G'day
I read the Quantify article in the book “97 Things Every Software Architect Should Know” ( Amazon sanitized link ), and I wondered how to scale it quantitatively.
I developed two systems for a large British broadcasting corporation, which are used for:
- determine the country of origin for incoming HTTP requests or
- determine the appropriate video formats for the screen geometry of mobile phones and the current type of connection.
Both options should provide scalability.
My projects for both systems scale horizontally by caching the load balancing levels that are used to process incoming requests for both of these services and distribute them across several servers that actually provide the service. The initial increase in throughput is achieved by adding more servers behind the load balancing layer, therefore the term horizontal scalability.
There is a limitation on the scalability of this architecture, however, if the load balance level begins to cope with the traffic of the incoming request with difficulty.
So, can scalability be quantified? Will there be an estimate of the number of additional servers that you could add to scale the solution?