TODD ROTH TEF 2012 Q&A
Todd Roth is Vice President of Technology for Harris Corporation, an international communications and information technology company serving government and commercial markets in more than 150 countries. As part of Harris’ Broadcast Communications group, he is responsible for incorporating and leveraging critical enabling technologies, including Ethernet.
Q: What is, in your opinion, the biggest connectivity challenge the broadcast communications industry faces today?
TR: There are a number of challenges that we face on a day-to-day basis, such as signal timing and deterministic performance of media, but it comes down to clearly communicating the advantages and benefits of a given technology choice to our customers.
Signal timing and deterministic performance of media are some of the most difficult things to address when moving from a non-buffered discrete coaxial cable to an IP media stream, particularly when you take into account QoS provisions. In transitioning to an IP-based environment, there needs to be a certain amount of provisioning at varying levels. Right now, we’re achieving this by bridging between the coax and IP environments, but the true challenge will be to eliminate the coax environment entirely, rather than having to rely on this bridging process.
This is not a new idea – conceptually, the idea of migrating broadcast signals from discrete coaxial to packetized IP has been around since Gigabit Ethernet was introduced. Adoption however has been slow, mostly due to inertia in the industry in terms of workflows, how people do things, and legacy issues, preserving existing infrastructure. It takes awhile to effect change. There’s a lot of education that needs to take place in order to alert customers that IP is a good thing. It’s a little bit of a catching-up thing for an industry that has been held back as much by its own legacy as technological issues.
Q: If educating others is critical to your business, then what is the one thing you want the people behind Ethernet to know or to hear about?
TR: It has to be compatibility and ease-of-use. There’s something intrinsically easy about having devices with dedicated “inputs” and “outputs”, installing a new piece of equipment is as simple and quick as connecting an output to an input to have it work correctly. It’s sort of like the plug-and-play model; when you’re setting up a TV at home, you just have to plug the cables in. But if you’ve ever configured anything with an Ethernet cord, it’s not the same experience. You need some form of console, or user interface, you have to know about IP addressing, you have to get your device to work with the router…there’s a certain amount of complexity that you have to deal with, and when things don’t work, it’s very difficult to troubleshoot.
Take for example, my Xbox. Microsoft wants the Xbox to be a media hub for the home, so one of the things it does is to allow you to access a home computer and play back media from that computer. It took me quite some time to get my Xbox to work because of a firewall issue. I wasn’t sure where the failure was occurring – I didn’t know if it was my router, my PC, Windows networking or the console itself. It was difficult to troubleshoot the situation, and I’m a highly skilled guy. Even the basic troubleshooting skill of isolating an issue to hardware, software or configuration is compounded when networking issues are added.
If a similar issue was to occur in a broadcast facility, where business and revenue are on the line, and it took them an abnormally long amount of time to solve, it would be a big problem. Broadcasting is a 24 by 7 on-air operation. Engineers and system architects love, and are willing to pay for, redundancy. However this doesn’t translate easily into a networked IP environment, where failures can cascade across even seemingly redundant systems. Everyone can see the benefits such as cost saving and increased flexibility of networked IP infrastructure, but they have to be able to realize those benefits without undue amounts of pain. The gap between what “theoretically” should work and what “practically” does work needs to be eliminated.
Q: What’s some of the biggest impacts of content on server connectivity?
TR: For broadcast media, there’s a lot of concern about how media is being protected, and whether it’s too easy to steal. There are media security and rights, as well as network and resource access aspects to this issue. We need to make our data more secure but not so much that it’s difficult to get to. A lot of this is being addressed of at consumer level and in-home implementations; less so inside the professional broadcast facility. Traditionally, there has not been lot of risk management principles being applied to the media itself. Security often means secured access to a given file or tape, however, if your content is being routed and delivered over a networked infrastructure, you have to have more built-in security features because digital theft becomes much easier. Someone from outside the country can now steal an asset that could before only been stolen by someone inside the building. Network configuration, VPN’s, provisioning, access and management are critical issues, and all have impact on workflow, accessibility and overall system reliability
Q: What is the one thing you have to say about Ethernet as a technology today?
TR: Right now, the industry tends to look at networking media and the management of networking as entirely separate applications. There are a lot of tools out there for managing and provisioning, but they need to be more tightly embedded into the overall solution. Having an engineer come out to set up the switch, then a security expert come out and set up the security…it’s inefficient. There are too many experts and specializations required to implement and manage networks today. We need to see more of a plug-and-play model.
Microsoft’s Windows 7 made big advances over previous versions in terms of networking – it’s way more intuitive than it was before. That type of intuitiveness needs to span any device that has an IP connection. You simply can’t have a situation where every manufacturer, device and application, does it a little bit differently. By the way, my XBOX – PC issue turned out to be a problem between Windows Media Center and the Windows 7 Firewall. Greater standardization would help alleviate this problem, but as this example shows, getting there is going to be difficult. In the long run though, interoperability is a primary barrier to Ethernet adoption.
Q: What other challenges would you like to see addressed across the Ethernet ecosystem?
TR: From the technical side QoS issues need to be addressed, either as dedicated “virtual” circuits or in some other manner where it is easy to see and manage how one device or application connects to another. I believe there should be a certain amount of self-discovery and self-awareness between devices that can connect together. Anything on a given segment or domain must be able to understand what else is on the network, what it can “connect” to and what the rules are for provisioning such a circuit. For example, if you have a network delivering video in real-time and somewhere else someone decides to run a browser while yet somewhere else a scheduled process begins a backup, it must not interfere with the video stream. We need to see greater coexistence. There also needs to be a better, more accurate way of knowing exactly what’s happening on a network, and not requiring someone with a master’s degree in order to do so.