BACKEND SUBSYSTEM DEVELOPMENT AND REAL-TIME PERFORMANCE MONITORING SUBSYSTEM ON 5G CLOUD GAMING PROTOTYPE
With the increasing fidelity of games developed in recent years, the computer specifications required to play new games have also increased. The result of this escalation is the rising cost that users need to incur to play a game. Therefore, a system is needed to provide access to consumers who d...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/82257 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | With the increasing fidelity of games developed in recent years, the computer
specifications required to play new games have also increased. The result of this
escalation is the rising cost that users need to incur to play a game. Therefore, a
system is needed to provide access to consumers who do not want to invest in a
gaming PC, allowing them to use a remote PC for gaming. This system is called
cloud gaming. The cloud gaming system to be designed will include several
subsystems, namely the backend subsystem, automation subsystem, network
subsystem, hardware subsystem, and finally the frontend subsystem, as well as
client and server monitoring subsystems. The subsystems to be implemented in this
document are the backend subsystem and the client monitoring subsystem. The
backend subsystem will be implemented with a microservice architecture design
where the API will be built using Golang. The database to be used is a relational
database (PostgreSQL). VPN will be used for client-server connection
implementation. For the client monitoring subsystem, Python will be used to
monitor the load of the Moonlight application and cloud gaming QoS. From the
implemented system, API performance results were obtained as expected with a
maximum latency of 300 ms and throughput of 40 requests/s. Additionally, the API
only needs to use two CPU threads with a total RAM of 1.5 GB to handle that load.
As for the client monitoring subsystem, metrics for cloud gaming QoS obtained from
Moonlight and application load from Moonlight, such as CPU usage, memory
usage, and bandwidth, were successfully obtained in the testing. From the research,
we can conclude that this Moonlight monitoring method has not yet been
implemented in other studies. From the data obtained, for a client PC with
specifications of 6 Core (12 Threads) CPU, 8 GB RAM, and Integrated Graphics,
with a preset of 1080p60, CPU utilization of 3% (36% for Single Thread utilization)
was obtained, RAM usage was 190 MB, with bandwidth usage of 2.5 Mbps (40
Mbps) for downspeed. The latency obtained was 1 ms (under normal conditions
and conditions with 200 Mbps background traffic load) Keywords: Backend, Rest
API, Cloud, Cloud Gaming. |
|---|