Media Links talks tech with IP, 4K, 12G SDI, 12-bit sampling, HDR and more

Where would the sports broadcasting industry be without the ability today to transport content all over the world, in real time? Media Links is a company that has helped to build some of the world’s largest video and media transport networks, including major installations in the US, Europe, Japan, and Australia. For example, Media Links is providing the core media networking technology for Telstra’s next generation digital video network (DVN2) for broadcasting. IP-based technology from Media Links has been the backbone of football World Cups, winter and summer Olympic events, and many other high profile global sporting events since 2002, allowing broadcasters to transport signals to destinations around the globe. Here we chat to Mark Podesla, senior manager for product management and marketing at Media Links, about the technical side of what this company provides, and how it is helping the industry.

Media Links has provided IP video transport solutions to the broadcast TV industry for more than a decade. What do you mean when you talk about providing a “soft-step SDI-to-IP migration”?

A number of newer SDI to IP transport solutions require large scale expenditures in terms of core IP routers and switches, orchestration and management platforms, and edge feeder equipment. To tackle this problem, Media Links takes a scalable hardware and software approach, allowing users to incrementally migrate a portion of the local area network (LAN) or wide area network (WAN) network at a time, helping them to control costs, move at their own pace, and gain valuable experience along the way.

Customers have been very comfortable with SDI operations and performance characteristics for a long time; we don’t want to install systems that disrupt familiar usage patterns. Instead, we keep IP mostly under the covers. We use it to replicate the SDI-like properties and workflows that our customers already know.

How does a pure IP 4K studio solution compare to a baseband SDI 4K solution? What are the operational differences? What about cost differences?

Both solutions may meet the same short term SDI switching goals but the baseband approach is tactical and users may find themselves ‘boxed-in’ with baseband SDI 4K routers. In studio or control room applications, the per-port cost of a baseband solution may be lower, but flexibility, end to end signal control, and workflow improvements are diminished or lost.

Customers also need to deploy and manage separate IP networks to handle their increasing remote production and data flows, a costly and cumbersome proposition. SMPTE 2110 leverages IP and goes way beyond basic switching to give customers an architected framework that allows them to switch, mix, edit, record, and store anywhere in the network.

This just isn’t possible with baseband SDI 4K solutions. Two or more years down the road, IP switching will win out big-time cost wise, because it is far more flexible, scalable, and extensible, thereby protecting the customer’s investment. As network operators and managers get comfortable with IP switching technology, ongoing OPEX costs also trend lower.

What technical and business factors are driving customers towards higher capacity switches? How does IP switching affect or improve workflows?

With the rapid increase in 4K UHD, 12G SDI, 12-bit sampling, and HDR video feeds, signal bandwidth is increasing quite sharply. Prior generation 10Gb Ethernet switches just don’t handle this type of traffic very well on a per port, trunk, or switch fabric basis and customers don’t necessarily want to use heavier compression to make it all fit.

Remote production and data feeds are also on the rise, further contributing to switch load. To meet service level agreement and redundancy goals, hitless protection and zero packet loss has become mandatory for most feeds, but this effectively doubles bandwidth consumption. Newer 40G and 100G IP switches are far better suited to these higher bandwidths feeds, aggregating signals to 100G core trunks for more efficient and non-blocking workflows.

Aren’t all COTS switches essentially the same? How important is a non-blocking architecture?

Some vendors would like customers to believe that all COTS switches are all the same but our experience does not bear this out. In terms of latency, jitter, QoS guarantees, per port queue shaping, traffic prioritisation, multicast capabilities, and rapid switching, video and media traffic is far more demanding than data traffic and must be handled differently. Within a switch, the choice of embedded LSI switch controller logic used to manage shared memory pools plays a major role in the proper non-blocking treatment of media traffic. Some LSI switch controllers do not allow full bandwidth on every port to every other port. Ingress virtual local area network (VLAN) translation and ‘match and forward’ flow management are also two important switch requirements.

As new network-centric and software-defined applications emerge, there is also a need to fine-tune the switch operating system (OS), add custom APIs, or create task-specific apps that run locally on the switch. A tightly managed non-blocking switch architecture and control plane ensures that ports cannot be oversubscribed and that every provisioned circuit and service gets its requested performance criteria, regardless of switch port or fabric load. A non-blocking video switch is optimised for zero packet loss, 100% throughput, maximum forwarding rate, and memory allocation to priority queues.

Media Links has promoted what it refers to as a “common LAN/WAN architecture”. What are the benefits of deploying that?

Many vendors are developing IP switching solutions that are constructed from the LAN core outwards. While fine for contained studio or control room applications, WAN traffic from remote stadiums and venues, edge devices, aggregation switches, and regional superhub sites may not be well integrated into the overall network management or control framework, particularly when ingressing/egressing the LAN core. This potentially creates provisioning and troubleshooting headaches, especially when things don’t work as planned.

Multivendor (different core versus edge components) installations are quite susceptible to these type of problems. The use of a common LAN/WAN architecture mitigates these problems by using common software orchestration, messaging, control protocols, and management tools throughout the entire network, on a true end to end basis. Our common LAN/WAN architecture can also transport legacy time division multiplexing (TDM)-based video signals (or its derivatives) over an IP network.

What experience does Media Links have using switches as IP video routers in studio and production centre applications, and using this switching technology, what is the typical end to end latency you expect to see in a studio application?

Media Links’ experience with large scale IP video routing systems (SDI baseband routing replacement,) dates back more than eight years and these systems are running at a number of network studios and production centres. After proving the system’s value and performance at their HQ location, one prominent global broadcaster has deployed similar solutions in other major cities. In anticipation of the upcoming 2020 Summer Games, some of these systems will be upgraded to support 8K UHD video. In a centralised studio and production centre switching environment, we expect to see latency in the range of 400 microseconds.

What customer successes does Media Links have with large metropolitan switching networks?

Across a major US city, we’ve installed a large, fault-tolerant multi-hub metropolitan switching network that is used by many of the city’s broadcasters, TV and film studios, and news centres, soundstages, and postproduction facilities.  The network uses a ring topology and each hub is interconnected using multiple 40Gb and 10Gb trunks. Two separate remote NOCs share management and operations responsibility for this network.

Some of the benefits realised with this installation include reduced operating expenses, greatly increased network capacity, improved network resiliency, new hitless services, streamlined workflows, and a new centralised signal switching and routing portal. Multicast capabilities make it easy to send video monitor streams to any target site in the network. As an available option, our Media Exchange Gateway module provides VLAN to IP Subnet translation, allowing end users to securely tunnel their private Layer 3 IP networks over this shared Layer 2 switched network, if desired.

We know that SMPTE 2110 products are starting to become available. How do the new Media Links Aggregation and Core switches support this standard?

Media Links’ new Aggregation and Core switches are part of a comprehensive system architecture that includes new signal, control, and system timing planes as well as various encapsulation, coding, and protection methods that support the SMPTE 2110 standard. SMPTE 2110 builds upon a lot of standards that we’ve already implemented, such as ST 2022-2/5/6/7 so we’re quite familiar with the territory.

Our leadership role in several groups within the Video Services Forum (VSF) has also added a lot of value. Within the switches and network, switching and flow control will be a combination of OpenFlow plus our own Media Rapid Flow switching technology using per-flow forwarding. These new switches also support newer spine and leaf network topologies that deliver lower latencies and more predictable and consistent traffic patterns. Critical unit and system redundancy for mission critical feeds is also built into these aggregation and core switches. We are also adding TICO mezzanine compression to our edge encoders and decoders to handle greater numbers of high bandwidth 4K feeds on a given port.

Our first SMPTE 2110 capabilities will focus on 2110-20 (Uncompressed Video), 2110-21 (Traffic Shaping), 2110-30 (PCM Audio), and 2110-40 (Ancillary Data). For IP Video Router (Studio) installations, 2110 encapsulation and de-encapsulation are key, while WAN installations will also need 2110 Gateway services, including RTP transcoding.