Return to Blog
November 30, 2017
3 min read time

Live video streaming secrets: Part one

Arianna Aondio
  • live streaming
  • how to stream live video
  • live video content delivery
  • scale up for unpredictable web traffic
  • continue serving traffic in a server outage

    • Leading up to our December 12th live webinar, Six secrets to successful live streaming, I wanted to share some thoughts on preparing yourself for optimal live video content delivery, introduce some of the topics I’ll discuss in the webinar, and at the same time invite you to register to participate.

    How can you seamlessly serve content even if your video goes viral?

    We should all be so lucky. We create content hoping it will go viral - but virality is unpredictable. How can you take steps - and what steps can you take - to protect yourself in the event that your content does go viral?

    Here you want to plan for the worst possible future scenario. Your architecture must be resilient and be able to cope with unexpected peaks.

    During the webinar I’ll present the use case of a company whose website got flooded due to (apparently) unrelated events.

    How can you keep your platform up and running even if some (or all!) of your servers crash?

    Symmetry is key: never, ever plan your architecture with a single point of failure. Use at least a couple of Varnish servers with cache replication so that if one of the two crash, you still have the other one, which can fulfill incoming requests. If ALL your Varnish servers crash (that would be very unlikely), you can still rely on the cloud which, if combined with persistent cache, will always be available.

    How can you protect your backend from being overloaded?

    Varnish can act as shield and origin protect for your backends. The reason for this is that CDNs and any other content delivery solution can send a lot of requests to your web servers and put a lot of pressure on them.

    Varnish can sit next to your backends and protect them from overload and, given that all the traffic will have to go through Varnish, it adds an extra layer of security to your architecture.

    How can you proactively scale for the unpredictable?

    This question is in essence related to the first. Your traffic is unpredictable - so you really need to be ready for anything before you even get started. What does that involve?

    Apart from what we have already discussed, and what I will discuss in greater depth during the webinar, part of scaling your platform properly involves understanding how to size your platform based on your assumptions about your live streaming audience size and demand and understanding how to calculate your needs from there. (My colleague, Guillaume, wrote a helpful blog post on sizing your Varnish platform for live streaming not so long ago.) Scaling requires having a solid understanding of what you need to deliver and then figuring out what you might have to deliver in the unpredictable, peak conditions you are probably hoping for. The proactive approach comes in because you plan for the unpredictable from the outset.

    Of course, I can’t reveal all the secrets here because I hope you’ll attend the webinar to get all the details.  

    REGISTER FOR THE WEBINAR