IP Production Forum: The Roadmap for Transitioning to IP Routing, COTS Infrastructure
Truck vendors, sports broadcasters are increasingly integrating IP routers
As the broadcast-production industry makes the transition from SDI baseband to IP, the sports-media sector is helping lead the way. And the transition starts with the router. Over the past three years, the market has seen mobile-unit providers like Game Creek Video and NEP roll out trucks built around IP routers, and several U.S. sports broadcasters — including NBC Sports Group’s Stamford, CT, facility and Fox Sports’ Los Angeles facility in the coming months — have integrated IP routers into their broadcast centers in various capacities.
With the publication of the first standards within SMPTE ST 2110 (Professional Media Over Managed IP Networks), the move toward IP-based routing is only likely to accelerate. At SVG IP Production Summit in New York this week, leading vendors took the stage to offer tips on how to navigate the transition to IP routing.
“A lot of customers [are asking], Are we really ready for IP? Are we there? When is [the transition] going to start? Is 2110 there yet?” said Fernando Solanes, director, solutions engineering and IP technologies, Evertz. “But there are a lot of large installs already. Evertz alone has about a hundred [IP-router] installs, so a big number of [M&E organizations] have already adopted. We can call them early adopters, but the reality is that it’s not a new space. IP has already [arrived].”
COTS vs. Proprietary Infrastructure: Finding a Balance for Broadcast
Much of the discussion surrounding the M&E industry’s transition to IP has been focused on broadcasters’ ability to shift from proprietary hardware to COTS (consumer off-the-shelf) infrastructure. By using COTS switches and open IP standards, broadcast-TV facilities could potentially lay an economical foundation for nearly infinite growth.
“I think the point is that you’re not going to need a forklift [upgrade],” said James Stellpflug, VP, product marketing, EVS. “If I need more routing somewhere, then I can add another switch. Maybe I need to change out my main [switch], which was originally my spine, and I move that down to become a leaf. Then I upgrade my spine. But, no matter what, I’m not throwing away 100% of my router.”
This spine/leaf architecture is the key to creating a scalable IP fabric within a broadcast facility, which allows further scalability, automation, and redundancy.
“We in the IT industry are not used to a monolithic system where everything home runs into a single piece of hardware,” said Ammar Latif, technical solutions architect, Cisco Systems. “We are used to build distributed systems with a lot of tie lines in between them; that’s how we build data centers. We are bringing that concept [to broadcast] because that brings scalability. We’ve been doing it for 20 years, and the idea is to bring this into the media space and maintain the COTS switches. That’s why you call it an IP fabric: you can start with one unit and scale it out to hundreds of units.”
However, it’s not as simple as just converting an entire broadcast ecosystem to COTS-based hardware: broadcast applications often require highly specialized needs. Many broadcasters find that, although many pieces of a broadcast plant can deploy COTS-based hardware and/or be virtualized in a data center, others must retain a more traditional broadcast-centric structure.
“There are some fundamental differences between what IT does and what broadcasters do,” noted Solanes. “You have to pick what’s best for you. The use of either COTS fabrics in certain scenarios or specialized fabrics for other functions — really, it’s picking the best tool for that particular job. We look to move things to a data-center approach for some customers; for others, it remains a traditional [broadcast] look and feel with new technology around it.”
Pick an Island and Go IP
For most broadcasters and live-production facilities, a forklift shift to IP is simply not practical. As a result, many organizations are choosing a small portion of their operation/infrastructure and transitioning it to IP as a first step before taking the full IP plunge.
“Some customers are comfortable with the existing [SDI] technologies,” said Latif. “They need to understand what IP brings them before switching. I notice them picking low-hanging fruit to create an IP island in an SDI facility. Create an IP island, maybe attack the master-control use case. I think that is the trend: picking an island and converting it to IP and getting their feet wet. Or it’s a brand-new facility, and that’s another different use case I’ve seen.”
It’s All About Good Timing: The Impact of PTP on IP Routing
One key element of SMPTE 2110 is PTP, a time-synchronization protocol that specifies how real-time clocks in the system synchronize with each other across an IP network. Already in use in the finance industry, SMPTE ST 2110 is using PTP to synchronize separate media workflows, as well as to synchronize equipment to a studio wall clock (replacing black burst).
As detailed in Cisco’s System Management Configuration Guide, a PTP system can comprise a combination of PTP (ordinary clocks, boundary clocks, and transparent clocks) and non-PTP devices (network switches, routers, and other infrastructure devices). These clocks are organized into a master/slave synchronization hierarchy with the grandmaster clock, which is the clock at the top of the hierarchy, determining the reference time for the entire system.
How all this fits into the traditional live–sports-production environment continues to be fleshed out.
“I think we need something that’s going be able to fit the needs like we have today in a model where genlock works, and it has to be able to be distributable,” said Stellpflug. “So transparent clocks are not going to work everywhere because, at that point, which grandmaster is going to be the actual timing reference when you start connecting a truck to a broadcast center? And when you bring 10 trucks together, having 10 different grandmasters in each truck is never going to work. So I think boundary clocks are the right way to go when we look at PTP.”
Cisco Systems sponsored this session.