May 6

Tim O'Reilly

Tim O'Reilly

The battle for the cloud

Andy Kessler has a great op-ed in the Wall Street Journal, The War for the Web:

Microsoft was smart to walk away (for now) from its $44 billion bid for Yahoo. It's never good to overpay. But the software giant - whose stock has flatlined for eight years - was onto the right strategy in looking to the Web for growth....

With the Microsoft/Yahoo deal breakdown, everyone assumes Google walks away with the prize. Not so fast. This contest is just starting. For Microsoft or Google or anyone else to win, they need four key elements of an end-to-end strategy:

- The Cloud. The desktop computer isn't going away. But as bandwidth speeds increase, more and more computing can be done in the network of computers sitting in data centers - aka the "cloud."...

- The Edge. The cloud is nothing without devices, browsers and users to feed it....

- Speed. - Speed. Once you build the cloud, it's all about network operations....

- Platform. ...Having a fast cloud is nothing if you keep it closed. The trick is to open it up as a platform for every new business idea to run on, charging appropriate fees as necessary....

Andy's analysis is all in those ellipses. Succinct, on-point, and refreshingly insightful about the true drivers of Web 2.0. And I can't help pointing out that the Wall Street Journal has now noticed the fundamental premise of our Velocity conference: "Once you build the cloud, it's all about network operations."

If Velocity were a movie, don't you think that quote might be on the movie poster?

Previous  |  Next

0 TrackBacks

TrackBack URL for this entry: http://blogs.oreilly.com/cgi-bin/mt/mt-t.cgi/6474

Comments: 14

  Anand Sharma [05.06.08 01:23 PM]

I can't for the life of me understand how Andy comes to this conclusion:

"...Only by opening up system internals to thousands of hungry developers can anyone truly create an operating system in the cloud. Google has made open announcements but is still quite closed. Advantage Microsoft."

How is Microsoft more open than Google in the world of Web 2.0


  Moe Glitz [05.06.08 03:56 PM]

Although current moves into this new 'Cloud' networking' has gotten a lot of businesses interested in this new online network, the overall premise behind this current 'Cloud' is rather, ahem - cloudy.

So far there is not one 'Cloud network' that has the vision or the scale to be a true market leader.
There is a huge opportunity for the first company that can develop the 'Killer App' of the 'Cloud network'.

I only wish that I had the opportunity to share my vision with a group of developers or a company, in showing just how powerful 'Cloud networking' can truly be.

  Alex Tolley [05.06.08 06:57 PM]

Andy kessler makes some good points, but I think he misses a bigger one in his attempt to suggest that a vigorous competition will still occur between Google and Microsoft.

Microsoft has made almost no headway against Google's business - mostly ad supported search. Google (and others) however is making some small headway against Microsoft's dominance in OS's and office applications. Microsoft is trapped in the classic "innovators dilemma". It cannot really migrate to the cloud as this will devalue the lock it has on the desktop. This happers them, allowing nimbler companies to ride the wave of weblications, leaving MS to create ever more bloated, "premium" OS and office apps. I don't see MS winning if this trend continues. At the same time, Apple is offering a viable alternative to Windows, nibbling at them from another side.

All the drivers of change seem to me to be operating against MS at the moment. It seems easy to me to visualize a scenario 5 to 10 years away when most high end apps run on a very basic Windows OS (XP lite?) and most lightweight and networked apps run on the web. The more the shift to socially aware apps continues, the more the apps migrate to web platforms, leaving MS somewhat stranded. It is looking more and more like MS is following in the demise of IBM as a major IT player.

  Artur Bergman [05.06.08 11:32 PM]

I am not sure I would say IBM is not a major IT player.

But even if so, it is ironic that MS is suffering from an migration from client to server. While IBM suffered from a migration from server to client.

  John A Arkansawyer [05.07.08 06:53 AM]

Yes, yes, Tim, this is all very interesting.

But tell us about Neil Young!

  Javier A. Soltero [05.07.08 09:09 AM]

"Once you build the cloud, it's all about network operations."

Ironically, a lot of people think that the cloud suddenly removes the need for any kind of operations. Just because you're consuming infrastructure or software from "the cloud" doesnt mean that things dont break. I wrote about this a while back:


I look forward to discussing this topic at Velocity!

  John Furrier [05.07.08 09:19 AM]

I agree great post by my friend Andy. Does your Velocity conference have a tie in to Andy (he ran Velocity Capital)

Just curious?

  Tim O'Reilly [05.07.08 02:09 PM]

@John Furrier --

No connection to Velocity Capital except the coincidence of the name.

  aj [05.09.08 12:10 AM]

How can you talk abt. cloud computing without mentioning amazon ec2?

  Tim O'Reilly [05.09.08 08:13 AM]

@Anand --

I wondered about that too. I think Andy may be referring to some of the Microsoft announcements regarding ASP.net generating cloud applications, but I hardly see that as more open or developer focused than the stuff Google has exposed. And your other comments about Google's advantages certainly make sense to me.

I did, however, think that Andy's analysis of the four factors to consider was really right on. Take "edges" for example. That's what Android is all about. (And Windows Mobile for that matter, but that seems like it's going nowhere fast.)

Really interesting talking with Jonathan Schwartz of Sun at Web 2.0 expo. He argues that their purchase of MySQL was another way to get "distribution" for cloud services -- that is, edges. It's important to realized that edges are not just non-PC devices. (For example, Firefox is an incredibly important "edge" point for Google.)

  steve [05.12.08 06:05 PM]

I understand the vision of networked computers in the cloud being a giant computer, as it is similar to the one Sun pushed for decades "the network is the computer".

The point that strikes me is that it's an over generalisation to the point of being a load of bollocks when being pushed as an inevitable end destination.

There is NO reason why anyone should have to put their information into someone else's systems. There is NO reason why programs and services have to live in the cloud when they can run on the end user's equipment.

If you look at Apple's model, they have network based services for very simple things like storage of data on their .Mac services, and then sell intelligent end-user hardware like Macs and iPhones.

Sure, Google have their large data centres to process through millions of web pages to MapReduce their way through terabytes of data... and a home PC isn't going to do that anytime soon.

On the other hand, very fast networks linking distributed home PCs would be able to blow the hell out of anything Google could do with their large data centres. The networks have to pipe to those centralised locations. Home PCs can explore their local connections much quicker.

I still don't like Web 2.0. I still think it's needlessly coupled to specific techniques, even though I agree with Tim's end vision of a networked world with emergent capabilities.

I believe that the networking alone is sufficient, and Google, Microsoft, Yahoo, Facebook and the rest are capitalising on a merely transient phase.

  Lee McColl Sylvester [05.13.08 10:23 AM]

I agree with Steve for the most part. In reality, cloud networks can and probably do work for some applications, but will likely not be the best choice for others. It's all about context.

My worry is that all of this is more hype than the outcome will actually provide, as has been true of many web related fads... The web in its entirety being one of them.

  Tim O'Reilly [05.13.08 11:41 AM]

Lee -- I don't know how old you are, but if you're old enough to remember the exact same comments about personal computers in the 1980s, recast your statement, and replace "cloud networks" with "personal computers."

No, the DOS-based 286 and 386s were not about to replace mainframes and minicomputers. But personal computers sure were, as they grew up, as most disruptive technologies do.

Do you really think that applications of the future will be more or less connected, more or less dependent on cloud-based features?

Take your iPod or iPhone for instance. Take away the cloud and it stops working.

I suspect you're equating "cloud" with S3 and EC2-like services. But even there, I'll lay odds that the future will see more applications based on services like this, not fewer.

Post A Comment:

 (please be patient, comments may take awhile to post)

Type the characters you see in the picture above.