Media Dojo

This year’s U.S. Open Tennis men’s final saw Juan Martin del Potro stun number 1 ranked Roger Federer in a four hour, five set drama. Potro’s first victory in a major tournament denied Federer a sixth consecutive U.S. Open title and provided tennis fans one of the most riveting finishes in recent memory. Aside from the buzz courtside and on television, the 2009 U.S. Open also broke new ground for live video streaming. Here are a few stats from the 2009 tournament which went from August 31 until September 13:

Live Match Streaming on USOpen.org:

• There were nearly 14 million (13,891,115) activated streams on USOpen.org.
• More than 2.5 million hours of live streaming were viewed (2,531,236 hours).
• 157 matches were streamed live.
• The interactive media console to access live streaming was launched 3.8 million times over the course of the tournament.
• The average length of stay on the media console was two hours and forty-five minutes.

While the 300 hours total video content of the 2009 U.S. Open tournament doesn’t match the 2200 hours of online video content connected with the 2008 Beijing Games, the U.S. Open tournament is among the single sport leaders in online video streaming. Behind the scenes, there was a clutch of online video companies that brought everything together. I spoke recently with Darcy Lorincz, CEO of Origin Digital about the U.S. Open as well as how they’re using Microsoft Azure to support their video encoding/transcoding efforts. For more info on Origin Digital, check out the Media Dojo Tear Sheet.

Media Dojo: First, let’s start off with the DNA of the company

Darcy Lorincz: We’ve been in the business of media processing for over a decade. The original company was started in 97 and called Live On Line. We spun out Origin Digital in October 2006 and were acquired by Accenture in May 2008. Where we concentrate is in encoding and transcoding, heavy lifting of video, audio and image assets. We started out like most video companies by racking and stacking boxes against customer workflows. Historically in the encoding and transcoding business, you put a box against a job and when the job is over you wonder where it’s going to be used next.

MD: Is that how you got into cloud computing?

DL: Well, it was a little more complicated. The whole game here is efficiency—in the hardware, in the workflow, in the delivery, in the people. At the first level of efficiency, you’re mainly talking about automating how you ingest video content and assign resources against it. So we built an automation layer that helped us get away from a lot of the bespoke operating systems for all the devices we needed to support. After you automate how you take in video content as a file or as physical media, the next level up is virtualizing the resources. This gave us the ability to load more customer jobs onto the same machines to boost utilization from the typical 10-20% level all the way to nearly 60% before we brought in more resources. Cloud computing is the third layer up in which we’re now bringing elastic resources into our data centers. When the automation layer detects that we’re running close to our internal capacity, it starts pulling compute resources directly from Azure depending on the specs of a given customer job. It might be that if a customer service level agreement (SLA) states that we’ve got to encode a job for X number of formats within a hour, we might spin up 200-300 Azure instances to crunch it in parallel. If the SLA says we’ve got 10 hours to do the same job, we’ll spin up 20-30 to run longer. But the beauty of cloud computing in general is that ability to dimension resources against a business relationship more than a technology constraint.

MD: What about the U.S. Open tournament?

DL: Two years ago, most broadcasters only cared about linear programming of video to your television. They still care about that but now want to guarantee what happens on your second and third screens because there are viable options for them to monetize their media in all of those mediums now. It’s real dollars now not funny money. So they need to figure how to scale that, how to maintain quality. Live events like the U.S. Open are a massive issue because you have these huge spikes in interest and activity. With typical professional sporting events, you might get 1-2m online and mobile viewers showing up. We had 13 million digital viewers over the course of two weeks with this tournament.

MD: You mentioned mobile video a couple of times. How big is that getting? Where do you see it going?

DL: 2009 has definitely been a huge growth year for mobile video. I don’t have exact stats but it’s been at least 10 fold growth from 2008 for sure. From a pure bit traffic POV, mobile growth doesn’t compare to online video. But it’s arguably more significant because mobile video is growing so fast from a base where we never saw that kind of consumption before. Mobile growth isn’t simply smarter devices. It’s also smarter consumers. Additionally, content owners are a lot more savvy about the mobile medium perhaps because of some lessons learned in online video beforehand. So they’re presenting you with content that’s optimized for mobile as opposed to content that’s been shrinkwrapped for mobile. It’s becoming a one click experience rather than the consumer needing to hack through various carrier decks to find something.

MD: Last question. Where is the opportunity going forward in online video in terms of new customers and new money?

DL: It’s true that we started mainly in the media & entertainment vertical. However, M&E isn’t where the majority of opportunity is today. The real opportunity is in enterprises using video as a storytelling medium for their internal and external corporate communications. Our customer base are the Fortune 500 companies. Video evolution is a big issue in that market. How does the marketing department, HR and other corporate arms use the new medium to create a more interactive experience for their constituents whether they are employees, partners in the value chain or even consumers? While enterprise customers need to be sophisticated in creating and positioning their messages, they shouldn’t need to worry about distribution. So we see our position and advantage with cloud computing as being able to put resources against that massive need. Enterprise clients can dump their video into our automation system and it just gets published and distributed where it needs to go in all the formats and bit rates that the people on the other end need to have. In that sense, we’re like a head end for enterprise video.

Comments are closed.