Velocity, Structure and CloudCamp — A Cloud Computing Coming-Out Party

150 150 CloudPulse Strategies

cloud_trend From essentially nowhere less than a year ago, Cloud Computing has quickly risen to the top of the IT industry agenda and has everyone talking about what amounts to an exciting new way to deliver applications and services.

With this growing buzz as backdrop, three brand new Cloud-focused events launched in San Francisco last week, bringing together a host of key players from the rapidly growing Cloud Computing ecosystem.

I had the good fortune of attending all three of these events — the O’Reilly Velocity conference, CloudCamp, and GigaOM’s Structure ’08.

Velocity 2008

The week began with Velocity 2008, O’Reilly’s inaugural "Web Performance and Operations Conference." Targeted at the folks who build and operate large-scale Web/Cloud infrastructures, Velocity addressed the Cloud from the bottom up. The conference was chaired by Jesse Robbins and Steve Souders, both of whom are brilliant and did a fantastic job with the event.

The conference explored many of the ideas, best practices and tools used by leading companies to manage extremely large Web applications. There was an Operations/Infrastructure track and a perhaps deceptively-named "Performance" track. (The Performance track seemed to focus not on server-side application performance but on end-user perceived performance at the client. Focusing on the end-user is a Good Thing™, but the sessions on what’s new in the browser world, page rendering, etc. seemed off-topic to me, a server-side kind of guy.)

There were a good number of very informative Cloud-oriented sessions. Here are some highlights from my Twitter stream from Velocity 2008:

  • Best session title: "Clouds are no Substitute for Competence," by Javier Soltero of Hyperic
  • Favorite line: Jesse Robbins’ description of the Cloud as a "Redundant Array of Inexpensive Data-centers"
  • Interesting tidbit: The average adult male consumes about as much energy as a 3 year old Dell; ~120 Watts

Big shout-out to Jesse for helping me make it to Velocity. Thanks!

Structure ’08

CIMG0415Whereas Velocity was a meeting place for work-a-day uber-geeks sharing best practices on how to weave the Web, Structure was the Cloud equivalent of the World Economic Forum at Davos, bringing together the "Clouderati" for a day of thought-provoking panels and presentations on where it was all going and what was needed to get there.

Om Malik, Alistair Croll, and the rest of the GigaOM team did a great job with the agenda and list speakers; there was very little fluff in the program. Fortunately, for those of you who couldn’t make it, the entire event is available for on-demand viewing. If you’re interested in Cloud Computing or the evolution of Web infrastructure and platforms, I encourage you to check out some of the sessions.

I found the most interesting presentations to be:

  • Jonathan Yarmis’ (VP, AMR Research) talk "The Platform Revolution: A Look Into Disruptive Technologies," which really helped set the tone for the day Jonathan made some very interesting points about the intersection of Cloud Computing, mobile, social networking, and the enterprise.
  • Werner Vogels (CTO, on "Ahead in the Cloud: The Power of Infrastructure as a Service". Werner argued that running data center infrastructure is "undifferentiated heavy lifting" that consumes lots of enterprise resources/energy but provides little unique value, and therefore should be in the Cloud. I was hoping for a bit more on where Amazon was headed, but still enjoyed the presentation.
  • Greg Papadoploulos (CTO, Sun) on "The Invisible Hand of Web Architecture," which attempted to present an economic argument for Cloud Computing a la Adam Smith: economies of scale will drive innovation which will in turn drive everything to the cloud, assuming openness prevails.

The panels were interesting as well.

My favorite line from the conference was Werner Vogels opening with "Hi, I’m the system administrator for a small bookshop in Seattle." Other highlights can be found in my Twitter stream from Structure 08.

CloudCamp San Francisco

Nestled between the the high-gloss conferences that were Structure and Velocity was the "little unconference that could," CloudCamp! When I say that CloudCamp was my favorite event of the week, it’s not just because I helped organize it. What made CloudCamp so cool was that it was totally participant driven, and boy did the 300+ participants "bring the funk." CloudCamp was pure grassroots energy — the enthusiasm about the topic was palpable.

CIMG0395 Unlike the other events with their set agendas, at CloudCamp the participants both created and facilitated the sessions. After some opening remarks, attendees were invited to volunteer to present or facilitate a discussion on a topic. We had 27 volunteers step up to run a session (without any prodding!) and easily filled the 30 available slots (10 rooms, compliments Microsoft x 3 one-hour-long sessions).

The list of presentation/discussion topics ran the gamut. A few were:

  • What is Cloud Computing?
  • 1 Billion Page Views: Hacks to Scale Big on the Cloud
  • Cloud Storage; Unifying Cloud Storage APIs
  • SQL Database Clusters in the Cloud
  • Simple DB or SQL: Which Will Dominate the Cloud?
  • Connect the Clouds: Integration Between Clouds and Local Applications
  • No Cure for Cancer: Managing the Expectations of Cloud Computing Today and Tomorrow

Here’s the entire session board:


I held a session (this one ^^^^^  ) entitled "How Cloud Impacts Enterprise Computing." I was joined by about 20 other CloudCampers to explore a range of topics including the emergence of Private Clouds, standardization of Cloud Computing environments, Cloud security issues, and more. One of the most quotes of the evening came from Geoffrey Routledge in this session when he asked "After all, what is the difference between a really well run virtualized environment and a Cloud?"

CloudCamp managed to attract its share of Cloud heavy hitters, including many of the folks that attended at presented at both Velocity and Structure. After the sessions, folks hung around networking, chatting and catching up for quite some time.

What the Future Holds

The energy level was extremely high at each of these events, and I was left with the distinct impression that we as attendees were witnessing a decisive turning point for Cloud Computing as a market — a coming-out party of sorts.

I expect each of these first-time events will continue long into the future., and I’ve already learned about three new Cloud-focused events just SINCE LAST WEEK! For our part, we certainly hope to make CloudCamp an ongoing tradition. The next CloudCamp, CloudCamp London, will take place on July 16th, and additional dates/locations are in the works. If you’re interested in bringing CloudCamp to your area, please do get in touch.

Finally, I’ll be bringing a little bit of Cloud-mojo to my session at the upcoming Next-Generation Data Center conference in August. Let me know if you’re planning to attend. Also, if you’re interested in attending and would like a 20% discount let me know and I’ll send you my Priority Code.


  • Rick Parker

    Really curious where Jesse heard of this. I thought of this concept about 6 months ago while runnning Bedouin Networks, a startup to offer Virtual Clouds for small businesses a bit ahead of its time, started about 2 years ago. I just started a resource for this at to help others with how to do this. My question here is what would be considered inexpensive? My answer is 3-5 racks of hardware max. A cookie cutter config that can be copied to 2 datacenters in the local area and 2 datacenters in a disaster recovery area. Why 2? so if one datacenter raises its prices you can move the virtual servers to DC2 move out of DC1 to less expensive DC3 and then move the previous DC1 servers to DC3. If you have all your servers in one big DC you have a lack of a negotiation position.

    The other idea is 100% virtualization of all components, routers, switches, servers, storage systems, monitoring systems, backups systems, etc and no servers in any offices. 100% virtulaization of all servers, domain controllers, email, file, etc. This can be done because I have done it. Small companies can virtualize across products or services, larger across divisions, etc

    Hope to hear feedback on this,


  • Jesse Robbins

    I’m really glad you enjoyed Velocity!

    @Rick Parker: I’ve been using the phrase “Redundant Array of Inexpensive Datacenters” to describe resilient systems design for years now, since at least 2004. I think I coined it. It is the building block of resilient design for modern web services, beginning with the assumption that “Failure Happens” and moves up from there.

    Good luck, your idea sounds like it is on the right track!


Leave a Reply

Your email address will not be published.