RESILIENT — High Availability & Load Balancing Infrastructure
High-availability web infrastructure using redundant load balancers, VRRP virtual IP failover and multiple Nginx backend servers.
About this project
# Project Overview
RESILIENT is a practical infrastructure engineering project focused on designing, implementing and testing a highly available web-serving environment.
The architecture uses multiple load balancers, a shared Virtual IP (VIP) and multiple backend web servers to reduce single points of failure and improve service continuity.
Its central objective is to demonstrate how infrastructure can continue delivering services despite individual component failures.
# 1. Infrastructure Architecture
RESILIENT uses a two-layer high-availability architecture.
| Component | Technology | Responsibility |
|---|---|---|
| HAP1 | HAProxy + Keepalived | Primary load balancer |
| HAP2 | HAProxy + Keepalived | Backup load balancer |
| Shared VIP | VRRP | Consistent client access address |
| Web-01 | Nginx | Backend web server |
| Web-02 | Nginx | Backend web server |
| Web-03 | Nginx | Backend web server |
**Shared Virtual IP:** `192.168.50.22`
Two load balancers provide primary/backup redundancy, while three Nginx servers provide backend web services.
# 2. Core Technologies
- **HAProxy:** Distributes incoming HTTP traffic across backend servers.
- **Keepalived:** Implements VRRP-based virtual IP failover.
- **VRRP:** Provides primary/backup election and a shared access address.
- **Nginx:** Hosts backend web services.
- **Linux:** Provides the operating environment for infrastructure services.
# 3. Features and Implementation
## Load Balancing
HAProxy distributes incoming requests across three Nginx backend servers using round-robin scheduling.
## Backend Health Management
Backend health checks allow HAProxy to identify unavailable web servers and remove failed servers from the active pool.
## Backend Recovery
Recovered backend servers can rejoin the load-balancing pool when they become healthy again.
## Load Balancer Redundancy
HAP1 and HAP2 provide primary/backup load-balancer capability.
## Automatic VIP Failover
Keepalived uses VRRP to transfer the shared VIP to HAP2 when HAP1 fails.
## Failback
When the primary load balancer returns, the environment can return to its preferred primary state.
The project demonstrates two distinct resilience mechanisms:
- **Application-layer resilience:** HAProxy manages backend availability and traffic distribution.
- **Infrastructure-layer resilience:** Keepalived and VRRP provide load-balancer failover.
# 4. Testing and Verification
| Test Scenario | Demonstrated Result |
|---|---|
| Normal operation | Traffic distributed across backend servers |
| Web server failure | Traffic continues through healthy backends |
| Web server recovery | Recovered backend returns to service |
| Primary load balancer failure | VIP transfers to HAP2 |
| Primary load balancer recovery | Failback demonstrated |
## Outstanding Verification
During testing, HAP2's HAProxy service required manual startup. Consequently, automatic HAProxy service startup and fully unattended recovery on the backup load balancer were not conclusively verified.
This limitation is documented to distinguish demonstrated failover behaviour from complete autonomous recovery.
# 5. Engineering Skills Demonstrated
- Linux server administration
- High-availability architecture
- HAProxy load-balancer configuration
- Keepalived and VRRP
- Virtual IP addressing
- Reverse proxy configuration
- Backend health checks
- Fault tolerance and failure testing
- Infrastructure troubleshooting
- Service continuity and recovery
# 6. Project Outcome
RESILIENT demonstrates practical experience designing and testing redundant web infrastructure.
It combines traffic distribution, backend health management and load-balancer failover to illustrate how multiple infrastructure components can work together to improve service availability.
The project also highlights the importance of validating recovery behaviour rather than assuming that configured redundancy guarantees fully autonomous service recovery.
**Project Status:** Completed
**Category:** Infrastructure Engineering
**Technologies:** Linux, HAProxy, Keepalived, VRRP, Nginx