Scalability and the front-end layer

Updated 12 Oct 2026

•
Scalability, not just speed
- •
  Scalability -> adjust capacity up/down to meet demand while keeping cost and user experience acceptable.
- •
  Performance -> speed/resource use under a workload; scalability asks how these behave as workload/capacity changes.
- •
  Three dimensions
  - •
    Data -> growing dataset.
  - •
    Concurrency -> simultaneous clients.
  - •
    Interaction rate -> request frequency.
  - •
    each can stress different resources; user count alone is not a capacity metric.
- •
  Vertical scaling
  - •
    action: strengthen one node: RAM, CPU, I/O, network throughput.
  - •
    often simple initially, but physical/cost limits and shared-resource contention remain.
- •
  Horizontal scaling / cloning
  - •
    action: add interchangeable instances and distribute traffic.
  - •
    downstream capacity and distribution still matter; growth is not unlimited or necessarily linear.
- •
  Functional partitioning / service isolation -> split responsibilities for independent operation/scaling, e.g. catalog and user profile; not identical clones.
- •
  Data partitioning / sharding
  - •
    action: split data by partition key; distribute storage and potentially reads/writes.
  - •
    Costs -> routing, skew/hotspots, cross-shard queries, transaction complexity.
- •
  Read replicas
  - •
    benefit: extra copies distribute reads in a primary/replica design.
  - •
    primary writes and replication remain bottlenecks.
  - •
    async replicas can lag behind primary writes, not receive those writes first and later update the primary.

Linked references 1

web-development-polban

◌ Explore connections in Graph view

Scalability and the front-end layer

Scalability, not just speed

Linked references 1