Again, maybe I’m a slow learner, but it dawns on me that service providers — Rackspace, Savvis, Terremark that was, and all kinds of MSPs for the SME/SMB market — are going to be the trusted providers of public cloud services in the near future.

Maybe not this year (it’s fashionable to sneer at public cloud nowadays, which means nothing except that it’s passing through the Trough of Disillusionment).  But in 3 years, or 5 or 7?  You betcha.

And, just like customers today won’t buy storage systems from startups, but buy them from VARs and OEMs today, customers won’t buy public cloud from startups.  They’ll buy it from service providers and MSPs.

Which puts these guys in a similar structural position to storage OEMs: they need to buy startups in order to get R&D, and startups them them to partner with in order to show a great balance sheet and other stigmata of credibility to their customers.

My impression — pretty uneducated, but I’ll look into it shortly — is that service providers are still somewhat green at the partnership/BD/M&A game.  That should change rapidly.

Your thoughts?

Maybe I’m behind the curve here, but it strikes me that we need to start seriously planning for the death of the PC.

Those of us who remember the Larry Ellison/Scott McNealy “network computer” chatter of the ’90’s will wince a bit, but this time it’s for real.

Consider:

1. The network is now ubiquitous enough and reliable enough for an almost-always-connected model to work.  Not perfect, but good enough.
2. New clients (tablets, to be sure, but netbooks and smartphones as well) are gorgeous and compelling, more usable in some ways than desktop PCs and creeping up on laptops.
3. We are beginning to understand how to connect cloud backends with multi-client front ends.  These will get better and better.

All of this is a recipe for multiplicity of clients, and a accelerated drop in share for PC clients.

HP’s move to spin off its PC business, and Dell’s flailing about with new acqusitions and acquisition models are evidence that the top two PC makers are planning for the deatfh of the PC.  Microsoft’s bluster about the robustness of its Windows and Office franchises is typical disruptee talk.

More later on how to invest.

A September 2010 post by Michael Stonebraker was the subject of some back-and-forth in a recent CACM.  Dr. Stonebraker was saying that enterprise customers were skeptical of non-SQL or anti-SQL approaches because, among other things, they lacked ACID coherence and didn’t have a uniform functional query language like SQL.

The discussion was reminisicint of many in the “disuptor/disruptee” wars of technology: noSQL approaches (lumping them together for the moment) are inadequate for the cutting-edge needs (OLTP) of the cutting-edge users (big enterprise customers).  Of course, there are a number of niche applications — processing Web logs, for example, or call data records, or maybe BI applications in general — where the noSQL-based solutions are “good enough”, or, maybe, even better.

The trick for the noSQL vendors — and their investors — is to find out which niches those are and suffice them, quick.

Seven years ago, AOL acquired Advertising.com for some $500M, a nice payday for brothers Scott and John Ferber, who had co-founded the company, named TeknoSurf.com, in 1998.  Valhalla founders backed the company as investing professionals at predecessor firms.

Estimates are that the company, before and after the AOL acquisition, has grossed some $3B in online products and services.  Understandably, the Ferbers (and other investors) felt that we left money on the table by what now seems a “premature” acquisition.

A made-to-order Serial Entrepreneur scenario.  Scott Ferber has taken another turn at bat with TidalTV, a video ad network founded in early 2008, and this summer John Ferber has stepped up as co-founder of Domain Holdings, a “domain-lifecycle management” company.

In both cases Valhalla has invested in its winners.  Why not?  These are brilliant guys with a great business sense and a technology-enabled advantage that should help them win — and win perhaps even bigger than before — in their new ventures.

In an earlier post on “software factories”, I touched on the question of why America’s software engineers were not, by and large, working on projects that would enhance American software competitiveness:

…the finest software minds of the current generation are not interested in solving the American productivity problem, but are interested in profiting from what I elsewhere call flash-fads, huge blockbuster moneymakers that last for the comparative blink of an eye but, like the Pet Rocks of my youth, make lots of money.

This is probably rational behavior on the part of these software engineers.  Sacrificing current income to make the income of the nation greater over time is a bit like voluntarily helping to pay down the national debt by giving extra money to the Treasury: patriotic, maybe, but certainly not a mass choice.  (One of my partners told me this morning that some $81M had been contributed to the Treasury in this fashion, versus a national debt service obligation several orders of magnitude greater.)

But how does the rational behavior of individual software engineers feed the public good?  Our market orientation in the U.S. gives us a touching faith in what we might call “the healing power of greed”, an exaggeration of Adam Smith’s point of about the “invisible hand” into the notion that individuals can do whatever we darn well please and somehow benefit the polity.  In this raw form, it probably ain’t so.

Yes, over time the drive for hundreds of flash fads has in fact made software development more productive.  It is probably an order of magnitude cheaper and quicker to bring an application to market today than 30 years ago, when I got started in the tech business.  But these benefits accrue to everyone, and don’t provide specific advantage for our country.  Which is what we need to remain competitive.

It seems we can’t escape from a policy that targets innovations which are game-changing and then invests in them directly from the public purse or incents the private sector to factor innovation projects in among the Pet Rocks.

Your thoughts?

I’m sure many people are on top of this, but isn’t there a big difference between different kinds of links in the social network based on their quality?

I’m thinking about this because I’m gearing up on Google+, which features the idea of circles, discrete collections of contacts with different characteristics: friends, colleagues, running buddies, frenemies.  And a lot of the discussion in the service about circles revolves around what to do with low-quality links.

In finance, margin is a measure of the quality of a business’s income.  If it costs a lot of money to earn your revenue, it’s not as high-quality as higher-margin revenue.  High margins inhere to businesses that create differentiation.  Low margins are the h**l into which weak, aging, or commoditized businesses descend.

So with links.  We all have 80-20 rules with our link forests, where some few are priceless to us and most are almost worthless.  Should be a weighting in measuring social network influence or the like.  Probably someone is doing it already.

Thoughts?  Want an invite to Google+?

I’ve been pleasantly absorbed getting an iPad up and running the last few weeks, and I’ve learned something that is applicable to any client device: apps gain most of their power from interacting with other apps.

Case in point: I use Tripit, and got another travel-oriented app called FlightTrack (in fact I paid for FlightTrack Pro, as much because I believe in paying for software as because I like the Pro features (although I do)).

Well you can type flight numbers into FlightTrack, and it’ll give you vital info about them, like mainly whether or not they’ll be on time. But if you integrate FlightTrack with TripIt you get automatic lists of all the flights in your TripIt trips, without lifting a finger. Power of integration.

So now, on the mobile, I get a flight itinerary from the travel agent, forward it to TripIt, who parses it and turns it into flight, calendar, and other relevant info, then feeds the flight info to FlightTrack.

This is the future of apps, this is the future of app-driven programming from the cloud.  Workflows of app(let)s dynamically assembled to accomplish a one-time or recurring task.

(Now in this case the apps were statically assembled, a big hassle with manually typed-in passwords and all kinds of other (shudder) legacy stuff.  But you get the idea.)

Incidentally, this is one place where Android shines.  The operating system embeds connectivity through a “publish and subscribe” facility exposed to users via the “Share” menu.  iOS jumps through hoops to accomplish the same thing, generally on a one-off app-by-app scheme that is, frankly, flakey.  (I’m not especially an Android zealot, but this is one case where it’s clearly ahead to my mind.)

How do you chain together or connect apps to accomplish tasks?