Prerequisites

• Redis server
• Python 3.7
• $ pip install -r requirements.txt

Description

Solution contains two independant services:

• webserver responds to http requests with the list of running tasks
• scheduler creates players and then periodically creates tasks for each

Start

$ python -m game.webserver.main

$ python -m game.scheduler.main

Results

Params: 20k players, 32x32 view, ~80k tasks total, ~350 tasks running

Request-Response time: ~0.04s

Possible improvements

• multi dimensional indexes: https://redis.io/topics/indexes#multi-dimensional-indexes

Redis structures

Could be configured from settings.py

Players

players:by_x - zset with scores by X coord and player_id members (<X>x<Y>)

players:by_y - scores by Y

Views

view:<xstart>-<xend>:<ystart>-<yend> - lists of players in x-y diapason

Individual tasks

<player_id>:t[1-4] (e.g. "396x22:t1", "160x169:t3") - key with ttl, per task

Task lists

tasks:<player_id>