Some thoughts on designing an excellent and stable queuing system in the game:
1. First, we need to consider the high availability (HA) issue. The queuing system cannot be designed as a new source of failure during design. Design cannot itself be a single point of failure. One of the stand-alone queuing systems is similar to a Rate Limiting Algorithms problem. It is possible to make some functional improvements based on leaky buckets or token buckets. But once it involves high availability issues, it's different. Because the system depends on a centralized queue data. But to design centralized data into a distributed high-availability system, such a design is very challenging.
2. What needs to be considered is load capacity. For example, if 100,000 or more people are online in a short time, how to bear such traffic while ensuring the stability of the service is also a big problem.
3. What needs to be considered is the response speed and circulation. An ideal platform-level queuing system must be able to support multiple servers and multiple games, and queue each server and each game separately. Then the total throughput of such a system is very large. How to ensure that all users can pass through in a stable and fast order is a great challenge.
4. What needs to be considered is robustness, or resistance to hackers. To ensure a certain amount of circulation, but not be blocked by attackers, such a system design is not simple. This is because before the system that logs into the game, whether it is a CC attack or a DDOS attack, it must have stronger resistance to ensure that the business system is not affected.
For more security technical articles, please follow the "Game Security Attack and Defense" official account, learn together and make progress together.