Blog Post

SXSW Cloud Computing Panel: Clouds Still Need Work

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Cloud computing and cloud services are real, but this is only the beginning. This was the message the guys who helped build Amazon (s AMZN) Web Services, Google’s (s Goog) App Engine and Microsoft’s (s MSFT) Azure clouds conveyed in Austin, Texas, this morning at South by Southwest’s only cloud computing panel. It was packed.

Given that between one-quarter and one-third of the audience identified themselves as Amazon Web Services users, the reality of the cloud doesn’t seem to be in doubt.  Because I was moderating the panel, I didn’t get the chance to take great notes, so I am taking the easy way out and posting some of the most interesting statements as bullets. You can also check out the Twitter stream using the #sxswcc tag. While Meebo hosted the official chat rooms for these panels, Twitter has become the de facto back channel for audience commentary during the event.

  • Microsoft is clearly thinking about the enterprise with its Azure cloud platform, according to Yousef Khalidi, distinguished engineer at Microsoft.
  • Khalidi also said he believes that there will be public and private clouds for the enterprise, with apps such as corporate databases and privacy constrained applications being hosted inside the corporation. Collaboration can be in the public cloud. Werner Vogels, CTO of Amazon Web Services, says enterprise test and development are already in the Amazon cloud.
  • Kevin Gibbs, tech lead and manager of Google App Engine, took flak for AppEngine’s programming limitations, but he defended those limits, saying that Google believes those constraints are one way to build web apps that can truly scale.
  • Vogels says Amazon is an AWS customer, and when he was asked why doesn’t go down when AWS services do, he implied that because it’s such a large customer it gets priority. However, he then said it wasn’t a paying customer and implied that other, paying customers had priority.
  • All the panelists dodged a question about how “green” cloud computing is by pointing out that since a large percentage of the costs associated with running platforms is power consumption, they’re all highly aware of it.
  • When it comes to hybrid models that combine dedicated servers and cloud, both Khaladi and Vogels talked about figuring out the best way to bridge the two. Vogels mentioned customers using VPNs.

On a personal note, listening to the three panelists, gauging audience reaction and following a Twitter stream, all while trying to take notes, may have been one of the biggest challenges I’ve faced as a reporter. I strongly recommend checking out the Twitter stream and ReadWriteWeb’s liveblogging coverage for the big picture.

8 Responses to “SXSW Cloud Computing Panel: Clouds Still Need Work”

  1. khalidi is awesome – you mean from MS right? the same guy who was part of the original mach team with rashid and tevanian, then did springOS at sun (amongst other amazing things)..

  2. Stacey, I was in the room today and I certainly didn’t hear what Vogels said in the same way that you did. Vogels did not imply that Amazon was a big customer and that is why it didn’t go down. He said that you must architect your application properly to handle these situations and that Amazon runs on the same infrastructure. He certainly said the priority goes to the paying customers of AWS. During the session, you grouped AWS as the service that went down. AWS has never gone down as a whole. Portions of the S3 service have been unavailable for periods of time (20 mins – 8 hours on the high end) but I have never seen multiple services fail all at once. Architecture is key here in preparing for these situations and AWS has the right pieces in place to prevent total failure and as such, our application has never been completely unavailable. I found it puzzling that you went on the attack with Werner over the “outage” and didn’t say a word to Khalidi from Microsoft about the 22 hour Azure failure this weekend. You also misrepresented the “Green” question. I do not think that any of them dodged this question as you state here. They each simply stated that it does go hand in hand and that it is really tough to measure because there are no common metrics to compare. If you thought they were dodging, then as a reporter and moderator it is your job to dig and ask for the specific “Green” metrics that you wanted to know about. I would have loved to hear those individual statistics and what you view as important to a green data center.