Archive for June, 2010
MD speaks with RightScale CEO
Last week, I went to San Francisco to be a part of GigaOm’s Structure conference. One of the best cloud events by far, Structure gave me an opportunity to sit down with Michael Crandell, the CEO of RightScale. For media players trying to wrap their head around cloud computing, the cloud management space RightScale leads is destined to become one of their first ports of call. It’s not simply a case that a media company needs a control panel for spinning up or shutting down servers. The bigger play hands down involves how to use a cloud management offer as the vehicle that carries a project from sketched on a napkin to hard-core deployment and operations without needing to change a lot of the fundamental plumbing. Bob highlighted some of the work RightScale does with media properties like ESPN and Zynga. As always, you can check the Media Dojo Tear Sheet–RightScale
Media Dojo : Please describe RightScale and the problem it solves
Michael Crandell: RightScale is a cloud management platform that lives on top of an infrastructure as a service cloud. The big problem we solve is allowing companies to get quick access to cloud infrastructure, namely fast provisioning, pay-as-you-go and dynamic workloads. It’s an entire cradle-to-grave environment for delivering cloud-based IT resources. RightScale itself is a SaaS platform that’s a web-based management system. From there, you get a portal if you will into all of your data center resources regardless of where they are, whether they are in the same public cloud, different public clouds and regions, private clouds or hybrids.
MD: How does that play out in practice?
MC: It’s everything from an operational dashboard that gives you real time information around lower level things like monitoring and alerts to a whole tracking and auditing function, which is critical in a cloud environment. In a world where resources come and go, people need to know “what happened on that server that was running last Monday and it’s gone now—and everything about it is gone.” The server may have been part of a cluster of 700 servers that launched to complete a huge batch job or a grid job. It may have been running in a transient, scalable app front end. Now it’s gone. So you need to be able to go back and look at log data. We also track cost data along with the operational stuff. We can cut it any number of ways depending on the goals of the customer. There’s a metering function within RightScale. Related to operations and cost functions, there is a whole set of tools to establish user roles and privileges. You don’t want everybody to be able to push the Big Red Switch that literally stops all servers. There needs to have a high level of admin access.
MD: That’s fine for the cost side of the ledger. What about the revenue side?
MC: There’s also the area of design and architecture. We have IP around a concept we call server templates. These are innovations on the idea of machine images that allow dynamic configuration of boot time servers so they can adopt their role effectively, whether you’re adding another app front end, load balancer, database slave etc. As that server is booting up, it can find out via the server template and RightScale that it’s part of a given load balanced array of web front ends, for example. The server knows where the other front ends are. It knows where the load balancer to register with is located. It knows where is the database that it’s talking to. All of that is set in a template. The configuration is predictable with the variable information being fed in at launch time.
MD: Let’s switch gears to talk about how you work with the media industry..
MC: We work with a variety of media clients like ESPN with their fan profile site during March Madness as well as Sony Music, which uses RightScale to power their artists fan sites all the way to e-commerce. We also work with Sling Media, which uses RightScale to do a lot of back end transcoding so that a slingbox can handle most anything you throw at it. Sling Media’s problem is that they have all this content coming in from publishing partners that needs to be prepped on-the-fly for a variety of devices ranging from a phone to a big screen TV with different resolutions and codings. That’s a big back-end grid transcoding operation. In the music space, we do a similar kind of job with Tunecore. Then there are the social apps and games, Facebook type stuff.
MD: What do you see really pushing the cloud in the media space?
MC: We see a couple of things. In the gaming space, the console games are now heavily networked, which means there’s a big cloud component anyway. Aside from enabling MMORGS, another important aspect is that games are increasingly being played on lightweight front-ends (browsers and phones). Asia is leading this push not only because of high end phones and networks but also because on average, the discretionary spend once you get out of Japan, Korea, urban China is often lower than what you find in North America and Western Europe. So you might end up spending $3-4 hard currency to go to an Internet café and purchase some Zynga bucks to play whatever game. Those are almost entirely cloud based, resource intensive and are driving a lot of adoption and innovation as the demand scales. Last time I checked, RightScale powers the top ten Facebook apps that aren’t published by Facebook itself.
MD: Last question, looking out 12 months, what are you seeing driving the bus for cloud computing and media?
MC: What’s becoming a reality are hybrid clouds. We started day one with running solely on AWS. More public clouds then came on line as well as low level cloud management software like Eucalyptus, Cloud.com, VMware,etc. It’s now possible to organize your own internal data center resources under the cloud model. From our POV, that’s about more choice for the customer. We’ve been doing a lot of work on the private cloud side. Doesn’t mean you repeal the laws of physics. You’re not going to have your database in Philly and your web tier in Austin or Seattle. But you might want to have a disaster recovery footprint on a separate power grid in another part of the country. In the media industry, I think there will be a lot more streaming simply because the competitive thrust will be about getting media to the customer however, whenever they want it. People will also own copies of their content. Bandwidth has become fast enough to make that a reality. The cloud will be more important for powering that. Never forget, the lighter weight the device you’re using, the more that power needs to be somewhere like the cloud.
On my way to Structure 2010
Of course I’m biased, but GigaOm puts on by far the best gathering of the cloud computing industry in June with Structure. All the C-level players who count will be there: AWS, Salesforce.com, Google, Microsoft, IBM, Rackspace, Cisco, AT&T, the list goes on.
I’ve got my 15min of fame on the second day at an “Ask the Analyst” session. Basically, it’s going to be the cloud equivalent of the bullring experience from high school football. For those of you who didn’t play, you get in the middle of a circle of players, coach nods at someone, and they charge at you from any direction to try to knock you on your butt. And as the good book says, it is truly better to give than to receive.
Given that it’s a hard-core tech-savvy audience in San Francisco, I expect some very hard hitters coming my way. Keeps one on their toes.
Equally important to my analyst duties, I’ll be prowling the guest list of behemoths, start-ups and VCs on behalf of my major media client. Cloud is fundamental to a HUGE transformation they’re funding with some serious dosh. I’ll be collecting contacts for a corporate version of “The Bachelor” or “The Bachelorette” depending on the particular execs.
Should be fun.
Post Structure there will be a lot more interviews on site plus some of initial thoughts from the ongoing research I’m doing for GigaOm Pro on the consumer media business in the cloud. This is bullring on steroids.
Up for Air Finally
I’ve been submerged the past few weeks for both personal and professional reasons. On the personal side, we’re selling my 89 year old mother’s house, which has 45 years worth of stuff crammed into it. For those of you who’ve walked in those shoes, you feel the pain.
Professionally, I’ve started a consulting project with the senior operations people at one of the top five global media companies. You love their music, movies and plug in their consumer electronics—get the drift? I can’t give a lot of specific detail for NDA reasons. But it’s no great secret that the physical supply chains for media (eg. from a master file to a CD/DVD sitting on the shelf at Wal-Mart) are undergoing profound change. I don’t think there’s a clean flip from all physical to all digital production and distribution of media. People still like to own *stuff* at the end of the day.
That said, the transition to a new media supply model has cloud computing written all over it. Go ahead and Google “Digital Supply Chain” to see why. The number of linkages between finished and semi-finished media goods, co-mingled professional and user-generated content, commerce and community functions suggests that the cloud just might be the only institution capable of taming that kind of complexity. There’s definitely no lack of challenge here.
I’ll document some lessons learned in a couple of months.

