English · Français

Main docs (HTML) · API reference · PDF manual

SysResEval (Système Réseaux Évaluation, shortened to SRE) is an open-source software suite for teaching and managing exercise sessions and evaluations in network and system administration courses. Each "lab" or project is a small virtual network of Linux machines, orchestrated through the Kathara framework on top of Docker and VDE. Students work through guided exercises in a GUI; instructors monitor sessions and run time-limited exams with automatic, reproducible grading.

It is developed at the IUT d'Orsay, Université Paris-Saclay.

• Two-tier interface. A desktop GUI (sysreseval) for students, plus a full-featured CLI (sre) for instructors and authors. The GUI shells out to the CLI through a small setuid sre-wrapper helper, so privilege boundaries stay clean.
• Rich lab structure. Every lab present several tabs:
  • Schema — the network topology (some machines may disallow connection (in red) or be hidden)
  • Informations — general background and instructions, written in Markdown.
  • Questions — a list of items the student must complete. Each question is one of:
    • a Markdown text describing a task to perform on the machines (e.g. configure the network on m1),
    • a form whose fields capture specific facts (e.g. what is the MTU of eth0 on m1?); fields may be free-text validated by a regexp (e.g. an IPv4 address), a dropdown, or a checkbox,
    • a free-form multi-line text answer.
  • Terminals — an embedded shell for each machine the student is allowed to connect to. A lab may expose these as plain root shells or, for example, as a login prompt that restricts students to a particular user account.
  • Machines — a status table for each machine: state, NAT network, exposed ports. The students can use it to launch separate terminals sessions to a machine (useful when several terminal sessions are needed on one machine)
  • Evaluations — lets the student trigger an automated evaluation of their work and view the resulting grade table.
  • Apply Configuration — lets the student put the project into a predefined state, for example a partial correction.
• Multiple labs open at once. Course content can be organised by topic rather than by session — students can keep several projects running and switch between them.
• Live classroom monitoring. Instructors can watch the whole class in real time, inspect each student's latest evaluation and grading errors, and see per-item min/max/average statistics across the class.
• Time-limited exams. sysreseval starts each student's project (immediately or at a scheduled time), shows a countdown, runs periodic evaluations, and posts an end-of-exam banner. Duration can be adjusted on the fly — useful for accommodations.
• Reproducible grading and post-exam tooling. Each evaluation is archived as a compressed msgpack file capturing the project data, raw command outputs, student answers, and per-item grades — inspect any archive with sre cat. If a grading bug surfaces after the fact, sre re-eval re-runs the scoring against an updated script; sre outline then produces per-student PDF reports and a recap ODS spreadsheet. During exams, every terminal session is also saved as an asciinema cast for later review.
• Internationalisable. Lab strings and GUI translations ship in French and English; tools (prepare-sre-translations, add-sre-translations) make multilingual authoring straightforward.

Download the full-quality video (MP4)

SRE is a Linux server-side install (Debian/Ubuntu, any other distros with manual dependency installation). Production deployments live under /opt/sre and are typically shared across a classroom via NFS.

git clone https://github.com/sysreseval/sysreseval /opt/sre
cd /opt/sre
sudo ./scripts/install.sh        # interactive — ~10 questions

The installer creates the sre system user, drops a sudoers rule, builds the C wrapper and Python venv, and (optionally) installs a .desktop entry, bash completion, and a sre-preload-images.service systemd unit.

See docs/sphinx/installation.md for the post-install set-up:

• raising inotify limits,
• configure X to listen to TCP port 6000,
• restricting student access to Docker during exams,
• sharing archive directories for live monitoring,
• pre-loading Docker images to avoid a network meltdown when an exam starts.

Full step-by-step setup, manual install, and post-install steps are documented in docs/sphinx/installation.md.

• Overview
• Installation & deployment
• Running exams · exam reference
• Authoring labs · translations
• CLI reference · GUI reference
• Runtime & internals