
Scalability is a system’s ability to keep throughput, response time and cost within acceptable bounds as demand grows, by adding resources or changing its structure. It does not mean the system is already large, and it is not simply a matter of installing more machines.
Suppose a library increases its loans desks from one to four. It appears able to serve more readers, but every desk still sends unusual loans to one staff member for approval. The queue merely moves to that shared point. Local capacity has grown while the whole flow has not. A scalable design identifies shared bottlenecks and allows data, work and decisions to be distributed instead of merely duplicating the entrance.
Scalability also differs from efficiency: low cost at a small scale does not show that performance will hold as demand rises. It is related to elasticity, but elasticity more specifically concerns acquiring and releasing resources as demand changes. The central point is that growth is not the old structure made larger; it tests which relationships stop working as the scale changes.
https://wa.aws.amazon.com/wellarchitected/2020-07-02T19-33-23/wat.concept.elasticity.en.html
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.