Capturing baseline network performance

Task Statement 1.4: Define logging and monitoring requirements across AWS and hybrid networks.

📘AWS Certified Advanced Networking – Specialty

---

When we talk about network performance, we mean how well a network is operating. This includes metrics like speed, reliability, latency (delays), packet loss, and throughput (amount of data transferred over time). Capturing a baseline means measuring these performance metrics under normal, healthy conditions. This baseline acts as a reference point to detect anomalies, troubleshoot issues, or plan for scaling.

Think of it as understanding what “normal” looks like for your network before you try to spot problems or optimize it.

---

Why Capturing a Baseline Is Important

1. Detecting Anomalies:
   If your baseline shows that normally latency is 10 ms, but suddenly it jumps to 100 ms, you know something is wrong.
2. Performance Optimization:
   Helps identify bottlenecks. For example, you can see if a VPN connection to your AWS VPC is slower than expected.
3. Capacity Planning:
   Knowing the baseline allows you to plan for future growth. If average throughput is 500 Mbps and peak traffic is 700 Mbps, you know when to scale.
4. Hybrid Networks:
   In environments connecting on-premises infrastructure to AWS (hybrid cloud), baselines help compare performance across different segments (like your data center vs AWS Direct Connect).

---

Key Metrics to Capture

For AWS and hybrid networks, the following metrics are important:

Metric	Description	AWS Tools
Latency	Time it takes for a packet to travel from source to destination	VPC Flow Logs, CloudWatch metrics
Throughput	Amount of data transmitted per second	CloudWatch, AWS Direct Connect monitoring
Packet Loss	Percentage of packets dropped	VPC Reachability Analyzer, CloudWatch
Jitter	Variation in latency	CloudWatch, third-party monitoring tools
Connection Errors	Failed TCP/UDP connections	CloudWatch Logs, VPC Flow Logs

---

AWS Tools for Capturing Baseline

AWS provides multiple tools for measuring and monitoring network performance:

1. CloudWatch Metrics and Logs
   • Collects performance data for AWS resources such as VPCs, EC2 instances, and Load Balancers.
   • Metrics like NetworkIn, NetworkOut, and StatusCheckFailed help understand normal traffic levels.
2. VPC Flow Logs
   • Captures information about IP traffic going in and out of network interfaces.
   • Helps measure traffic patterns, latency, and abnormal drops.
3. AWS CloudTrail
   • Tracks API calls and control plane activity.
   • Useful to correlate network events with administrative changes.
4. AWS Network Performance Tools
   • VPC Reachability Analyzer: Confirms connectivity between AWS resources.
   • AWS Direct Connect Monitoring: Tracks hybrid connectivity performance.
5. Third-Party Monitoring
   • Tools like Datadog, Zabbix, or SolarWinds can capture baseline metrics from both AWS and on-premises networks.

---

Steps to Capture Baseline Performance

1. Define Scope
   • Decide which networks, regions, and resources to monitor (VPCs, VPNs, Direct Connect links, etc.).
2. Select Metrics
   • Choose metrics like latency, throughput, packet loss, jitter, and errors.
3. Collect Data Over Time
   • Measure performance under normal workloads (ideally over 1-2 weeks).
   • Include different times of day to account for peak and off-peak usage.
4. Analyze and Document
   • Calculate averages, peaks, and trends.
   • Store this baseline for reference.
5. Set Alerts and Thresholds
   • Compare future metrics against baseline using CloudWatch Alarms.
   • Example: Trigger an alarm if latency exceeds 150% of baseline.

---

Best Practices for Baseline Network Performance

• Baseline Both AWS and On-Premises
  • For hybrid environments, monitor end-to-end performance.
• Repeat Baselines After Changes
  • After deploying new resources or changing network configurations, capture a new baseline.
• Use Visualizations
  • Use CloudWatch Dashboards or third-party tools to easily spot trends.
• Correlate with Application Performance
  • Baseline networking alone isn’t enough; ensure that applications hosted in AWS also meet performance expectations.

---

Exam Tips

For the exam, remember:

1. Definition: Capturing baseline means measuring network performance under normal conditions.
2. Purpose: Detect anomalies, plan capacity, optimize performance, troubleshoot issues.
3. Key Metrics: Latency, throughput, packet loss, jitter, connection errors.
4. AWS Tools: CloudWatch, VPC Flow Logs, CloudTrail, Direct Connect Monitoring, Reachability Analyzer.
5. Hybrid Networks: Baseline applies to both AWS and on-prem networks.
6. Best Practices: Repeat baselines after changes, use dashboards, set alarms for deviations.

---

Capturing a baseline is essentially knowing what “normal” looks like, so you can detect what isn’t normal. For AWS exams, they want you to understand what metrics to capture, which tools to use, and how to maintain baselines across hybrid environments.