Compute Requirements
The following table outlines the recommended server specifications for deploying the core components of the MLAngles LLMOps Platform. These specifications are optimized for stable performance, scalability, and efficient resource utilization across the platform's frontend, backend, database, pipeline processing, and VPN access layers.
Note: Instance types are shown in both AWS format and equivalent hardware specifications for on-premises or alternative cloud deployments.
| Server | OS Type | Instance Type (AWS) | Equivalent Specs (vCPU/RAM) | Root Volume (GB) | Secondary Volume (GB) |
| Frontend | Ubuntu 22.04 LTS | t3a.large | 2 vCPUs, 8 GB RAM | 8 | 10 |
| Backend | Ubuntu 22.04 LTS | r5a.xlarge | 4 vCPUs, 32 GB RAM | 15 | 50 |
| Database | Ubuntu 22.04 LTS | m6a.large | 2 vCPUs, 8 GB RAM | 10 | - |
| OpenVPN | Ubuntu 24.04 LTS | t3a.micro | 2 vCPUs, 1 GB RAM | 8 | - |
Hardware Specification Details:
- Frontend Server: Requires moderate compute resources (2 CPU cores, 8 GB memory) suitable for serving web interfaces and handling user interactions
- Backend Server: Memory-optimized configuration (4 CPU cores, 32 GB memory) to handle API requests, data processing, and model inference tasks
- Database Server: Balanced compute and memory (2 CPU cores, 8 GB memory) for efficient database operations
- VPN Server: Minimal resources (2 CPU cores, 1 GB memory) sufficient for secure network access management
Storage Notes:
- Root volumes host the operating system and application binaries
- Secondary volumes provide additional storage for logs, temporary files, and application data
- All storage volumes should use SSD or equivalent high-performance storage for optimal I/O performance