Web 2.0, Microsoft, and the Costs at Scale
Nat just pointed me to a year-old article from Fortune about the strategic thinking that Ray Ozzie was bringing to Microsoft. Nat pointed in particular to Ray's thinking about infrastructure costs as Web 2.0 reaches scale, something we've written about previously on Radar:
As massive as that infrastructure is now, Ozzie says it pales in comparison to what will be needed when everyone is using the Internet for high-definition video and other data-rich goodies. He says he has a 300-page printout on his desk that shows where telecommunications and power assets are located globally, country by country.
"Just think about where there are windmills, dams, and other natural power sources around the world, and that's where you're going to see server farms," he says.
Though he won't get very specific, Ozzie says that he is amazed at the amounts Microsoft is spending, and that the cost of building the physical infrastructure for Web services will be a major barrier limiting the number of players in this business.
"The people who could build a viable services infrastructure of scale," he says, "are companies that have both the will and the capacity to invest staggering amounts of money - staggering amounts." Think billions, many billions.
This is the point so elegantly summed up by Microsoft's Debra Chrapaty with the statement that we quoted in the Radar post linked above: "In the future, being a developer on someone's platform will mean being hosted on their infrastructure." Yet despite Microsoft's insight into this trend, it's Amazon's S3 and EC2 that seem to have actually begun to capitalize on that opportunity.
The article also highlights another key strategic point, namely that we've hardly scratched the surface of advertising as a business model:
But the companywide excitement about the potential of online advertising is palpable. MSN's Blake Irving calculates that annual worldwide advertising spending amounts to about half-a-trillion dollars, vs. total software industry revenue of about $120 billion.
"Only 3.6 percent of that half-a-trillion today is being spent online," he says with relish, "even though 20 percent of all media viewership - including instant messaging, et cetera - is online now. So just assume that 3.6 percent grows to match the media opportunity. We want to be part of as much of that 20 points as we can."
Again, despite the strategic insight into the power of advertising as a model, it's not Microsoft that is seizing the palm. It's far too early to count Microsoft out, but a year after the fanfare of Ray's hiring at Microsoft and the justified celebration of his strategic insights, Microsoft continues to struggle. As science-fiction writer Frank Herbert once said to me, "Ideas are a dime a dozen. It's execution that counts."
tags:
| comments: 14
| Sphere It
submit: 
0 TrackBacks
TrackBack URL for this entry: http://orm3.managed.sonic.net/mt/mt-tb.cgi/2042
Comments: 14
[05.16.07 07:32 AM]
Utility computing and storage is changing the way web services are built. Amazon S3 now has 5+ billion objects stored in it. What's more, it had 920 million S3 requests on its peak day, and 16,607 requests in its peak second!
http://innowave.blogspot.com/2007/04/amazon-two-guys-0-next-youtube.html
[05.16.07 08:33 AM]
So I guess we really are going back to the mainframe model? Only this time it's MS instead of IBM at the helm.
Interesting read for someone trying to get into the software-as-a-service model.
[05.16.07 08:34 AM]
So I guess we really are going back to the mainframe model? Only this time it's MS instead of IBM at the helm.
Interesting read for someone trying to get into the software-as-a-service model.
[05.16.07 08:41 AM]
Perhaps Microsoft should go and use Amazon Web Services? Might be cheaper.
Nonetheless its gratifying that Microsoft are investing so much in the web.
[05.16.07 11:03 AM]
In the original Web 2.0 article "operations as a core competence" was one of the key characteristics listed. Forward looking service developers are realizing that means they can't own their own infrastructure any longer - it consumes resources and offers no advantage. Utility computing is already a reality for small Web 2.0 and SaaS companies and within a year will be viable even for large services.
[05.16.07 02:09 PM]
Any thoughts about governments kickin' in on infrastructure anywhere?
[05.16.07 03:11 PM]
and here i am, thinking that advertising
is just about ready to bust its own bubble,
since they can't turn the shrillness of it
up many more notches -- as one datapoint,
our l.a. billboards are quickly turning into
video display panels -- yet most of us have
already learned how to tune it all out...
silly me.
microsoft says it's a growth opportunity.
-bowerbird
[05.16.07 03:18 PM]
New trends would tend to say that server space as a utility is about where this Web 2.0 enabled world is heading.
It is impossible for your average startup with very little working capital and the dream of being the next big media company to cost effectively host rich-media files. In the future, choosing which data farm to locate your web service may be just as important choice as choosing your domain name is today.
Steve,
Howtosplitanatom.com
[05.16.07 06:46 PM]
The re-centralisation limitations of resources drive drive P2P options. It won't necessarily be dominant, but an alternative to the vertically integrated model.
Break outs like Amazon (selling their Service and Infrastructure), and full-blown infrastructure plays like Sun's will support more than the stove-pipes of Google and Microsoft. But eventually even they will run out.
With the long history of distributed storage and distributed batch processing and now the arrival of P2P streaming, it seems plausible that consumer broadband, storage and processing at the edge may provide a way to support SaaS outside the data centre.
[05.16.07 07:39 PM]
Hamish -- very good point about the P2P alternative. Joost, for example, is betting on it.
[05.17.07 05:52 AM]
I like the last statement. It's execution that counts. Every leader and manager can have his vision , but only ones with strong execution win at last. Not the vision itself.
[05.19.07 02:46 PM]
Tim-san,
Google does not mean the company that only offers a search engine.
The company also develop the new software service business-model, I mean.
One is the viewpoint of marketing and a distribution cost.
Usually, suitable cost is needed in order to supply software to ISV.
The second point is the new software business model.
Mostly, the purchase from a customer became the bottom line and enterprise application is developed and producted. However, in the case of Web application, marketing cost is free fundamentally. It is distributed via the Internet very simply, as the a delivery model.
If there may be a defect on software, since it is Web application, update cost does not start at all.
Furthermore, the sales of Google are consisted of advertisements.
Finally, I mean that the emergence of Google is import having made us the new idea not only from technologies but also business models.
[05.19.07 05:09 PM]
"Ideas are a dime a dozen. It's execution that counts."
very much so
http://blogs.business2.com/sloan/2007/05/why_microsofts_.html
Why Microsoft’s purchase of Aquantive is So Smart
part of the bigger picture is why all this infrastructure is needed by Microsoft. The bottom line is Microsoft is going after revenue streams from Advertising in "web 2.0" media.
read my comment on the blogpost (link above) to see how it links together.
the one chink in their armor is not being able to get their new plug-in (for web video) as ubiquitous as the Flash plug-in currently is.
but all that will take is an automated windows update as well as support of the Mono project for non-MS operating systems.
Post A Comment:
Radar RSS feed
davidm [05.16.07 06:52 AM]
"He says he has a 300-page printout on his desk that shows where telecommunications and power assets are located globally, country by country."
Ahh-ha-ha-ha.. that made me laugh. Is it a Web 2.0 300 page printout?