You are on page 1of 1

Topic 2 – Performance Analysis inside a docker container using an NXP

i.MX 8M mini
More important, start here.

We aim towards performance penalties and bottlenecks for applications running inside a docker
container.
1. Memory (disk) usage/requirements of the docker runtime
2. Memory (RAM) footprint of the docker runtime
3. Memory (disk) usage (in addition to 1. and 2.) for a docker container (only the container without
its application)
4. Memory (RAM) footprint (in addition to 1. and 2.) for a docker container (only the container
without its application)
5. Startup time of applications/processes inside vs. outside a container. (A corresponding and
meaningful application has to be developed)
6. Runtime measurement of applications/processes inside vs. outside a container. (A corresponding
and meaningful application has to be developed)
7. Timing of inter‐process‐communication (IPC) events inside vs. outside a container
7.1. Before this task give a short overview on the available IPC‐mechanisms for our platform

Freedom of Interference
How do interfaces such as ethernet or storage drivers interfere with each other and with
applications?
1. Directly (e. g. writing / reading from disc)
2. Indirectly (e. g. parallel access of the ethernet of two applications; what is the network
throughput?)
3. What else can go wrong with this hardware architecture (i.MX 8M mini)?
4. Analysis / timing measurements in real world scenarios; e. g. a denial of service attack (via
ethernet)

Literature:

https://www.embedded‐software‐engineering.de/echtzeitfaehigkeit‐von‐container‐loesungen‐am‐
beispiel‐docker‐a‐857072/

https://scoutapm.com/blog/restricting‐process‐cpu‐usage‐using‐nice‐cpulimit‐and‐cgroups

https://embeddedbits.org/using‐containers‐on‐embedded‐linux/

https://www.whitesourcesoftware.com/resources/blog/docker‐container‐security/

You might also like