.jpg)
It seems that every Tom, Dick and Harry is in the Video Optimization space nowadays. Seriously, every vendor with video expertise claims to deliver a smoother video experience with little interruption. As a consumer, I applaud this attempt at creativity. As an industry insider, I’m amazed by the confusion in the market. Let’s review the relevant technologies and separate the hype from the substance.
First, what kind of videos are we talking about? Well…what kind of videos are you watching? YouTube! YouTube and other internet video sites typically deliver content over HTTP using Flash or MP4 encoding. While we watch the video ‘stream’, it is technically being downloaded in small chunks rather than truly being streamed. This means that a video optimization product must ‘inhale’ HTTP Progressive Download (that’s the techie term for the small chunks) and then ‘exhale’ HTTP Progressive Download toward the client. Otherwise, it wouldn’t be able to affect YouTube and YouTube-like content.
Many self-proclaimed video optimization companies focus on completely different content – that is, content streamed using RTSP to devices supporting 3GPP-compliant video players. This is useful to content providers (or network operators acting as content providers by offering subscriptions to premium video channels from their portal), but not to operators who are trying to deliver the content that you and I watch on popular internet sites through their network.
Second, what kind of optimization are we talking about? Many vendors are talking about transcoding, which technically means conversion of content from one format to another. Sometimes transcoding doesn’t change the codec (format). Instead, it performs the function of reducing the number of frames or the resolution. By definition, these content conversions can only reduce the quality of the video – frame numbers or resolution cannot be increased in the process of transcoding. In fact, video quality will be reduced in any conversion process, even if the reduction is unintentional.
Hold on…so what’s good about transcoding? It seems that it can only diminish the user experience. Well…that depends. The wireless network is shared by many people using many applications. If the network is congested, your video may stall. Alternatively, your video may play, but my video will stall. If there is no congestion, then video transcoding can only reduce quality. But when there is congestion, video transcoding could help both of us watch videos with fewer interruptions (albeit at somewhat reduced quality).
This can get complex. Some users on the base station may watch videos, others may browse the web, yet others may interact with apps, and without even knowing it, some may be downloading updates to operating systems, media players, apps, and anti-virus software. As a result, one video user may negatively impact several users engaged in other activities. Should that one video user experience reduced video quality, or should the other users experience slower web downloads and less responsive applications?
Also, how do video optimization vendors know that we are all served by the same base station (shared network)? And how do they know that you are, I am or someone else is experiencing the effects of congestion? As you have probably concluded, optimization algorithms and policies have little to do with transcoding. The trick is to figure out if and when to optimize, and what optimization technique to apply to each transaction. Transcoding itself is just a small piece of the puzzle – a piece which happens to be a completely commoditized technology.
So…what is different about Bytemobile? We provide a holistic video optimization solution – not a discrete technology for an isolated problem. We religiously improve the user experience by averting, detecting and preventing congestion. Other vendors provide a piece of technology with the expectation that the operator will miraculously figure out how to apply it. We have proven the performance of our technology – not only on limited-capability wireless phones, but much more importantly on high-end smartphones such as Android and iPhone devices, as well as full-scale laptops, netbooks and, soon, iPads.
Unlike legacy vendors that designed their systems to handle low-bandwidth devices, our system was originally designed for high-end devices – those that you and I use to watch YouTube videos. This requires a unique architecture that enables us to make complex optimization decisions at the speed of the Internet. It requires tight integration within the core of the wireless network, and it requires visibility into users’ activities and experience. Our video optimization technology – like our web optimization technology before it – has been first to market and first to deploy in commercial network environments.
We humbly call it “optimization – video included”.
-Joel Brand
Image courtesy of Altair Libre via Creative Commons Attribution License.