Delivering Millions of Notifications within Seconds During the Super Bowl

Video with transcript included on InfoQ: bit.ly/3AGjOmV
Zhen Zhou discusses how they built/test an on-demand notification system, what it takes to manage cloud resources/site-reliability at the same time, and how to mitigate reliability issues.
#CaseStudy #Duolingo #InfoQDevSummit
-----------------------------------------------------------------------------------------
Follow InfoQ:
Mastodon: techhub.social...
Twitter: / infoq
LinkedIn: / infoq
Facebook: / infoqdotcom
Instagram: @infoqdotcom

[@fiercyfirefox]

very good story telling, average tech talk

[@cp-petar]

Great video ... The craziest part though is the engineering at Apple and Google back-ends, which handle the billions of messages sent by companies large and small like you :) I want to see how that is architected :)

[@ReadThisOnly]

lots and lots of machines!

[@poke_champ]

throwing shade. Who are you 😂

[@fiercyfirefox]

likely they don't/can't have sla for the same

[@VoltLover00]

Why are you paying AWS for all this? Don't build your house on someone else's land

[@Dom-zy1qy]

Then they'd have to build their own land, then hire people to build homes on it. Even if they just had 2 data centers that's still a large undertaking. I don't really know what the total cost would be, but they'd need to employe people to maintain these in person, and more engineers if their infrastructure uses managed services or services that are AWS proprietary. Plus ironically enough, they'd need to build their data centers on other people's land anyways (buying land in foreign countries), and that process probably necessitates more legal and other types of employees that I can't think of right now. I don't even think this company is particularly profitable to the point where that's in their best interest. But maybe im uninformed, honestly don't know.

[@dhruveshk9457]

from day 1 company should think about cost saving and reliability. for fast prototyping aws is fine.