Skip to main content

Cybersecurity Lab · Live in production

A production-grade cyber range. Built, secured, operated.

A self-contained cybersecurity lab where learners run real attack-and-defend exercises in the browser — no client software, no exposed infrastructure. Dozens of isolated systems, single sign-on, encrypted access, continuous monitoring, and an automated nightly reset. Designed, deployed and operated end to end by ABOSS.

  • Live in production
  • Browser-based access
  • Single sign-on
  • Segmented & monitored

Download the one-pager (PDF)

Key capabilities

Everything a class needs. Nothing the campus network doesn't.

Browser-based access

Terminal and remote-desktop sessions to every offensive and defensive system open in a tab. Nothing to install on the learner's device.

Single sign-on

Federated to the institution's existing identity provider. One login opens every permitted application; access is granted or revoked in one place.

Segmented architecture

Vulnerable targets and live attack traffic stay inside the range. A hardened reverse proxy is the single front door.

Encrypted everywhere

Every service is reached over HTTPS/TLS, with certificates that renew automatically.

Blue-team monitoring

Network and host-based detection, intrusion-detection sensors, real-time alerting and automated daily health reporting.

Nightly reset

Training systems return to a known-good baseline every night, so each class starts clean regardless of what happened the day before.

Project overview

Not a website deployment. A platform.

Secure front door
1
Federated identity
SSO
Every service, auto-renewed
TLS
Reset to a clean baseline
Nightly

ABOSS was engaged to build and run a complete cybersecurity training platform for a higher-education institution — an environment where students and staff practise real offensive and defensive techniques safely, from any browser, with nothing to install.

The result is a production cyber range: a fleet of purpose-built virtual machines — attacker workstations, vulnerable web applications, a Windows Active Directory environment, malware-analysis tooling and full blue-team monitoring — consolidated onto controlled, owned infrastructure and presented behind a single secure web address.

It combined application engineering, DevOps, infrastructure design, network security, identity, monitoring and day-to-day operations into one dependable system that is live and in daily use for teaching.

The challenge

Many vulnerable systems. One campus network. Zero leakage.

A hands-on security curriculum needs something ordinary IT can't easily provide: many deliberately vulnerable systems running side by side, safely isolated, and reachable by a whole class at once — without turning the campus network into a target.

Complexity at scale

Dozens of specialised systems — attack tools, vulnerable targets, monitoring stacks — each with different access methods (SSH, RDP, web) had to coexist and stay manageable.

Safe isolation

Intentionally vulnerable machines and live attack traffic must never leak onto the wider network. Segmentation was a hard requirement, not a nice-to-have.

Frictionless, secure access

Students needed everything from any device without installing clients or juggling separate passwords — while keeping strong authentication.

Operational reliability

A lab that breaks mid-class is worthless. Damaged machines had to recover automatically, and staff needed clear visibility into platform health.

Trustworthy exposure

A single, encrypted, professionally branded front door — not a sprawl of raw services on random addresses.

Cost control

Enterprise-grade range capability without an enterprise-grade cloud bill.

Architecture

One secure door in. Strong identity. Isolated services. Eyes on everything.

The platform is built in clean layers, deliberately understandable by business owners, CTOs and engineers alike. Internal addressing, ports and configuration are deliberately omitted.

  1. 01

    01 · Edge

    Users

    Students & staff

    Any device, any browser. Nothing to install.

  2. 02

    02 · Secure web layer

    Hardened reverse proxy

    The single front door

    Terminates HTTPS/TLS, presents one friendly address and routes each request to the right internal service. Nothing internal is exposed directly.

  3. 03

    03 · Identity

    Single sign-on broker

    One login, short-lived sessions

    Federates the institution's existing accounts. Users authenticate once; each connected application receives a short-lived session. No per-app passwords to manage.

  4. 04

    04 · Application services

    Containerised apps

    Isolated services

    The browser-based remote-access gateway, the challenge and scoring platform, and the training web applications — each an isolated container.

  5. 05

    05 · Lab targets

    Segmented internal networks

    Reachable only through the gateway

    Attacker workstations, vulnerable hosts, a Windows Active Directory environment and analysis tooling on segmented internal networks.

Application engineering

Every component maps to a teaching function.

The platform brings together several classes of application, each deployed as an isolated, containerised service, integrated with single sign-on where appropriate and placed behind the secure front door. Nothing was added for effect.

Browser-based remote-access gateway

Full terminal and remote-desktop sessions to lab machines, inside the browser. The single path through which learners reach every operating-system target.

Challenge & scoring platform

Security challenges with flag submission, scoring and a live leaderboard for assessments and competitions.

Vulnerable training web applications

A curated set of deliberately insecure web apps for teaching web exploitation — including a per-learner variant so each student gets an independent instance and a clean scoreboard.

Offensive & defensive systems

Attacker workstations, vulnerable hosts, a Windows Active Directory environment, malware-analysis tooling and full network and host monitoring stacks.

Cybersecurity

Security was designed in, not bolted on.

Infrastructure

  • Network segmentation isolates vulnerable systems and attack traffic from the wider environment.
  • Least-exposure model — internal systems are never published directly.
  • Administrative access is controlled and kept off the public path.

Application

  • Every service is reached over HTTPS/TLS with automatically renewing certificates.
  • A hardened reverse proxy centralises and standardises secure access.
  • The public surface is a single, well-understood entry point.

Access

  • Single sign-on federates the existing identity provider — strong authentication enforced centrally.
  • Short-lived, session-based access to each application instead of shared standing credentials.
  • Time- and class-based controls limit who can use the lab, and when.

Detection & containment

  • Network security monitoring and host-based detection observe lab activity for threats.
  • Dedicated intrusion-detection sensors support signature-based exercises and real visibility.
  • Because the lab deliberately contains vulnerable machines, containment is a first-class control.

DevOps & operations

Repeatable and recoverable — not hand-built and fragile.

The environment runs day after day with minimal babysitting. The emphasis throughout is on outcomes: a lab that is dependable, observable and self-healing.

Version-controlled configuration

Infrastructure and service configuration tracked in Git — a clear history and a path to review every change.

Containerised services

Applications run as containers for consistent, portable, repeatable deployment and clean upgrades.

Automated recovery

Training systems reset to a known-good baseline nightly. Staff can restore an individual system on demand from a snapshot.

Scheduled operational jobs

Routine health reporting and maintenance run on automated schedules without human intervention.

Daily health reporting

An automated summary of system and service state, flagging anything that needs attention.

Real-time alerting

Operators are notified of significant events as they happen, backed by continuous network and host monitoring.

These patterns extend into full CI/CD pipelines — automated build-and-deploy via GitHub Actions — for teams that want every change to ship through an automated, tested workflow.

Security testing & validation

Validated before it was trusted for teaching.

  1. Assessment
  2. Findings
  3. Prioritisation
  4. Remediation
  5. Validation
  6. Improved posture
  • End-to-end verification confirmed every service was reachable only through the intended secure path.
  • External-exposure checks confirmed internal systems were not publicly reachable.
  • Access and authentication flows were validated through the identity provider.
  • Issues found during bring-up were remediated and re-verified before go-live.

We do not publish exploit paths, findings evidence or sensitive configuration. Validation is described by outcome, not by method.

Project journey

Discover to operate — the full engineering lifecycle.

01

Discover

Translated the security curriculum into a concrete set of systems and access needs.

02

Design

Planned a segmented architecture with one secure front door and centralised identity.

03

Build

Provisioned the virtual fleet and deployed containerised services.

04

Secure

TLS everywhere, single sign-on, and network isolation for vulnerable systems.

05

Automate

Built the nightly reset and scheduled health reporting.

06

Test

Validated access paths and confirmed nothing internal was publicly exposed.

07

Optimise

Right-sized infrastructure and consolidated onto owned compute.

08

Deploy

Cut over to the production domain and professional brand.

09

Operate

In daily use with monitoring, alerting and self-recovery.

Before vs after

From a risky sprawl to one operated platform.

Before
After
  • A curriculum needing many risky, hard-to-manage systems
    One consolidated, professionally operated platform
  • Manual, install-heavy access
    Browser-based access, nothing to install
  • Separate logins per system
    Single sign-on through the existing identity provider
  • Vulnerable systems near the wider network
    Segmented, isolated, contained lab environment
  • Services exposed ad hoc
    One encrypted, hardened front door
  • Manual recovery after damage
    Automated nightly reset to a clean baseline
  • Limited operational visibility
    Continuous monitoring, alerting and daily reporting
  • Uncontrolled infrastructure cost
    Right-sized, cost-efficient owned infrastructure

Results & impact

What changed.

Stronger security posture

Least-exposure design, encrypted access everywhere, centralised authentication.

Reduced attack surface

A single controlled entry point; internal systems never directly exposed.

Operational reliability

Self-healing training systems and continuous visibility keep classes running.

Frictionless access

Any device, any browser, one login.

Repeatability

Containerised, version-controlled, automated operations.

Cost efficiency

Enterprise-grade capability on right-sized, owned infrastructure — rather than renting equivalent always-on cloud capacity.

Production-ready

Live, validated, branded, and in daily use.

Impact is stated as capability and outcome. No unsupported numbers are claimed.

Technology stack

What it runs on.

Only technologies actually used on the project are presented as delivered. Cloud breadth and AI serving are transferable or extension capabilities.

Infrastructure
Linux virtualisationNginx reverse proxySegmented internal networking
Application & services
Docker containersBrowser-based remote-access gatewayChallenge / scoring platformTraining web applications
DevOps
Git-based configurationContainer-based deploymentScheduled automationCI/CD-ready (GitHub Actions)
Security
HTTPS/TLS, auto-renewedSingle sign-on / federated identityNetwork segmentationNetwork & host intrusion detectionSIEM / monitoringFirewall controls
CloudTransferable capability
AWSGCPDigitalOceanLinode
AI infrastructureExtension capability
Self-hosted model servingGPU-based inference
Roadmap option

Extend it with self-hosted AI infrastructure.

Environments like this can be extended with AI/LLM models running on the institution's own compute — for cost efficiency, data control and independence from external APIs. Controlled model serving, self-hosted inference and GPU-efficient deployment are offered as a roadmap option, not as work delivered on this project.

Build with us

What we build for institutions and enterprises.

The same architecture — secure front door, federated identity, segmented services, monitoring and automated operations — applies to training platforms and production application environments alike.

Cybersecurity labs & cyber ranges
Secure application development
Cybersecurity infrastructure & network segmentation
DevOps implementation & CI/CD automation
Cloud deployment & server hardening
Infrastructure modernisation & cost optimisation
Monitoring, alerting & operational automation
Self-hosted AI & model-serving infrastructure
We don't just build applications. We build the infrastructure, security, automation and operational foundation required to run them reliably in production.

Have a platform that needs to be built, secured, deployed and operated properly?

Let's build the infrastructure behind it.

Download the one-pager (PDF)

We build to strong, honest security standards. We don't promise "100% secure" — no one credible does.

FAQ

Questions institutions ask first.

Yes. The vulnerable systems exist by design for training, and they are contained through network segmentation and a least-exposure architecture so that risk stays inside the range.