William A.

I designed, developed, and deployed Voxa Center, an open-source VoIP telephony platform powered by Asterisk 20. This project leverages my skills in Linux server administration, network administration, telecom protocols (SIP/RTP), infrastructure security, and DevOps automation.

Linux Server Administration

I deployed the infrastructure on Debian 12 / Ubuntu Server using native installation: compiling Asterisk 20 from source, orchestrating all services via Systemd (Asterisk, PHP-FPM, Nginx, MariaDB, Redis), server hardening with Fail2ban, iptables/nftables, SRTP/TLS encryption, and Let's Encrypt certificates. Performance tuning: PHP-FPM, MariaDB (InnoDB), Redis. Centralized log management with journald and logrotate.

Network & Telecom Administration

I configured a production SIP/SRTP architecture with multi-operator interconnection (OVH, Twilio, Vonage): SIP trunks, G.711/G.722/Opus codecs, call routing. Deployed a WebRTC softphone with Nginx WebSocket proxy, ICE/STUN NAT traversal, and DTLS-SRTP. Developed an application-level SIP firewall with dynamic IP filtering, rate limiting, and Fail2ban synchronization. Proficient in TCP/UDP flows, SIP/RTP ports, DNS, and voice QoS.

DevOps & Automation

I wrote an automated Bash installer (curl | bash) that provisions a bare-metal server in 5 minutes: compilation, configuration, SSL certificates, and systemd services. Fully reproducible and version-controlled infrastructure on GitHub. Integrated monitoring dashboard with live supervision, CDR logs, and statistics. Embedded update mechanism with database migration and rollback.

A client hosting multiple websites (WordPress, PrestaShop) on a shared infrastructure was experiencing critical issues: high TTFB (> 800 ms) impacting SEO and user experience, performance degradation under load, lack of isolation between sites ("noisy neighbor" effect), and complex maintenance due to a single PHP stack.

As a systems and network administrator, the goal was to stabilize the infrastructure while significantly improving performance.

The solution involved deployingOrbiterto containerize each website, isolate PHP-FPM processes, and allocate dedicated resources (CPU, RAM). Concurrently, server and network optimization was implemented: Nginx HTTP caching, optimized reverse proxy configuration, reduction of backend calls, and PHP performance tuning.

The environment was administered on Linux Debian in production, with multi-version PHP management (7.4 to 8.x), resource monitoring (CPU, RAM, I/O), and system parameter adjustments (workers, buffers, ulimit).

The results were significant: TTFB reduction from ~800 ms to 120–180 ms (up to -80%), improvement in loading times of 40% to 60%, ability to handle +70% more simultaneous traffic without degradation, complete site isolation (no cross-incidents), and a 30% reduction in technical intervention time.

This project directly improved SEO (Core Web Vitals), provided a better user experience, and made the infrastructure more resilient and scalable, while reducing operating costs.

It exemplifies a comprehensive systems and network administrator intervention, combining system optimization, network performance, and infrastructure modernization with concrete and measurable results.

Complete IT management of an AWS cloud infrastructure with over 20 instances. Implementation of custom automation, containerized backup solutions, and real-time monitoring.

🖥️ AWS Infrastructure

Administration of 20+ production EC2 instances

S3 storage management: retention policies, lifecycle rules, encryption

RDS database administration: maintenance, tuning, replication

IAM rights management, security groups, and VPC network architecture

⚙️ Bash & Python Automation

Development of Bash scripts for recurring tasks: log rotation, cleanup, health checks

Development of Python scripts for orchestration: provisioning, EC2 snapshot management, S3 synchronization

Scheduling and orchestration via cron and systemd timers

Documentation and versioning of all scripts

🐳 Containerized Backup with Docker

Design of an automated containerized backup solution

Dedicated Docker containers: RDS database backup, EC2 snapshots, S3 export

Backup orchestration with dependency management and retries in case of failure

Automatic backup integrity verification

📊 Grafana + Prometheus Monitoring

Deployment of a containerized Prometheus + Grafana stack

Exporters on each EC2 instance: CPU, RAM, disk, network metrics

Custom dashboards per service: EC2, RDS, S3, backups

Real-time alerting: Slack/email for anomalies or backup failures

✅ Results:

20+ instances supervised in real-time. 100% automated and daily verified backups. Reduced incident detection time thanks to Prometheus alerts. Time savings of several hours per week on recurring administration tasks.

Technologies: AWS (EC2, S3, RDS, IAM, VPC), Linux, Bash, Python, Docker, Prometheus, Grafana, automation, backup, monitoring