Movierama is a social movie sharing platform where users can share their favorite movies, like or hate movies, and explore submissions from other users.
- Users can sign up or log in.
- Registered users can add movies with a title, description, and date.
- Users can vote (like or hate) on movies submitted by others.
- Votes are limited to one per movie per user.
- Users can change or retract their votes.
- Users cannot vote for movies they submitted.
- Movie listings display:
- Title
- Description
- Name of the submitting user (clickable to filter by user)
- Date of publication
- Number of likes
- Number of hates
- Movies can be sorted by likes, hates, or publication date.
| Layer | Technology |
|---|---|
| Backend | Rust, axum, sqlx |
| Database | PostgreSQL |
| Frontend | Vue.js, Bun |
| Containerization | Docker, Docker Compose |
Movierama is fully containerized with Docker. The backend, frontend, and database run in separate containers on a shared network.
- Docker
- Docker Compose
docker compose up --builddocker compose downAfter starting, open http://localhost:5173.
Perfect — based on your new benchmark data and Docker resource metrics, here’s the updated and polished README section for your Performance Comparison, including insights from your latest results and your dashboard image (comparison_dashboard.png):
As part of the migration from the original Java (Spring Boot) backend to the new Rust (Axum) backend, I performed detailed load and resource usage tests using Grafana k6.
The tests simulated a consistent load of 70 requests per second for both backends over the same duration, capturing latency metrics and Docker resource usage (CPU and memory).
| Metric | Rust (Axum) | Spring Boot (Java) | Notes |
|---|---|---|---|
| Average latency (avg) | ≈ 1.7 ms | ≈ 6.2 ms | Rust handled requests ~3.5× faster on average |
| 90th percentile (p90) | ≈ 4.0 ms | ≈ 9.6 ms | 90% of Rust responses completed under 4 ms |
| 99th percentile (p99) | ≈ 7.2 ms | ≈ 17.6 ms | Rust remains low-latency even at tail load |
| Requests per second (req/s) | 70 req/s | 70 req/s | Identical test load for both |
- 🚀 Rust (Axum) delivers much lower latency and tighter consistency, thanks to its zero-cost abstractions and async runtime efficiency.
- ☕ Spring Boot, while robust and feature-rich, incurs overhead from the JVM, garbage collection, and heavier threading, resulting in higher average and tail latencies.
- ⚙️ CPU usage: Rust’s CPU footprint was only a fraction of Spring’s, remaining under 5 % even during steady load, whereas Spring Boot frequently spiked above 100 %.
- 🧠 Memory usage: Rust consistently stayed around 5 MiB, while Spring hovered between 300–360 MiB, mostly due to JVM overhead.
These results demonstrate that Rust’s Axum-based backend is dramatically more efficient, making it ideal for high-performance, resource-constrained, or latency-sensitive systems.
This dashboard summarizes:
- Top: Latency comparison (avg, p90, p99, req/s)
- Middle: CPU usage over time
- Bottom: Memory usage over time