The conference MC was none other than Joe Weinman, author of Cloudonomics. I scored a free autographed copy of his book and he recognized me from when we briefly met earlier this year at one of the UBM conferences in Santa Clara. What can I say, folks; I'm becoming extremely recognizable among the technorati. I missed the talk from Kevin Slavin, a TED figure from the MIT Media Lab.
The rapid session from TW Telecom made me think about virtualization for the first time since the VTUG conference I attended this past August. IMHO clients who rely heavily on virtualization are better off using a cloud solution instead of expanding their proprietary data centers. It's so obvious there's even a Dummies article to point the way.
David Kidder's keynote didn't merely recap his lessons from The Startup Playbook, and (oh yes) I scored an autographed copy of that one too. He filters a lot of entrepreneurial stories through a lens / instincts / impacts pedagogy. His formula for startup success boils down to proprietary gifts, extreme focus, painkillers over vitamins, building 10x better solutions, and monopolistic customer capture aggressively designed in from day one. David recommended the 2006 HBR study "Eager Sellers and Stony Buyers" for insights into why humans adopt new products. This synopsis of that HBR study makes me think consumer resistance to change can be plotted in a 2x2 matrix. David also said that the entrepreneurs he interviewed estimate 75% of their success is from luck! That confirms what I've witnessed and experienced. I should make a 2x2 matrix for that insight, with three whole quadrants plotted as not worth the effort. I agree with David that the psychology of playing to win increases one's chances of success in iterative moves. The key is staying in the market long enough to find an opportunity for enormous growth. In simpler terms, don't run out of money and don't ever quit.
One of David's observations about risk led me to ask my only question of this conference. He thinks extreme accountability and permission to empower progress must be as pervasive in an enterprise as permission to fail, and that's why larger companies have difficulty with innovation. The preeminence of Six Sigma means outliers aren't tolerated. I asked David if there's anything good about Six Sigma in smaller organizations. He said it is useful in the right place, specifically performing QA/QC of complex things. The problem he sees is that it forces de-risking across an enterprise in areas that have nothing to do with process control. That's what kills innovation. He wants C-suites to turn off those de-risking forces in their internal reviews of innovation and staff intrapreneurial projects with cross-functional teams. I want to believe him when he says CEOs should champion change at big companies, but my perception of most CEOs is that they are too ego-driven to want to be anything other than celebrities within their industries. Only a few who are outliers themselves will have the guts to unleash experiments unimpeded by Six Sigma.
The other afternoon panels tended towards technical specifics but I did get the impression that subsea cables are a very resilient way to deliver bandwidth. In case anyone is really into subsea cables, feel free to attend the SubOptic conference, peruse TeleGeography's Submarine Cable Map, or join the International Cable Protection Committee. One panelist mentioned that the finance sector used to pay premiums for microsecond latency advantages, but that business has dried up as brokers are becoming subject to more regulatory scrutiny. Advertising networks are the new source of hot bidding for data speed. Tolerance for latency is tied to some verticals more than others and the unit cost of data delivered matters to price-sensitive customers. Guess what, folks. The ones that aren't price sensitive, like the ad networks, will pass the price-inelastic premiums they pay along to you.
I learned a new phrase from another presenter: "hypervisor-agnostic." I did a Google search to find out what the heck that means. A hypervisor is something that runs virtual machines, so an agnostic solution must be something that presents a paravirtualization technique to its host machine regardless of the host's hardware and software configurations. That's my story and I'm sticking to it. These cloud services need to have low latency to be competitive, virtualization or not.
The final panel tried to predict the future of the cloud. The IEEE Intercloud Testbed Project is formulating standards for interoperability in multi-cloud, federated cloud, and intercloud environments. If you haven't had enough of the cloud yet, watch those clouds on the horizon. Fans of the Terminator movie series will recall that SkyNet became self-aware and nuked all of the humans it could find. Cloud providers need to make sure that they don't infuse any of these cloud virtual machines with autoimmune responses.
I'm not a cloud technician but my blogging, web hosting, and email services all reside in clouds. I've been using this tech without realizing it for as long as I've been self-employed. It's enough for me to know the outlines of the cloud sector and its major players. The finance community definitely tracks the cloud, partly thanks to the efforts of Yours Truly at Alfidi Capital.