« Television for Software Engineers | Main | Architecture, What Is It? »

Sunday, June 21, 2009



"Scale measures how much work the system can perform within a given set of performance parameters"

isn't this more capacity/efficiency than scale? i've always thought of scalability as what happens when i add *additional* resources (servers) to my system. if my system currently can handle 200 requests/sec with 100ms max response time, my system is said to be "scalable" if adding another server lets it handle 300 requests/sec with 100ms max response time. it is not scalable if adding another server keeps things at 200 request/sec, 100ms max response time.


I don't see any conflict here ;) "Scale" (or "throughput") is the measurement Mr. Pritchett is talking about, and "scalability" is the property (of being able to increase "scale") you described. It's interesting though how using different terms ("performance" for "latency" and "scale" for "throughput") puts a bit different twist on this classic trade-off.

Forex Advice

There is actually no conflict, it's all a matter of identifying the key factors and slowly analyzing them.. this is true to most economic landscape.

The comments to this entry are closed.