I benchmarked boxr vs podman on container startup. Here are the honest numbers.
TL;DR: On the same VM, starting a minimal cached container 20 times each: boxr 0.1.44 median 134 ms, podman 4.9.3 median 217 ms. boxr was ~38% faster on median startup. The full methodology and every caveat are below — please read them before quoting these numbers, because the caveats matter more than the headline. What I measured The simplest meaningful container operation: start a container…
Starting a minimal cached container 20 times each on the same VM, boxr 0.1.44 had a median startup time of 134 ms, while podman 4.9.3 took 217 ms on average. Boxr was approximately 38% faster in terms of median startup. The methodology and important caveats are detailed below. The test focused on the simplest container operation: starting a container from a pre-pulled image and exiting immediately.
Both boxr and podman used the same Alpine image, with identical configuration digest sha256:320994c3... The wall-clock time measured included CLI startup and captured the user experience. Five cold runs were performed on each engine, followed by 20 warm runs each, alternating between boxr and podman. No runs failed, and all data points were kept.
Warm results (n=20 each) showed a median of 134 ms for boxr, 150 ms for podman, with minimum and maximum times ranging from 107 ms to 262 ms and 181 ms to 310 ms, respectively. Cold runs (first 5 invocations) indicated a median of 128 ms for boxr and 212 ms for podman. The test environment consisted of 2 vCPUs, 7.7 GiB RAM, running Ubuntu 24.04.5 with kernel 7.0.0-38-generic.
Boxr was the prebuilt v0.1.44 Linux x86_64 release binary, while podman was installed through Ubuntu's packages with crun 1.14.1. Caveats include the fact that this benchmark does not compare rootless podman, as the VM itself is a user-namespaced container with restrictions. The engines ran as root, potentially giving podman a slight advantage due to skipped user-namespace setup.
Networking was disabled for both engines, and podman required --network none due to setns: Operation not permitted in this environment. Podman's default storage chose fuse-overlayfs, but it was pointed at tmpfs to use native overlay with idmapped mounts, likely helping podman. These results suggest a ~80 ms gap between boxr and podman, which is consistent with architectural differences rather than tuning.
Boxr's design should benefit short-lived containers, while podman may be more suitable for long-running services. Future comparisons could include rootless podman, real workloads, cold page caches, and additional engines like Docker and nerdctl to provide a more comprehensive comparison.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
