Performance

This section presents how the architecture addresses performance related issues.

This section presents various design patterns used to achieve performance.

Basic performance metrics are latency and throughput. Latency is measured as the time elapsed between request and response and throughput as the number of requests handler per second. In an ideal world, the latency should not increase and throughput should scale linearly as the load increases.

Performance related issues needs to be investigated at various points of the architecture. Some common elements that should be subjected to performance tuning are:

• Middleware – Middleware technologies like ESB or other distributed technologies are primary candidates for performance tuning as issues like network round-trips and network latency could become critical.
• Database – Database access and processes like joins and sorts are candidates for performance tuning.
• Search Engine – Solr caching is an candidate to improve search performance by leveraging cached queries and results