In response to my "web hosting can no longer be considered a core competency" (it's true!) post, Santosh Dawara brought up Greg Matter's bold statement that the world needs only 5 computers.
By computer, Greg (who works for Sun) means a "hyperscale, pan-global broadband computing service". The Google grid, for instance. Microsoft's Live platform. Salesforce.com - especially given its plans to build an iTunes like store for third party SaaS apps. They are sooo going to start offering hosting! Can you imagine iTunes not hosting the music they sell?? Amazon and eBay are possibilities as well. Greg also thinks we'll see the emergence of a Great Computer of China, and a commenter puts Akamai in the running. (BTW, Akamai disagrees with Mirror Image's claim that it has a lower cache-hit ratio. Stayed tuned for their story in separate post.)
Each "computer", Greg says, will consist of "millions of processing, storage and networking elements, globally distributed into critical-mass clusters" of possibly 5,000 nodes each. (That does sort of sound like Akamai, only much bigger.) Because emerging trends such as SaaS and online video will grow faster than Moore's Law. As he puts it:
"A company like Salesforce.com sees hypergrowth not in the form of intrinsic demand for CRM, but rather the sum of consolidation of CRM systems across thousands upon thousands of companies."
It takes tremendous capex (such as Microsoft's $980 million Texas data center project and Google's rumored $750 million facility in South Carolina) to accommodate this kind of spectacular snowballing growth. Greg thinks only the Big Five will be up to task. As for everyone else? "Either you will grow to become one of the big Computers, or you'll be acquired and be Borg-ed into one of them!" (Well, there's also the possibility that they'd push you out of business?) Of course, some large enterprises will want to operate their own infrastructure. Greg thinks those folks will get bunches of Blackboxes instead of building traditional data centers.
Thought provoking stuff, no?
You make a great point. *IF* a small company drives value innovation, it'll be able to comfortably coexist next to the hyperscale pan-global clusters. (Or get acquired at an astronomical valuation.) BUT - if the smaller company is in the business of (over)selling commodities like bandwidth and storage space, it won't be able to compete against more complete solutions from the Big Computers. Think about it...
* Google can leverage its search traffic to acquire hosting customers, offer free Adwords credits to drive adoption of Google Checkout, provide easy integration of Google-hosted sites on Google Base and Google Maps...
* Microsoft can give users an integrated online/desktop experience
* Salesforce can offer 3rd party developers not just hosting, but also a distribution channel
* S3/EC2 are part of Amazon's web services stack, which also include the Mechanical Turk, the Alexa Search Platform, Amazon's ecommerce service, etc. Also importantly, Amazon recently make its ecommerce logistics infrastructure available to customers. They'll warehouse, pack and ship physical goods for you.
Customers will gravitate towards these large ecosystems because the Big Computers will fulfill more needs than just hosting. In addition, their hosting will be cheaper and better than any hosting-only provider can offer. So the challenge is on for today's hosting providers to innovate - or else!
Next to his father's place of business was a store that sold only shoes, next to that was a picture frame store and then a lawn mower store.
The young student thought " why not knock down the three walls and sell everything through my father's store?"
The young man's name was Sam Walton and if we use the dollar as the common denominator for business success - and most of us do - then Sam must have been right. Because if you look at the list of the 10 richest people in the world, 4 or 5 of them are named Walton.
My point? Oh yes, actually it is an analogy, an analogy that is about to happen;
Google and/or AOL and/or Yahoo, IBM, Microsoft, HP will become the Wal Mart of the software industry, offering virtually any software program ever written in a SaaS mode, via their software storefronts.
ISVs will, ultimately, need not go anywhere else to market their programs to the world and EUs (end users), including the largest enterprises in the world, will, ultimately, need not go anywhere else to access any program. All that is needed is the "secret sauce" - the middleware - that enables any program ever written to be offered live, on-line.
This secret sauce must be capable of "wrapping", or provisioning any app ever written, on any platform, without access to the ISV's code (zero piracy) and be robust enough to empower a Google to offer all apps on-line, to one, one million or 100 million concurrent users. Only the "biggies" ( Wal Mart Super Stores) will have the humongous hosting capacity that will be required.
This secret sauce must also allow the ISVs to upgrade their apps, at any time, without taking the 100 million EUs off line.
This paradigm shift in the manner in which software is distributed and accessed will, ultimately force all ISVs to abandon their traditional business models, including shrink wrapped discs which, by their very nature, become obsolete as soon as they are burned.
Also, a individual software developer, with one lonely app, who was heretofore unable to bring her/his "mousetrap" to market, will - with the secret sauce - be able to offer their app to the world.
The secret sauce - aha - it is real, it took a truckload of vision and 6 years to develop and it is available (a colleague sent me a wrapped version a very popular software application yesterday, in the form of a attachment in a Email, which I accessed and used with one click of the mouse!) Watch for it - very soon!