Build Streaming Solutions for Tomorrow, Today
Ask the folks at any web startup or any technology company that offers a cloud-based service, and they will tell you they play the "What happens if?" game all the time. What happens if one of our major data centers goes dark? What happens if network operators throttle bandwidth? What happens if all of our users sign in at the same time? The reason most of these companies ask these types of questions is because they are built on breakage models. (The most well-known breakage model is actually your local gym; what happens if everyone shows up at the same time?) Software service companies, like all companies built on finite resources, rely on subscribers paying monthly or annually, but they typically have only built their services to withstand a percentage of user concurrency. Of course, as their user base grows, hopefully, so does their infrastructure: servers, network capacity, customer service personnel.
Streaming is no different. But being around for 25 years has given streaming a lot of breakage moments to learn from. Each time there's an "internet record" of simultaneous users at a certain bitrate, platform operators have an opportunity to assess their own infrastructure (especially when those breakage moments actually cause something to break) and make changes to ensure the same situation won't happen again. And each moment builds off the last, leading toward a future in which streaming video operators can realistically support tens of millions of concurrent users at broadcast-like quality.
What if that future were today, though? What if it was happening right now?
The COVID-19 pandemic and stay-at-home orders from local and national governments have created a perfect storm for the streaming video breakage model: more people watching streaming video at more times during the day, contending with more people using every other internet service. No platform operator could have predicted an exponential growth in viewership overnight, compounded by global launches (Disney+ entering the U.K. and other European countries), theatrical releases being pushed to digital early, game drops being fast-tracked, and the launch of new streaming video platforms like Quibi, HBO Max, and Peacock. It's not just a perfect storm for the internet and streaming video; it's digital Armageddon. Governments have even intervened, asking major streaming providers like YouTube and Netflix to lower bitrates so that work-based internet usage, for all those people who work from home now, doesn't get throttled by everyone wanting to watch the latest episode of their favorite show in 4K.
But the current situation is also a unique opportunity. Again, this kind of concurrent usage and flattened traffic patterns was most likely predicted by the majority of video platform operators. What they couldn't really predict, though, without having the sort of scale we are seeing now, is where the breakage would happen. What systems would fail? Would it be physical infrastructure getting overwhelmed (just too much I/O)? Would it be the pipes filling up? Would it be the cloud services stretched beyond the capacity of their elasticity? You could say we are in a giant petri dish right now, the ultimate streaming experiment to determine how future architectures and technologies need to be deployed, configured, and optimized to ensure that when this unique situation, this perfect storm, does become the norm, the streaming providers are ready for it.
Thinking of this all as just "weathering the storm" isn't the correct approach, though. Instead, you're preparing for the future. Yes, the current situation will get resolved. Things will go "back to normal." And when they do, streaming operators can take a deep breath and know that they've built enough to handle what they had projected for today's demand. With this opportunity, platform operators should take a deep look into their architectures and technologies to get a jump on the future. For example, by implementing solutions to measure end-to-end delivery, they could fix workflow components that are not performing well under the current load, so that when the current load becomes the norm, their architectures are already future-proofed. Of course, it's hard to juggle all the balls and examine them at the same time. Still, there might not be a better opportunity to at least gather data on the impact of mass usage. If platform operators can implement the appropriate measurement tools for their entire workflow (most video distributors have been hyperfocused on the player for gathering data), then critical information can be used to influence the future-proofing after all of the fires have been stamped out.
The streaming industry probably isn't going to see the level of demand it's experiencing now for years to come. Hunkering down and just riding out the storm are like hiding in the back room when all of the members of the gym descend upon the front doors. The only way to face adversity and learn from it is head-on. I only hope that when the dust settles, all the video distributors, service providers, and network operators share where their infrastructure broke so that we aren't all trying to weather the next storm by ourselves in the back room.
New data from Mux Video customers shows just how much video delivery has grown for B2B and niche markets during the COVID-19 crisis.