Optimize Networks Based on Performance and Monitoring Metrics

Network performance optimization involves analyzing cloud network monitoring metrics and adjusting architecture configurations to improve system response speed, throughput capacity, and stability. Key measures include the following:

• Redeploy application components across different regions to reduce cross-region communication latency.

• Optimize data stream paths to reduce unnecessary cross-network domain data transmission.

• Optimize network service access patterns. Use private network access modes, such as PrivateLink, instead of public network access to improve transmission stability.

• Configure network service levels and scale appropriately to avoid resource bottlenecks.

• Adopt more efficient network resource types to replace less efficient solutions.

Priority

High

Practices to Avoid

• Failing to promptly address bandwidth overages causes public network packet loss.

• Expanding new business without addressing overseas user access to domestic services impacts overseas user experience.

• Underutilizing monitoring data—collecting extensive monitoring data without effective analysis or action—leads to recurring issues.

Expected Outcome

Establish a closed-loop network performance optimization process based on monitoring metrics. Continuously improve the network transmission efficiency and stability of business systems.

Implementation Guide

1. Identify network performance bottlenecks. Based on network anomalies and issues reported by monitoring, use tools to analyze network links and pinpoint bottlenecks. Combine Cloud Monitor data to analyze root causes of network performance issues. Determine whether the issue stems from network product configuration, architecture design, or business model. Focus on abnormal metrics such as high latency, high packet loss rate, and high retransmission rate.

2. Optimize network architecture design, product configuration, and product billing model selection. Typical optimization items include the following:

   • Increase public bandwidth or switch to a pay-as-you-go model to enhance public network access capabilities.

   • Adjust Server Load Balancer (SLB) load balancing policies to optimize access layer capabilities. Select appropriate load balancing algorithms, such as round-robin, weighted round-robin, or least connections, based on business characteristics. Fine-tune session persistence time. For API Gateway scenarios, enable HTTP/2 and QUIC protocols to improve transmission efficiency. Regularly check the health status of SLB backend servers. Promptly remove abnormal nodes to prevent performance degradation caused by uneven traffic distribution.

   • Choose suitable hybrid cloud connectivity solutions to optimize hybrid cloud access capabilities. Based on business bandwidth requirements and latency sensitivity, select an Express Connect circuit, a VPN Gateway, or a third-party SD-WAN solution. For high-bandwidth, low-latency scenarios, prioritize Express Connect over VPN connections to reduce encryption overhead and network hops. Regularly evaluate connection performance and adjust bandwidth configurations promptly.

   • Optimize server-side network parameter settings. Based on service traffic characteristics, adjust the network parameter values of ECS instances to improve network throughput capacity.

3. Optimize network architecture design. Re-evaluate application component deployment locations based on monitoring data. Deploy frequently interacting components within the same zone or region. Typical optimization items include the following:

   • Optimize zone selection to reduce cross-zone business latency. Based on in-region network performance data provided by NIS, optimize and adjust the zones for business deployment to reduce cross-zone business latency.

   • Use Cloud Enterprise Network (CEN) to optimize cross-region service invocation access. For applications deployed across regions, use CEN instead of traditional public network access to reduce network hops and latency. Use CEN's bandwidth plan feature to allocate cross-region traffic bandwidth as needed and avoid single-connection bandwidth bottlenecks. For critical business paths, enable CEN's Quality of Service (QoS) feature to ensure network service quality.

   • Use Global Accelerator (GA) to accelerate cross-region public network access. For cross-region, especially cross-border, public network access, use GA to optimize customer access experience, reduce latency, and improve public network access quality.

   • Use PrivateLink for secure and efficient service access. Migrate service invocations across VPCs or accounts from the public network to PrivateLink private network connections. Eliminate public network transmission uncertainties and latency fluctuations.

4. Implement network performance tuning. Use Alibaba Cloud network diagnostic tools to analyze network issues. Optimize the protocol stack. Based on business scenarios, select appropriate protocols and enable SSL offloading to improve business performance.

5. Establish a continuous optimization mechanism. Set network performance baselines. Regularly evaluate network architecture performance. Establish association analysis between monitoring metrics and business metrics. Translate network performance issues into quantifiable business impacts. Conduct A/B testing to verify network optimization effects. Ensure each adjustment delivers actual performance improvements.