When creating a stretch cluster the most common questions are usually, will this work with the latency between my sites and if so, what do I need to tune? In this session I’ll explain the most common levers we use to combat increased latency in stretch clusters. We will cover operating system level changes, broker side socket and buffer sizes, replication level tuning and touch on client optimizations. For each area I’ll dive into the three M’s. The Mentality (reason why we look at this), the Metric (what specific metric do we use to observe the impact of our changes and the Measure (what is the sweet spot we are looking to find for each optimization) . At the end of our trek, you’ll be ready to roll out clusters that are tuned to combat latency for any workload you may need to run.