When planning a live-streaming event like a sporting or e-sporting event, an award show, a corporate announcement, or a political event, you want to ensure that it gets delivered to your audience without excessive latency, quality degradation, or interruption. Delivering a poor experience can lead to upset customers, advertisers asking for credits or refunds, and social media backlash.
With today’s high quality-of-experience (QOE) expectations, no matter what type of content you’re delivering, you can’t afford not to have the right plan in place in advance of your live event. Here are the crucial steps to take when planning your live-streaming event.
Conduct an initial tech stack assessment
Take stock of what you have and what you need. This includes making sure your architecture can scale to meet demand, that you can stream to any device and have capacity in regions with your audience, that you have the ability to collapse requests to reduce origin egress costs, and that you have the ability to manage a multi-CDN or multi-cloud environment.
Other considerations include:
What do you have currently in your end-to-end technology stack?
Can you identify the capacity limitations of the components in your stack?
Do your vendors require a heads up or capacity reservations for big spikes in traffic?
Who are your supporting personnel and vendors?
Will your vendors provide real-time support during the event? How?
What analytics capabilities do you have ready?
Scope your event
Put together a “best case” estimate (including spikes) as well as a “most likely” estimate for your event’s audience to ensure that you have the needed capacity reservation in place. Estimate high. Better safe than sorry. Ask yourself:
What is the expected number of peak and concurrent viewers?
What is the expected geographic distribution of your users?
Do you expect waves of users coming and going or staying on for most of the event
What will your bandwidth and request volume expectations be? Consider things like highest resolution, bitrates, and segment length.
What is your event duration?
What happens if the event is more successful than you expect?
What are the limitations of the infrastructure and what is the backup plan?
Move everything you can to the edge
You’ll improve your overall user experience by having everything closer to your viewers — you’ll also likely save money on origin trips. Look for solutions that complement your existing infrastructure investments. Then think about what you can move to the edge. For example:
Device detection/device-specific logic
Empower your team
Create a team of internal stakeholders and personnel from your vendors that knows your configuration and is available to quickly respond during the event. You don’t want to wait on hold or have to go through various tiers of support when every minute counts. Instead, you’ll want both proactive notifications of everything going smoothly and of problems – complete with the context of whether it’s an issue unique to you or of internet weather. Now is not the time to bottleneck. When troubleshooting on the fly, agility is key.
Create onsite decision-making trees and empower those onsite to make decisions.
Get team agreements in place on how to politely and clearly indicate potential issues with other vendors in your stack.
Make sure you have the ability to roll out or roll back configs as needed.
Make sure you have backups for all critical roles.
Create contact lists and escalation points.
Foolproof your plan and conduct a dry run
Once your delivery pipeline is set up, test a config change to see how long it takes to take effect, and test your backup plans as well, to ensure they work if needed. You’ll also want to establish a multi-vendor communication plan for how you’ll relay information to the various vendors during the live event.
Conduct a full dry run with the technology and the services that you expect to run with a smaller event prior to your large live-streaming event. Make sure this includes all components, like
Your delivery pipeline;
Meaningful logging, both for troubleshooting and for post-event measurements;
Request collapsing/mid-tiers (Verify that origin offload is good enough or add an additional layer like a shield);
Your connectivity (If you’re using a public cloud, ensure your edge delivery has the connectivity you need to your origin or if you have your own bare metal origin consider doing interconnects/peering);
Test for what could go wrong to ensure your resilience plan works as intended;
Your live-event monitoring services ; and
Post-event review, summary, and learnings.
These tips can help you deliver a quality live-streaming event to your viewers. With so much on the line, you want to be as prepared as possible and make sure you have the support you need. With real-time monitoring, streaming delivery, request collapsing, capacity planning, and flexible deployment, Fastly Live Event Services gives you insights into your live-streaming performance and the ability to troubleshoot on the fly.
Learn more in our January 21 webinar, Navigating 2020 with success: A live Q&A with Fox TV & Sports, including how Fox pivoted their content delivery ecosystem in order to continue bringing high-quality streaming entertainment and live coverage to a worldwide audience last year.