Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

It says they ended up at 100 million daily active users. The fact that they could deploy that capacity in 60 days should terrify their competitors.


I don’t think so, as their competitors already know Google is Google and has virtually unlimited resources. As far as I understand from the article it was mostly a matter of “hey can we have extra capacity”, which shouldn’t surprise any competitors.


The competitors scaled by similar factors from similar baselines. Here’s a comment I wrote in April, collecting the then-current info from the three big ones: https://news.ycombinator.com/item?id=23029151


Google Search had 1.2 billion users back in 2013, now it probably has 3 billion.

Youtube has 2 billion users.

Gmail has 1.5 billion users.

Android/Play Services has 2.5 billion users (well Android has 2.5, Play Services probably has at least 1.5 billion users).

For Google a service with 100 million users is a drop in the bucket :-)


1.5 billion monthly active users sending short text messages is a really different network capacity problem compared to 100 million daily active users sending high-definition video streams to each other.


But it's still small compared to YouTube video content served.

Also, the WebRTC stuff all browser based videoconferencing apps are built on supports peer to peer Comms in most cases, so worst case they could have moved all the users over to that in an overload situation.


Not sure about the comparison with YouTube. YouTube is one to many, and can be heavily cached. Meet is many to many with no caching.


Right, and YouTube doesn't need high priority traffic because it has a deep buffer on the receiver, whereas videoconferencing will be disrupted by even a single dropped frame.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: