Pokemon Go players might disagree. Niantic's Pokemon Go had a very rough launch in July as the viral popularity was not anticipated. Now the Google Cloud team gives some insights what happened in the background.
In the first days of Pokemon Go server downtimes and failed login attempts have been part of the life as a Pokemon trainer. Niantic even had to cripple the app to get to the point the servers can stabilize and allow Niantic to expand into more countries. The server stability was such a big issue that the community build tools to check if the servers are up or not.
Don't Miss: MWC 2017 Highlights
A blog post by a member of the Google Cloud team paints a heroic picture of how Niantic and Google Cloud developers managed to keep the service up during the launch. The key problem was the dramatic and unexpected amount of traffic Pokemon Go was driving. When planning the cloud resources Niantic and Google Cloud put the worst case traffic at 5x the anticipated traffic. The teams got to confronted with 50 times the estimated traffic.
Now clearly 50x is a huge factor. On the other hand this is Google. I understand that scaling a cloud infrastructure by big factor takes time, but it took about two weeks until Pokemon Go servers settled to deliver near 100% uptime.
The blog post gives some insight in what has caused the instability, but does not cover all issues. There was more to it than just scaling the server resources. The application logic for Pokemon Go runs on Google Container Engine (GKE) powered by the open source Kubernetes project.
Google Cloud and Niantic engineers had upgrade to a newer version of Google Container Engine that would allow for more than a thousand additional nodes to be added to its container cluster, in preparation for the launch in Japan.
Pokemon Go was the largest Kubernetes deployment on Google Container Engine ever. Due to the scale of the cluster and accompanying throughput, a multitude of bugs were identified, fixed and merged into the open source project.
Niantic and Google engineers also replaced the Network Load Balancer, deploying the newer and more sophisticated HTTP/S Load Balancer in its place. The HTTP/S Load Balancer is a global system tailored for HTTPS traffic, offering far more control, faster connections to users and higher throughput overall.
Since the launch in Japan, Pokemon Go servers have been up for the most time. A DDOS attack on July 18 has knocked out the service, but since then there are no reports that hackers have been able to take Pokemon Go down despite threats.
Pokemon Go appears to have a state-of-the-art architecture that will be a solid basis for going forward. Why it takes Niantic now that long to bring back Pokemon tracking is unclear.
Now the question is if the Google Cloud can make Pokemon Go also secure. It is actually essential for Google Cloud that the upcoming new encryption is holding up against the 3rd party developers.
The blog post by Luke Stone, Director of Customer Reliability Engineering has not too many technical insights, but gives some idea of what has been going on behind the scenes during the massive launch of Pokemon Go.
Today is the last day you can win a Pokemon Go Plus in our giveaway. I have the Pokemon Go Plus on my desk and it is ready to ship after the give away ends today. Enter the Pokemon Go Plus Giveaway for your chance to own a Pokemon Go Plus a month before the next wave of inventory arrives. More Pokemon News.