Read about how Polydelic architectured an app that has over 1 billion server requests per month.
Geo-hashing. This is a way to query locations around the planet using only a string of characters. It has several speed and storage advantages over 2 dimension lat, long queries. We can query every single user in one location with a radius, even if there are millions of users in that location.
This is scalable. 1 million users or 100 million. Doesn’t matter.
When you're doing real-time dating it's extremely important to protect users against maliciousness data leaks.
We developed a modular flat data structure making it easier for our developers to implement security rules.
Every major release we hire penetration testers and quality assurance professionals to ensure that the app is secure.
We put into place pre-trained machine learning models to detect potentially harmful photos and nude material. We also developed a technique for filtering faces only for profile pictures. Instead of hiring a dedicated team of people to go though each and every photo we saved the client a ton of time & money. We also gave the users are more streamlined experience.
Users may report, flag, block and unmatch other users. Our client can login to a backend admin panel and view these reported users. Using the admin panel they can block, search and directly message other users.
We used the Google Cloud Platform, it's runs on 100% energy renewable tech. Read about some of the advantages we have using this architecture.
If you have 1 million new users overnight in a new location traditional architectures with operational managers and backend developers might struggle. Using Google Cloud it will assign resources automatically to that geolocation. Stress-free cost effective scaling.
Using lambda technology we can send push notifications to any device (Web, Android & iOS). When users receive messages inside the app they'll receive a push notification instantly.
When opening the message screen users can reply to messages they receive in real time. We use Firestore (by Google) to achieve this. One thread can support up to one million users real-time.
Cloud testing allows us to test on many different Android devices at the same time to check for bugs and crashes. Also allows us to stress test making sure services will run smoothly when it's live.
Our client is sometimes featured in the media resulting in our app peaking in usage. We will see spikes of millions of requests within hours. Our client's app has had 0 seconds of downtime since the release.
We have real-time server and device crash monitoring. So far with this app, we are 99.9% crash free. (Extremely good).