AWS Traffic Management: Latency, Geography, and Weighting Strategies

This study guide explores the mechanisms within AWS—primarily Amazon Route 53 and AWS Global Accelerator—used to influence and control how end-user traffic is routed to application endpoints. Mastery of these methods is critical for the ANS-C01 exam.

Learning Objectives

After studying this guide, you should be able to:

• Differentiate between various Route 53 routing policies (Weighted, Latency, Geolocation).
• Configure weighted records for canary deployments and blue/green migrations.
• Explain how AWS Global Accelerator improves performance compared to standard DNS-based routing.
• Evaluate the impact of health checks on traffic management decisions.
• Identify the limitations of latency-based routing regarding application-level delays.

Key Terms & Glossary

• Latency-Based Routing (LBR): A routing policy that directs traffic to the AWS region that provides the lowest network latency for the user.
• Weighted Routing: A method of distributing traffic across multiple resources in proportions that you specify (e.g., 70% to Region A, 30% to Region B).
• Geolocation Routing: Routing traffic based on the geographic location of your users (continent, country, or state).
• Geoproximity Routing: Routing traffic based on the geographic location of your resources and, optionally, shifting traffic from resources in one location to resources in another by specifying a "bias."
• Jitter: The variation in the delay of received packets, which can impact real-time application performance.
• Throughput: The actual amount of data transmitted over a network in a given time period.

The "Big Idea"

Traffic management in AWS is about moving beyond simple "one-name-to-one-IP" mapping. By using Intelligent DNS, we transform the network from a static path into a dynamic system that responds to user location, network health, and deployment logic. The goal is to maximize availability and performance while enabling controlled changes (like software updates) without downtime.

Formula / Concept Box

Hierarchical Outline

1. Route 53 Routing Policies
   • Weighted Routing
     • Assign numeric weights (0–255) to records.
     • Weight of 0 stops traffic to a resource (unless all are 0).
   • Latency-Based Routing (LBR)
     • Based on RTT measurements over time.
     • Note: Measures network latency, not application/database latency.
   • Geolocation Routing
     • Used for localized content or licensing restrictions.
     • Uses a "Default" record to handle unmatched locations.
2. AWS Global Accelerator
   • Uses Anycast IP addresses.
   • Ingresses traffic into the AWS global network as close to the user as possible.
   • Reduces latency by bypassing the public internet "middle mile."
3. Health Checks & Monitoring
   • Integrated with routing policies to bypass unhealthy endpoints.
   • Tools: CloudWatch Metrics, Network Insights, Reachability Analyzer.

Visual Anchors

Traffic Flow: Latency-Based Selection

Weighted Distribution Visualization

This diagram represents a 90/10 split for a canary deployment using a circle to represent the total traffic volume.

\begin{tikzpicture}[scale=2] \draw[thick] (0,0) circle (1cm); \fill[blue!20] (0,0) -- (0:1cm) arc (0:324:1cm) -- cycle; \fill[orange!60] (0,0) -- (324:1cm) arc (324:360:1cm) -- cycle; \node at (0.2, 0.4) {Production (90%)}; \node at (1.2, -0.2) {Canary (10%)}; \draw[->, thick] (1.1,-0.1) -- (0.8,-0.1); \end{tikzpicture}

Definition-Example Pairs

• Weighted Routing

  • Definition: A policy where you assign a relative weight to multiple records for the same name.
  • Example: You have two versions of a website. You assign weight 252 to Version A and weight 3 to Version B to send approximately 1% of users to the new version for testing.

• Geolocation Routing

  • Definition: A policy that serves different DNS responses based on the physical location of the requester.
  • Example: A streaming service that serves "The Office" to UK users but shows a "Not Available" page to US users due to licensing.

Worked Examples

Scenario: Configuring a 90/10 Canary Migration

Problem: You need to migrate traffic from an existing Application Load Balancer (ALB) to a new ALB version. You want to start with 10% of traffic on the new version.

Step-by-Step Breakdown:

1. Open Route 53 Console: Navigate to your hosted zone.
2. Create Record A: Select the existing ALB. Choose "Weighted" routing. Set Weight: 90 and ID: Old-ALB.
3. Create Record B: Select the new ALB. Choose "Weighted" routing. Set Weight: 10 and ID: New-ALB.
4. Verification: Route 53 will now sum the weights (90+10=100) and return the new ALB IP address approximately 10% of the time.
5. Scaling: As confidence increases, you change weights to 50/50, then finally 0/100.

Checkpoint Questions

1. Does Latency-based routing account for slow SQL queries in your backend database? (Answer: No, it only measures network latency to the AWS endpoint).
2. What happens in a Weighted routing policy if you set all record weights to 0? (Answer: Route 53 treats all records as having equal weight to avoid returning no records).
3. How does Global Accelerator differ from Route 53 Latency routing? (Answer: Global Accelerator uses the AWS private backbone for data transfer, whereas Route 53 only influences the initial DNS resolution).

Muddy Points & Cross-Refs

[!WARNING] LBR vs. Geolocation: Do not confuse these. A user in New York might have lower latency to a region in Europe than one in Oregon due to fiber paths, even if Oregon is geographically closer. Use Geolocation for compliance; use Latency for speed.

• Cross-Ref: For deeper troubleshooting of traffic paths, see Reachability Analyzer and VPC Flow Logs.
• Note on TTL: DNS-based traffic management is subject to TTL (Time To Live). If a user's local DNS resolver caches a record, changes to weights may not be immediate for that user.

Comparison Tables