The MLB Network is baseball’s premier broadcasting channel. It features non-stop 24-hour coverage of one of America’s favorite pastimes including complete game coverage, live look-ins, and in-depth analysis. The amount of data that the sport creates throughout a single season is impressive. The responsibility of storing and organizing that data falls on the shoulders of the MLB Network’s digital asset management team. In a recent video Tab Butler, Senior Director of MLB’s Media Management and Post Production joined us to discuss how the MLB Network manages all of its data and why the organization views tape as a vital component of the process.
The MLB Data Management Process
The MLB data management challenge is more than just one of capacity, although that is no small feat. The network also needs to track, log and identify all of the information within that data. MLB analysts need to be able to find “the needle in the haystack,” like a walk-off home run,” that tells the story.
On a typical day, fifteen stadiums are hosting a game. The broadcast of that game generates eight to ten streams from cameras, and fifteen to twenty streams during major events like the all-star game, playoffs and the World Series. For each hour of baseball, the broadcast has 10 hours of content. All of this data needs to be made available for editors. The MLB network can have as many as 250 content editors working on that content.
From a primary storage perspective, the organization can store 180,000 hours’ worth of content. MLB knows however that yesterday’s content doesn’t need the accessibility that today’s does, so it writes all the previous days content to tape. MLB uses an SL8500 tape library and a software application called DIVA to manage the content. The network writes 50TBs of new content to the library per day!
Tape though isn’t a dead end repository for the MLB Network. They interact with it daily. For example, if a player hit a walk-off home run in last night’s game, the MLB on-air personality may want to show that same player hitting a walk-off home run a couple of weeks before. The MLB software can perform a partial file restore from tape and restore just that previous walk-off home run, which is edited into last night’s home run to create a stunning visual highlight. MLB typically performs 5,000 partial restores per day.
There is a misconception that tape isn’t reliable. The MLB network can prove that tape technology is very reliable. All of its data is critical, so the MLB network makes multiple tape copies each day in case of failure but despite all those tape copies and data restorations, which involves hundreds of petabytes of data and tens of thousands of tape mounts the MLB network experiences a less than .09% error rate.
Adding up all the content that the MLB Network is responsible for adds up to over 70PB worth of data and growing. The cost to create a disk-based storage system or systems to store 70PBs worth of data plus the cost to create a disk-based disaster recovery breaks almost any organization’s budget. Once IT planners factor in a disk system’s power, cooling, and data center floor space into the cost, the problem gets even worse.
The MLB architecture not only saves the organization millions compared to an all-disk solution it also works. Editors get the files and data they need at the speed they need it, and MLB has a reliable foundation that can store data for decades before needing a migration to a next-generation format. MLB is already proving the generational capabilities of tape, it is currently migrating 1.2 PB of data per day from LTO4 tape media to the Oracle T10000D. The LTO4 tape had been in service for over ten years, and the migration between the two formats have been virtually error-free, once again proving tape’s reliability.
MLB is an excellent example of an Active Archive. The organization doesn’t wait until data is ice cold to archive to tape, it does it every day with confidence that it can quickly restore the data it needs to create the content required by the network’s on-air personalities. The MLB use case exemplifies the value of including tape, actively, in the organization’s infrastructure while at the same time debunking some of the myths surrounding the technology.