Measure the difference between a container and a VM

applied · 60 min · Objective 1.4

Task

Run the same trivial application as a container and as a virtual machine, and measure start time, image size, memory overhead and density. The lesson asserts the differences; this lab produces your own numbers for them.

Steps

  1. Pick a trivial HTTP service. Package it as a container image and record the image size. Build a minimal VM running the same service and record the disk image size.
  2. Measure cold start for each: from the start command to the first successful HTTP response. Take five measurements each and record the median in lab/compare.csv.
  3. Measure memory: the resident set of the container versus the memory allocated to the guest.
  4. Start as many containers as you reasonably can and record the number before the host struggles. Do the same for VMs. Record both densities as rows in lab/compare.csv.
  5. On the host, find the container's process in the ordinary process list and record its PID as seen from the host and from inside the container, as a row in lab/compare.csv. Write one sentence on what that proves about the isolation boundary. Then remove every container, image and VM created here.

Verify

awk -F, 'NR>1 && NF>=3 {n++} END {print n" measurement row(s)"}' lab/compare.csv
grep -Eci 'pid 1|pid=1|process list|shared kernel' lab/compare.csv
grep -Eci 'density' lab/compare.csv

The PID observation is the important one: seeing the same process with two different PIDs is the clearest demonstration that a container is a namespaced process rather than a machine.

This is an independent study companion for CompTIA Cloud+ CV0-004 and is not produced by or endorsed by CompTIA.