Network

Builds on Common Methodology

Times the in-sandbox command runtime to clone a pinned GitHub repository or install pinned npm or PyPI dependencies from a cold cache, by the sandbox's own clock.

Whenever a developer or coding agent initializes a fresh sandbox without a pre-baked project image, the workflow waits on the sandbox's outbound network connection and disk to clone the repository and install dependencies from public package registries.

The Network tab (Package Registries) measures the guest-clock runtime of 3 everyday cold-cache installation commands, git clone from GitHub, npm ci from the npm registry, and pip install -r from PyPI.

View the Procedure View Package Registries Results Download Raw Data View the Parameters

Metrics

Package Registries is the in-sandbox runtime of the selected cold-cache installation command (Install), measured with $EPOCHREALTIME.

Timed Part	Starts	Ends
Install	The workload's install command starts, right after setup.	The command exits, by the sandbox's own clock.

Measurement

Each trial uses a sandbox created from BoxRouter's pinned standard image with outbound network access to github.com, registry.npmjs.org, pypi.org, and files.pythonhosted.org.

1. Create sandbox, run readiness check, and upload pinned lockfiles. Each trial uploads the pinned lockfiles (axios-package.json, axios-package-lock.json, or fastapi-requirements.txt) to /tmp.
2. Run a background script as non-root user box that clears ~/.npm and ~/.cache/pip and prepares /workspace (setup, timed, not charted).
3. Time per Install Records /proc/net/dev received bytes, runs the selected install command (Install), and records received bytes again. Timed A background script runs the variant's install command timed by $EPOCHREALTIME and /proc/uptime (Install, polled every ten seconds) and verifies the checkout commit or installed package exit code.
4. Destroy sandbox.

Every repository and lockfile is pinned by commit and SHA-256 hash across three cold-cache installation variants:

• git clone performs a full clone of fastapi/fastapi and checks out the pinned commit.
• npm ci installs 687 locked packages from the pinned package.json and package-lock.json of axios/axios.
• pip install -r installs into a new virtualenv from a hash-pinned requirements file exported with uv export --frozen --no-dev --group tests --extra all --no-emit-project from fastapi/fastapi's uv.lock.

Scope Limits and Caveats

• End-to-end install time combines network download speed from public registry CDNs with in-sandbox decompression and filesystem writes; shared registry slowdowns outside a provider's network can affect all providers.
• Comparing the received bytes, recorded on every trial, with the Disk and Filesystem tabs helps tell a download-bound time from a CPU-bound one.
• Both npm ci and pip install make a brief registry check for a newer version of their own CLI at the end of a run, as a developer sees with default flags.
• Trials run in parallel per provider, up to 10 at once, so up to that many sandboxes on one provider can download from the registry at once.

Failures and Exclusions

Outcome classifications, top-up rounds, and outlier rules follow Common Methodology; the first failed step ends the trial without in-trial retries.

A trial succeeds only when cache clearing and the install command exit 0, the guest clock passes the twenty-millisecond /proc/uptime gate, and the sandbox is confirmed destroyed.

Step	Time Limit
Create	5 min
Setup	2 min
Install	10 min
Destroy	5 min

Statistics

Based On

Common Methodology

Per-unit percentiles, bootstrap confidence intervals, and tail rules follow Common Methodology.

The All view aggregates each unit's selected statistic across the workload using the unweighted geometric mean and missing-unit substitution penalty defined in Common Methodology. A provider is ranked on a variant of Package Registries once it has at least 5 successful trials.

Sandbox and Trials

Image Digest

sha256

Image Version

0.3.11

Image Registry Page

View on the Registry

Pull the Image

docker pull ghcr.io/boxrouter/standard-image@sha256:89d4c97048dbdfb8b6d85ad91833962ff38d1c334a5f47e6de80a9762310191b Copy

Lifetime

15.0 min

Idle Connection Time

5.0 min

From Common Methodology

Idle Connection Exceptions

Provider	Rule	File Transfer Idle Time
Daytona	The SDK has no idle-timeout setting. Its command channel keeps its connection open, and its file-transfer channel reconnects after 5 s of idle time.	5.0 s

From Common Methodology

Network Path Attempts

10

From Common Methodology

Mode

Single

Sandboxes per Trial

1

Parallel Trials

10

Concurrent Sandboxes

1

At a Glance

Quantity

Time

Per

Install

Unit

Milliseconds (ms)

Direction

Lower Is Better

Statistic

Median

Clock

Sandbox clock

From Common Methodology

Machine Size

2 vCPU / 4 GiB

Trials

20

Variants

• git clone
• npm ci
• pip install -r