Architectural Design of Compute and Storage Clouds
At the heart of every cloud data center are two massive subsystems:
- Compute Clouds: Systems designed to execute code, run calculations, and host virtual machines.
- Storage Clouds: Systems designed to store petabytes of data reliably without loss.
1. Architectural Design of Compute Clouds
A Compute Cloud takes thousands of physical multi-core servers and turns them into an elastic pool of virtual processing power.
Compute Cloud Architecture
Web Traffic / Users
➔
Load Balancer (HAProxy/ELB)
➔
Auto-Scaling VM Cluster (Node 1, 2, 3...)
Key Components of Compute Architecture:
- Load Balancer: A smart traffic controller that distributes incoming user requests evenly across multiple backend virtual machines so no single server gets overloaded.
- Hypervisor & VM Scheduler: Decides which physical server has enough free CPU and RAM to launch a new virtual machine.
- Auto-Scaling Engine: Monitors CPU utilization. If average CPU exceeds 80%, it triggers the creation of 2 new VMs automatically.
2. Architectural Design of Storage Clouds
Cloud storage is completely different from a regular computer’s hard drive. It is divided into three major architectural types:
Three Types of Cloud Storage Architecture
1. Block Storage
Raw disk volumes mounted directly to VMs (like an internal hard drive)
2. File Storage (NAS)
Shared network folders accessible by multiple servers (NFS / SMB)
3. Object Storage
Files stored with unique IDs & metadata; accessed via HTTP URLs (AWS S3)
Master Comparison: Block vs. File vs. Object Storage
| Feature | Block Storage | File Storage (NAS) | Object Storage |
|---|---|---|---|
| How Data is Managed | Fixed raw blocks (no file system until formatted) | Hierarchical folders, directories, and files | Flat structure of objects with unique IDs & metadata |
| Access Method | Direct SCSI / Fibre Channel bus connection | Network file sharing protocols (NFS, CIFS) | Web REST APIs (HTTP GET, PUT, DELETE) |
| Performance | Ultra-fast; lowest latency | Fast; moderate latency | Slower latency; massive throughput |
| Max Capacity | Gigabytes to Terabytes | Terabytes | Virtually infinite (Petabytes & Exabytes) |
| Cloud Example | AWS EBS (Elastic Block Store) | AWS EFS (Elastic File System) | AWS S3, Google Cloud Storage |
| Best For | Running databases (MySQL, Oracle) and boot OS drives | Shared company documents, team project files | Photos, videos, PDF backups, mobile app assets |
💡 DAVV Exam Tip
Question (10 Marks): “Explain the architectural design of Compute and Storage Clouds. Differentiate between Block, File, and Object Storage.”
- Draw the Compute Cloud flow diagram: Users -> Load Balancer -> VM Cluster.
- Write the 3 storage types with the comparison table.
- Mention AWS S3 as the prime example of Object Storage—examiners reward real cloud service names.