A powerful new load-balancing algorithm only available in Buoyant Enterprise for Linkerd, HAZL provides significant cost-saving combined with high-availability for multi-zone Kubernetes clusters. HAZL dynamically adjusts traffic distribution to eliminate cross-zone traffic under healthy conditions, while preserving system reliability in the event of zone failure or system stress—even for for cross-cluster calls. HAZL provides significant reliability benefits over Kubernetes's Topology Hints approach.

In this technical white paper, you'll learn how HAZL works, how it compares to Topology Hints, including as implemented by open-source Linkerd and Istio. You'll learn HAZL's implementation and its advantages for reliability, cloud spend reduction, and cross-cluster communication.