servers ○ GPUs #IstioCon Application Deployment: Cloud Layout ● Region: A metro region ● DC: One or more Data Centers in each Region ● AZ: One or more Availability Zones in each DC ○ Independent globe peering with the Internet closer to the customer ○ PoPs are mini AZs Region R1 AZ 1 AZ 2 AZ n Data Center DC1 Region Rn #IstioCon Application Deployment: Cloud Layout ● Multiple K8s Clusters HA ○ In all regions ○ In multiple AZs in each region ○ Capability to run all applications from a single region or AZ in a worst-case scenario Region R1 AZ 1 AZ 2 AZ n Data Center DC1 K8s Cluster

实现原理：控制面根据 consumer 和 provider 的地域属性为provider实例配置 不同的LB优先级，优先级的判断顺序如下： 1. 最高：相同 region，相同zone 2. 其次：相同 region，不同zone 3. 再次：不同 region #IstioCon Aeraki Demo: 地域感知负载均衡(Dubbo) 1. 在 dubbo: application 配置中为 Provider aeraki_meata_locality 自定义属性 2. 在 provider 的 deployment 中通过环境变量设置其所属地域 3. 在 consumer 的 deployment 中通过 label 声明其所处的 region 和 zone 4. 通过 dr 规则启用 locality load balancing https://docs.qq.com/doc/DVnlqUVB1ek1laFBQ #IstioCon

Visibility ✓ Traffic Shaping ✓ Latency ✓ Single Point of Failure Adoption Challenges ● Multi-region deployments ● Non-flat networks ● Multi-tenant configuration ● Management of Istio installation cluster registration ● Higher Level Logical Service for Developers ○ Multi-cluster Identity ○ Multi-region Endpoint ○ Istio config integrated with gitops deployment ○ Init modifications to prevent proxy

egress bandwidth ~2 PB/month #IstioCon Architecture Overview ● User traffic infrastructure - TW region, all 3 zones ● REST APIs for client traffic ● gRPC for inter-service traffic ● Around 100+ microservices Capability for service authentication and authorisation ● Improved Observability ● Extendable to multi-region setup #IstioCon Approach #IstioCon Rollout - Istio setup and Microservices ● Split rollout in

Square DC -> Cash App EKS Internal Presentation New in-mesh s2s Internal Presentation New cross-region s2s Internal Presentation