I have been working in the video industry for more than 15 years and have seen a lot of change over that time. In 2000, TV-based video on-demand (VOD) was in its infancy. There was no YouTube, no Netflix, and no Internet capable of supporting video streaming. Fast-forward to today and it is hard to imagine a time when you couldn’t binge watch your favorite show on any screen instantly at the click of a button. Change in the video industry has been constant. These days, it is my job to characterize that change and put it into context, to help guide my company’s strategy and build more relevant solutions for our customers. Needless to say, I spend a lot of time thinking about it.
In case you haven’t read the papers lately, the video industry continues to change and this time it is more fundamental. I don’t just mean the way people are consuming video. Yes, people are watching more video over the Internet. They are watching in more places within their homes, at the coffee shop, and at the airport. Some people are cutting the cord and relying entirely on Internet-based services. We all know this is happening, but that’s not the kind of change I am talking about. The industry is fundamentally changing from one where video has been a specialty to one where video is increasingly viewed as a feature. “So what does that mean?” you might ask. Well, when video was considered a specialty, it seemed entirely reasonable to use purpose-built technologies to support it. How else could one deliver reliable video performance to customers and use the least rack space and power? When video is a feature, reliability, density, and power consumption are no longer the sole measuring sticks of success. Clearly, whatever application runs in the data center must have these attributes, but there are several other factors that must also be considered. For example, how easily can video capacity be deployed, scaled, upgraded, and managed? How quickly and how frequently can new features be released to customers? How can data center resources be shared across applications to more gracefully account for changing consumer demand? To address these questions, one must begin to view video as a feature of a much broader IT and data center strategy.
Let’s explore this thought through some scenarios. VOD was first deployed commercially to TV set-top boxes in 1999. At the time, VOD required specialized server hardware. Video streams were delivered to consumers over a proprietary video network to proprietary set-top boxes, running proprietary software. Everything about VOD in 1999 was specialized and purpose built.
Let’s fast forward to 2010, when IP-CDN technology was first being deployed by service providers to support commercial video applications on mobile phones, tablets, and PCs. While commodity hardware was increasingly being used instead of special purpose servers, the software running on these servers was proprietary and the networks, while IP, remained proprietary as well. In fact, the first IP-CDN deployments looked remarkably similar to classic VOD deployments, in that service providers built monolithic solutions designed explicitly to support commercial video and nothing else. Video was still considered a specialty. But the tide is beginning to shift.
Today, more service providers are looking at the network holistically and beginning to see video as a required feature of their core IT and data center strategy. Rather than buying specialized technology for each video application, the trend is toward deploying a more flexible IT-centric solution based on common hardware and a common software infrastructure, along with software-defined video solutions that can be tailored more easily to suit the specific application requirements. Software-defined solutions are easier to deploy, easier to scale, and easier to integrate, as they follow pre-defined standards mandated by the underlying software infrastructure. This means operators can do more, more quickly, with less support staff.
So what’s the catch, you might ask? While there is growing interest in migrating away from proprietary technologies towards IT centric, software-defined architectures, video remains a challenge for most general purpose data center technologies. Video files are large, requiring not only scale, but a higher degree of performance than is required for general data. Video delivery also requires very low-latency communications across the network and since it is often streamed live, reliability for video is even more critical than it is for other data types. On top of all that, most video networks today are messy. A variety of technologies, some dating back to the 1990’s, are still in use inside of cable networks. One cannot deploy a video solution without some level of integration into the existing networks.
These are the challenges that companies like Concurrent are working to address for our customers. How do we build solutions that support our customer’s next-generation data center strategies, while also supporting the unique demands associated with video? By changing our perspective on how video should be delivered in the future and how it gets deployed within the framework of the software-defined data center, we are able to build solutions that are more relevant and valuable to our customers across the enterprise.