Well it was a big day for Disney a couple of weeks ago when they launched the much anticipated Disney+ streaming service.
And of course as widely reported, they had a really bad outage.
We’ve heard the old refrain before. We never saw traffic like this before. We were absolutely buried.
Join 35,000 others and follow Sean Hull on twitter @hullsean.
Indeed, Disney, in their own microcosm never saw traffic like that before. But does cloud computing know how to solve this problem? Has the software, techops and devops profession solved these problems?
Answer: Yes. Just look at Facebook, Google, Netflix or a million other high traffic sites.
Interestingly, I got a letter from a recruiter over at Disney just the other day!
I wanted to reach out again to see if you’ve had an opportunity to think about exploring a future role with Disney Streaming Services?
The company is in the midst of a massive engineering expansion, as Disney+ launched today in their US, Canadian and Dutch markets. While phased with natural issues, given the pure scope of their launch, paired with unforeseen customer engagement, the team is continuing to actively to build out their engineering team in order to ensure seamless launches in Western Europe and the Asian Pacific countries in late 2019/ early 2020.
While you may not actively be on the market, I would be interested in having a preliminary chat to see if joining and impacting one of the largest scale technical projects the international video streaming industry has ever seen, could align with future career goals?
Looking forward to your response!
1. Management maturity
When I was reading the article above I stopped at this line:
“This is a new ballgame for Disney”
And that tells the whole story doesn’t it? Just because the industry has solved a problem, just because there are technologists out there who know how to do this, doesn’t mean they work at Disney! And further doesn’t mean Disney’s management is ready for their new reality.
But they learn quickly!
2. Streaming is a solved
The challenge of streaming content on the internet is not new. The pipes are there, the cloud can scale seemlessly. But yes there are a lot of moving parts. That’s what testing is for. And for a launch like this, one could easily launch a million test clients, using aws regions and zones around the world. And point them all at your new streaming service. Don’t want to spend the money? Scale down your stack, and send a proportional amount of traffic.
If you’re not seeing the autoscaling happen quickly enough, spinup *lots* of spare compute before launch time. That’s another option. You can always scale back after the flash hits.
Sites like these deal with an explosion of traffic in a short period each day. Typically 90% of their traffic occurs in half to one hour of the day. So it’s a real herd that pummels the site.
Related: 6 Devops interview questions
3. Have doubts? Ask experts
Yes we can monitor. Yes we can test. Yes we can automate. Yes we can react. But still systems fail.
As a small pitch, I’ve helped companies like Hollywood Reporter (100m uniques per month), AppNexus, ideeli and SoulCycle scale their systems for hypergrowth. It’s not easy but it can be done!
4. Good problem to have?
Oscar Wilde said…
The only thing worse than being talked about is not being talked about.
And if today’s political climate is any indication, their is some prescient irony buried there.
So even though Disney+ and many other big names have failed…
Perhaps it’s a good problem to have?
Get more. Grab our exclusive monthly Scalable Startups. We share tips and special content. Our latest Why I don’t work with recruiters