Cloud Cover

The old joke in the IT community is that the highest bandwidth method of communication in existence is a 747 full of backup tapes – too bad the latency is so high.

Dan Golding looks at Amazon’s latest Web service: a hard drive in a FedEx box.

I was thoroughly gratified to be asked to moderate a panel at TiEcon last week. It was an all-star lineup, including long-time 451 friends Tim Guleri of Sierra Ventures and Biri Singh of IBM, as well as Preeti Somal, who inherited Richard Sarawal’s mantle as head of R&D at VMware, and Ravi Mhatre, of Lightspeed Venture Partners.

A couple of things struck me: first, the very good sportsmanship of my distinguished panel when I threw some very confronting questions (”So, VMware, IBM, talk to us about vendor lock-in”); and second, the very widespread acceptance and understanding of different kinds of cloud - not only IaaS and PaaS but the subtler and important distinction between commodity and enterprise-grade public clouds.

Perhaps most striking, though, was the fact that our largeish room was packed solid for the entire hour-and-a-half we were on stage. When we collected questions on index cards I had two fat handfuls of excellent, insightful queries, on security, licensing, telecommunications, smartphones, standards, market sizing; far more material than I could possibly cover in the available time. It is proof if proof were needed that the opportunity around cloud computing is the single most exciting prospect available to ambitious entrepreneurs right now. Many thanks to TiEcon and to Lightspeed, my generous hosts.

There are around 6000 hospitals and 15,000 or so school districts in the USA, according to Wikipedia. And there are fewer banks every day. I was checking the FDIC Failed Bank List this morning, as you do, and realized that we’re closing as many banks per month as we used to in a bad year.

So I am optimistic but not, you know, delusional.

LinkedIn founder Reid Hoffman has a provocative op-ed in the Washington Post: Let Our Start-Ups Bail Us Out. I don’t entirely buy his arguments - a payroll tax on H1-B workers just moves the disincentive to hire away from the visa caps and into the P&L statements. Nevertheless I do subscribe to his core thesis, which boils down to: only ingenuity is going to get us out of this mess.

Actually, let me refine that. Ingenuity with a great dash of hubris and a blithe disregard for financial reality got us into this mess. Only ingenuity with humility and a sober eye towards financial reality will get us out. Still, that’s exactly what I am seeing from entrepreneurs running virtualization and cloud firms, and from IT directors who are also struggling to make their systems more useful and efficient. Human intelligence wasn’t crushed in the dot-com crash and didn’t burst with the housing bubble. The tide goes in and out but the diligent fisherfolk are still mending their nets. The real work goes on.

ETA: We can also learn a lot from Mike Mulligan.

And this one from Jean-Lou Dupont is a real beauty. Very gratifying to see The 451 Group as the only research firm listed, of course! Much credit to William Fellows there, with Antonio Piraino and Dennis Callaghan doing more and more inspired work around the cloud.

I never saw a taxonomy I didn’t want to argue over, and good as it is, this one is no exception. I think Evident Software and vKernel should be included under billing, Bluestripe and vmSight under monitoring, Appistry, DataSynapse and Univa UD alongside their longtime rivals 3Tera and Cassatt. I wouldn’t call that section Private Cloud, either - my term for it is application fabric, to distinguish it from another kind of application virtualization aimed at the desktop. Private cloud enablement is shorthand for production VM automation in my book, so I would put Q-Layer there, alongside DynamicOps, Embotics, Fortisphere and ManageIQ (and indeed I did so in our November report Virtually Automatic.) Good to see VMware and Sun up there, but the workload/provisioning/scheduling VMware does at least is fairly basic and hypervisor-specific compared with what the production VM automation guys can do. And you can bet your life that VM security types like Altor, Catbird, Reflex and HyTrust will be reinventing themselves as cloud security players, especially now that VMware has bought their rival Blue Lane.

Otherwise I love it!

Tarry links to a fascinating study of global services, identifying the rising and falling stars. In: Rio, Toronto and Mexico City. Out: Perth, Australia. Coolest topline finding is that the choice of city outweighs the choice of country. I’m tickled to see this, as I’ve always sort of viewed the 20th century obsession with nation-states as a bit of an historical anomaly. Aaand sure enough, it’s back to the days when Venice and Bruges were world powers…

In a clear sign that cloud computing hype is beginning to stand out from background noise, Larry Ellison and Richard Stallman have come out and strongly denounced the hype around cloud computing and perhaps the idea of cloud computing itself. My colleague at Gartner helpfully (and tautologically) explains the situation:

“The term cloud computing has come to mean two very different things: a broader use that focuses on ‘cloud,’ and a more-focused use on system infrastructure and virtualization,” said David Mitchell Smith, vice president and Gartner Fellow. “Mixing the discussion of ‘cloud-enabling technologies’ with ‘cloud computing services’ creates confusion.”

Ellison nonetheless executes a brilliantly cynical maneuver: "We’ll make cloud computing announcements. I’m not going to fight this thing. But I don’t understand what we would do differently in the light of cloud computing other than change the wording of some of our ads." Stallman also rails tautologically: “It’s stupidity. It’s worse than stupidity: it’s a marketing hype campaign… somebody is saying this is inevitable – and whenever you hear somebody saying that, it’s very likely to be a set of businesses campaigning to make it true.”

Now, as an analyst, I’ve seen all manner of products and services being plied to me as cloud computing. I’ll freely admit that at least some of these instances are disingenuous. I will also agree that a fully cloud-powered world is very far off, if not altogether unlikely. But to discredit cloud computing itself as an idea is taking the backlash too far.

We have succinctly defined cloud computing as an economic organizing principle underpinned by several key enabling technologies. Hordes of successful startups are using cloud infrastructure providers to undergird their core business. The bandwidth consumed by Amazon Web Services surpassed Amazon.com’s retail bandwidth earlier this year, which makes Amazon.com CTO Werner Vogel’s quip sound not too fanciful. Moreover, Amazon.com is just one cloud infrastructure provider.

I don’t mean to come off sounding too breathless about cloud computing, but it’s undoubtedly real (i.e. more than just pure marketing hype) and is in the process of standing several long-held assumptions about the technology industry on their heads. A few instances of bandwagon-jumpery don’t mean we ought to take at face value the word of a prophet who has been wrong several times before. If anything, this seems to me to be a case of ‘protesting too much’, which only means that cloud computing is a real phenomenon.

The liveblogging action around ICE Summit is mostly over at Twitter. CohesiveFT’s Alexis Richardson is monadic and Terracotta’s Ari Zilka is ikarzali. I tweet as rachelchalmers and for some reason, Vishy’s Twitter handle is midtownninja. Thereby, I imagine, hangs a tale.

If you’re twittering, do add your handle to the comments.

I signed non-disclosure agreements before I got on the plane to Phoenix, so I can’t get into the details of what Citrix has disclosed at its annual analyst event. What I can point out is how remarkably congruent Citrix’s cloud vision is with that of (say) Hewlett-Packard or, indeed, The 451 Group.

Yet we’ve all proceeded from remarkably dissimilar basic axioms. Citrix’s current engagement with the cloud clearly shows the company’s heritage as a star of remote hosting, just as HP’s cloud vision is systems-inflected and ours is influenced by our deep engagement with grid and high performance computing. Despite our differences, we’ve all ended up in the same place, like a dolphin, a shark and an ichthyosaur

What we agree on: enterprise datacenters aspire to be clouds (as, indeed, do clouds); so much so that the distinction between the enterprise datacenter and a service-provider back-end begins to blur. The ultimate goal is cloudbursting - the ability to spill over spikes in demand to an elastic resource pool, without having to maintain the physical underpinnings of those resources year-round. It’s still years away, but its outlines are emerging from the cloud.

Hurricanes seen from orbit.