SREcon24 Europe/Middle East/Africa

Lerna Ekmekcioglu, Clockwork Systems


Is the app slow or the network lagging? When it comes to latency in distributed systems, it can be hard to identify where exactly the issue is. As businesses increasingly adopt diverse deployment environments —on-premises, cloud, or hybrid— the complexity grows, obscuring visibility into system health. Join me to hear why clock synchronization is key for identifying the true culprit when latency is due to contention in the network. I’ll demo how network contention impacts tail latencies followed by an overview of clock synchronization protocols to date, their pros and cons, and best practices in disciplining clocks, as well as recent algorithms from Stanford Research. With high precision clock synchronization at scale, we gain back visibility into useful one way delay metrics, which act as an early signal for network congestion that help us prevent impact to response times for our end users!


https://www.usenix.org/conference/srecon24emea/presentation/ekmekcioglu