Blogs

RedMonk

Skip to content

Beyond the Books: Amazon’s Next Generation Application Platform

“Right now IBM is very well set up for selling to Enterprise 1.0. But selling to Enterprise 2.0 is going to mean selling to the active end-points, or at least having a conversation with them. IBM needs a ubiquity play. It needs an storage cloud play. With Amazon on board it would have these things. IBM would be part of the internet backbone, and that’s where it has to be.” – James Governor

Fanciful as the suggestion that IBM should buy Amazon was considered at the time, it’s getting more plausible by the day.

Isabel may be entirely correct when she suggests that “Amazon’s core competency…is in observing and learning from customer behavior, and using that knowledge to put whatever products people might want right on the counter,” but there’s little question in mind that Amazon is a technology firm first, retail firm second.

It took me some time, I’ll admit, to digest how transformative and disruptive Amazon’s Hardware-as-a-Service play could be, but eventually I got it. And eventually, I’d bet, everyone else will too.

Just as Google continues to escape the notice of larger, established competitors in the collaboration space that should be taking notice, so too are businesses threatened by Amazon apparently slow to recognize the threat. Consider, for a moment, that hosting firms’ primary reaction to Amazon is to dismiss them as a competitor to Isabel by claiming that it’s “not a hosting company.” As if that’s a bad thing, as if “hosting” is a badge of honor to be worn proudly.

It’s not, if my 10 years of experience with dozens of hosts is any indication. With the rare exception of a Johncompanies or a Joyent, just about every host I’ve ever dealt with has coupled inflexible, overpriced offerings with positively abysmal customer service (1and1 once knocked our MySQL DB’s offline for four days with no explanation).

So whatever. Maybe these folks wake up and smell the train headed at them, and maybe they don’t. Either way, I won’t shed a tear.

In the meantime, Amazon refuses to sit on its laurels and hope that the basic building blocks will be enough to entice developers. The big announcement that Jeff was coyly hinting at on Twitter for days finally dropped, as did no fewer than five emails this morning asking the same question: “have you seen this?”

Indeed I have, how could I miss it? The newly hatched Flexible Payments System (FPS) is deceptively simple: it allows you to integrate with Amazon’s payment system. Big deal, you might think, I’m fine with PayPal or perhaps even Google Checkout. It’s more than that, however, trust me. Isabel points out the customer context as a differentiator, and I’d agree. It’s intriguing – like a not-as-evil Acxiom.

But more compelling to me is the big picture: Amazon has so many pieces of the puzzle in place now that it’s scary.

To illustrate this, let’s consider a concrete example – and one near and dear to my heart, my “network offering.” Without going into too much detail on the subject, the basic premise is about connecting users of Linux package management applications with individual application experts, efficiently. Ironically, my off-the-cuff explanation was “think Red Hat/MySQL network (platform bits) + AppExchange/Ubuntu Marketplace (commercial apps) + Portage / Apt (community apps) + Craigslist (job postings / resource brokering) + eBay (bids + rating system).” But with today’s announcement, how much of this quasi-services marketplace could be built on top of Amazon? Let’s see.

To make this a reality, one would need several things:

  1. Application Repository: to be selected from and installed
  2. Hardware: providing for dramatic spikes in demand
  3. Identity Management system: sufficient for providing context – history, payment info, ratings, etc – for individual transactions large and small
  4. Jobs Marketplace: allowing users to easily submit tasks and experts to discover and complete them
  5. Payment System: sufficient for everything from micropayments for very small jobs to recurring fees for retained relationships

Of those, the only one I don’t see being duplicated by Amazon is the Application Repository – and for that, there’s no need. Instead, you’d let Portage, Synaptic, Yum, and so on, worry about that part, and push the rest of it out to Amazon via an API.

Scalable hardware? EC2/S3 have you covered, relatively cost effectively. Identity management system? I’m not entirely sure, but I suspect that the identity systems behind FPS and would meet most of the requirements. What about the jobs marketplace? Hello, Mechanical Turk. And of course, FPS for payments – micro, recurring or otherwise.

Still a non-trivial endeavor, to be sure, but it really is remarkable how far along Amazon can get you – with limited barriers to entry, in either time or money. When I first discussed the above idea, my assumption was that this could only be done within the auspices of a larger entity. Reconsidering in light of Amazon’s offerings, however, I’m no longer sure of that. If I win Powerball tomorrow, as seems likely, I know what I’ll be doing in my spare time.

What can you use Amazon for? My bet is more than you think.

Disclosure: Of the mentioned firms, only IBM is a RedMonk customer. Amazon is not, while the Joyent principals host a personal website of mine gratis and Johncompanies is RedMonk’s current webhost.

Categories: Hardware-as-a-Service, Marketplaces, Software-as-a-Service.