Can BeReal Scale?
I never did any real software engineering, just read stuff, so i wanted to ask HN engineers, what they make of BeReal. The app is really hyped right now, but I thought that it might be really difficult for them to scale beyond a certain point, because of the apps nature and concept. Essentially BeReal has virtually no traffic all day long, with a very high peak at one point with a lot of data. Could this be an issue?
It would be a good case for cloud servers. They could have a script that starts a bunch of servers just before the event and then shuts them down afterwards.
Issues with scaling typically occur when confronted with unbounded traffic issues.
BeReal has a known traffic cap, say, 3 or 4 billions peak connections at the same time.
This cap has already been solved by other tech companies if i’m not mistaken.
Yes, it can scale. The ideal way to do it is to queue all the updates sent to their server and process them by spinning up extra VMs right when they are about to get the burst of traffic (easier to do since they themselves determine what that time is). That way they don't have to pay for resources to sit idle the rest of the day, and a slight delay (say, a few seconds) before the photos actually post and are visible to everyone is acceptable.
I think scaling BeReal is probably significantly easier than a lot of services.
1. You have a huge burst but you pick when that happens. That means you can provision the appropriate hardware just in time.
2. Storing and serving images is really easy now with a multitude of satisfactory cloud services that are reasonably priced.
3. You can move stuff to colder storage pretty aggressively.
I'd never heard of it until I saw this thread and looked it up. The idea is interesting, but... I don't think technologically they'll have much issue scaling. Sure, there's peak traffic, but they can easily queue traffic to spread the load if needed, especially knowing that most of the time will be idle time. The question is whether scaling is even worth it for them. On the one hand, I like the idea of deglamorizing social media. On the other hand, encouraging an inherently cross-referenceable combination of people with their surroundings every day, along with GPS data, means this might be even more of a privacy nightmare than most social media. And even the reactions are apparently selfies, increasing the amount of location data collected.
Plus, people like the overly glamorized ideals of other social media -- even though we all hate it because it's an unachievable ideal, it's more exciting than reality. That's why "reality TV" is simultaneously absolute mind-numbing dog shit and also watched by millions regularly. It's that cognitive dissonance between "this is ridiculous and unrealistic and probably unhealthy to see all the time" and "but it'd be fun to pretend it's real, though".
So I'm not sure they have a platform that's feasible to grow in the first place.
They already spread the update notifications by timezone (so you don’t get one in the middle of the night). I guess it should be possible to spread out the notifications even further by building clusters of friends, that receive it at the same time, instead of sending it to everyone in one timezone at the exact same moment.
Anything social that has not embraced the self promotion aspect of it hasn’t been successful in the long run at all…that I can think of. Seems like a tough sell.