/ingress rates and payment model. Anyone who fails to balance the three in a sort of perfect triple point will fail at it.
Overall I love the concept of cloud computing for consumers. All my stuff in some, ahem, cough; cloud. Able to connect all my devices to that sweet data and have it all synced effortlessly. The mere vision is drool worthy.
But wait a minute! If you scrutinize the core cloud computing scenarios, you'll find that there is a couple of core stuff in there: anywhere availability and any device access. The cloud itself has to be a bit more endpoint aware and able to sync your stuff to your multiple devices. And when you boil it down to this, you find that there are definitely other ways to solve this problem. For instance, personal peer to peer content and application delivery is a possibility. The typical personal computer circa 2008 sits at home at the end of a reasonably quick broadband connection, doing very little except bloating your energy bill; sipping on some watts and juice, if you'll pardon the hip hop pun. Imagine being able to put down a fairly robust set of web services on that PC and the being able to access that data from anywhere in a way that scales to the expected demand. Pictures, video, contacts, email – all your stuff you already store on your computer. There ARE several challenges to this vision today: uplink speeds in broadband connections, reliability of broadband connections, reliability and uptime of home computers1, a software licensing model that allows easily customizable web services on residential computers and finally (and the hardest) easy naming and discovery of the web services on a personal computing device – it really does have to be as dead easy as typing in http://hotmail.com in your browser to get to stuff stored on a computer at home.
We just have to redefine the 'cloud'; away from that uber grid owned by some wunderkind to that underused supercomputing capacity stored in your home. Uplink speeds WILL grow, Home based PCs WILL become more reliable and energy efficient. They WILL become more powerful, powerful enough to serve all your pictures to anyone who wants them, including your grandma. We still have to figure out how to connect to web services that are going to be naturally behind home routers/firewalls using easily understood naming techniques (some of which have not been invented)2. However once we do, this redefined 'cloud' will be more efficient and cost effective than that cloud farm in the Nevada desert - for one you don't have to pay for network access and egress rates, the customer does. And it will let software vendors still sell traditional software licenses; freeing them from choosing between a rock and the web 2.0 hard way3.
Heck, I expect my phone to become an internet accessible server within 10 years. Fully available and scalable – mom will be checking out the pictures I just took at the concert with the phone, using her web browser.
1All those NAS servers and Microsoft's Windows Home Server are the first wave of servers pushing services from outside the home firewall to the outside.
2Also ISPs and network operators have to stop being total *******s and get with the network neutrality program.
3Home web services can even use 'clouds' as backup and redundancy, for only those times that the homer web service is unavailable for some reason.