Research of Telemetry API Server Latency for Offline RPG Video Game

Abstract

Telemetry is an important element of the modern software support process, involving the collection of diagnostic data for the prompt resolution of technical defects. A classic example of telemetry usage is the submission of critical error reports: in the event of an application crash, the system automatically transmits a memory dump to developers to analyze the root cause of the failure. In scenarios where a program operates without continuous network access, analytics are accumulated locally. The transmission of this data to servers is triggered automatically as soon as the internet connection is restored or when specific system events occur. Video game publishers, particularly of offline RPGs, actively leverage telemetry to analyze difficulty balance, evaluate gameplay, and understand their audience’s preferences. This requires a server application deployed in cloud data centers, featuring a database structure typical of RPGs, and capable of ensuring real-time data ingestion with low latency. As part of the study, an experimental testbed was deployed using five virtual machines running the Ubuntu GNU/Linux OS on a VMware ESXi hypervisor. A comparative analysis was conducted on five architectural configurations that combine native deployment and Docker containerization for the API server and MySQL database. The methodology relies on load testing scenarios that simulate the peak burst workloads characteristic of game clients reconnecting to the network. The results figured out that the most efficient architecture for a game telemetry server utilizes host networking, whereas standard read and authorization operations are successfully executed within isolated container networks. The developed method enables the selection of an optimal server environment, minimizing network latency and ensuring infrastructure scalability to support a large number of concurrent connections.