CAMBRIDGE, MASS.—The old adage in sports is that defense wins championships, but streaming giant Akamai wants to play both sides of the field. The company—one of the world’s largest content delivery network service providers— recently built a Broadcast Operations Command Center at its Cambridge, Mass., location that is designed to play both offense and defense, providing real-time monitoring of OTT content, pro-active troubleshooting, and issue detection and resolution. The overall goal, according to Akamai’s Director of Media Operations Matt Azzarto, is to make the distribution of online content “greater than or equal to the TV experience” for the end user and content providers.
Azzarto joined Akamai about six months ago after a decade at NBC Universal in its broadcast operation center and network origination center. When he joined, Akamai was already working on the development of its BOCC facility with what he described as “considerable momentum.” With Azzarto on board, the project progressed from a cubicle space to an operation center in 100 days. Along the way he hired a team of 20 staffers, some with backgrounds in traditional broadcast, others with networking skillsets. “We’re bringing together what I call your ‘next-gen operations and support team’ for media distribution,” said Azzarto, “specifically addressing the OTT challenges that broadcasters are facing today.”
Akamai's BOCC Facility
The BOCC facility is currently in beta, with all 20 staffers working with the system simultaneously. However, as Azzarto points out, they hope to expand in the future so that they can have three to five staffers in the facility on shifts 24/7 to effectively monitor and address potential issues before they become larger problems. This proactive response differs from Akamai’s previous, more defensive, set-up, which could only react to issues when customers who deal with live event streaming and linear channels complained.
In addition to the extra manpower, the new BOCC now features “BOSS,” (broadcast operating support system), a primarily Akamai-designed toolset that gauges performance and identifies potential issues and where in the workflow they may affect quality or performance. Some of these tools—and the BOCC system overall—are based on Akamai’s Network Operations Command Center (NOCC). While the BOSS is new, the BOCC also uses tools like Akamai’s Monocle and Alert Messaging System, which originally came from the NOCC. The BOCC uses these tools to redefine the alerts to be media specific, according to Azzarto. “We’re looking at different components from encoder through entry point through mid-tier through archiver and end-to-end user; all the way through the workflow at a component level we’re monitoring with custom alerts and then we’re reacting to those alerts,” he said.
“The idea is detect an issue before it becomes an outage,” Azzarto explained. “The ultimate goal is to maintain that uptime, maintain that availability. And then beyond that it’s to improve it. How do we reduce start-up time for your player? How do we reduce re-buffering on the player side? How do we reduce latency through the network? How do we increase the bitrates that you’re able to provide? Those are the things that we are focusing on with this team.”
EXPANDING THE SYSTEM
Azzarto and his team already put the system through its paces during television’s biggest annual event, as part of CBS’ support team for its Super Bowl 50 broadcast. Azzarto described it as a good experience to see what the team could provision in the tools and what they could do from a pro-active monitoring standpoint. Akamai hopes to use the BOCC for similar large events, in which case it could scale up its support team anywhere from 30-50 people to help monitor. Between 20-30 staffers were on hand to work the Super Bowl broadcast.
While the new facility is primarily focused on the OTT, live linear market right now, Azzarto and his team plan to include an on-demand component for phase two. Akamai is also hoping to expand its BOCC services with a second location most likely outside the U.S., according to Azzarto. Although no location has been determined, the two facilities should be able to share relevant information with each other.
“Suffice to say what we are trying to do is open up this ‘black box’ of the CDN for our media customers and say ‘here is what’s happening at a component level,’” said Azzarto. “For me, it’s similar to what we would do within our team in a broadcast environment: Looking at what we would call a ‘channel chain,’ or everything from the control room through to the encoder to the uplink. We would look at everything at a component level and try to alert and respond to those issues in real-time. We’re trying to take a similar kind of approach, just do it across the CDN.”