Lesson 2: Scalability & Performance#

What is Scalability?#

Scalability is the ability of a system to handle increased load without performance degradation. It’s not just about “handling more users”; it’s about doing so cost-effectively and reliably.

Vertical Scaling (Scaling Up)#

Adding more power (CPU, RAM) to an existing server.

• Pros: Simple to implement.
• Cons: Hardware limits, single point of failure, expensive at the high end.

Horizontal Scaling (Scaling Out)#

Adding more servers to the pool of resources.

• Pros: Theoretically infinite scale, better fault tolerance.
• Cons: Increased complexity (load balancing, data consistency).

Performance Metrics#

Latency#

The time it takes for a system to process a request.

• Goal: Minimize latency (e.g., “API response < 100ms”).

Throughput#

The number of requests a system can handle per unit of time.

• Goal: Maximize throughput (e.g., “Handle 10,000 requests per second”).

🛠️ Sruja Perspective: Modeling Scalability#

In Sruja, you can represent horizontal scaling using the native scale block. This allows you to define minimum and maximum replicas and the scaling metric.

system ECommerce "E-Commerce System" {
    container WebServer "Web App" {
        technology "Go, Gin"
        
        // Define horizontal scaling properties
        scale {
            min 3
            max 10
            metric "cpu > 80%"
        }
    }
    
    container Database "Primary DB" {
        technology "PostgreSQL"
        // Vertical scaling example
        description "Running on a high-memory instance (AWS r5.2xlarge)."
    }

    WebServer -> Database "Reads/Writes"
}