CDN Multi-CDN Strategy & Failover 5 — Questions and Answers
Question 1: A financial services firm requires that its multi-CDN failover completes within 30 seconds of detecting a CDN failure. Which combination of settings best achieves this?
- DNS TTL of 300s + health check interval of 60s
- DNS TTL of 30s + health check interval of 10s with 3-failure threshold (Correct answer)
- DNS TTL of 60s + health check interval of 30s
- DNS TTL of 5s + health check interval of 60s
Correct answer: DNS TTL of 30s + health check interval of 10s with 3-failure threshold
A 30-second DNS TTL and 10-second health check interval with a 3-failure threshold allows failover detection and DNS propagation to complete within the 30-second SLA window.
Question 2: Which approach best prevents cache poisoning attacks when operating a multi-CDN environment with multiple origin-pull paths?
- Use different cache keys per CDN provider
- Validate all origin responses with TLS certificate pinning and enforce strict Cache-Control headers (Correct answer)
- Restrict origin pull to a single CDN at a time
- Use HTTP instead of HTTPS between CDN and origin to reduce handshake overhead
Correct answer: Validate all origin responses with TLS certificate pinning and enforce strict Cache-Control headers
TLS certificate pinning between CDN and origin verifies server identity, while strict Cache-Control headers prevent unauthorized content from being cached and served.
Question 3: A multi-CDN deployment uses JavaScript-based RUM to measure CDN performance. Which limitation must operators account for when using this approach?
- JavaScript RUM cannot measure HTTPS connections
- RUM data is skewed by users with ad blockers or script-blocking browser extensions who may not report metrics (Correct answer)
- JavaScript RUM requires a dedicated origin server to collect beacon data
- RUM measurements cannot distinguish between CDN-A and CDN-B performance
Correct answer: RUM data is skewed by users with ad blockers or script-blocking browser extensions who may not report metrics
Ad blockers and privacy tools often block RUM beacon requests, creating a self-selection bias where performance data underrepresents a subset of users.
Question 4: What is the purpose of 'last-resort' origin failover configuration in a multi-CDN architecture?
- To allow the origin to bypass CDN caching during low-traffic periods
- To ensure user requests reach the origin server directly if all CDN providers simultaneously fail (Correct answer)
- To route premium users directly to the origin for lower latency
- To enable A/B testing by sending a subset of users around the CDN
Correct answer: To ensure user requests reach the origin server directly if all CDN providers simultaneously fail
A last-resort origin failover path guarantees site availability even in catastrophic multi-CDN failure scenarios by allowing direct origin access.
Question 5: A multi-CDN operator wants to test failover procedures without impacting production users. Which technique allows this?
- Run failover tests only during off-peak hours on the live production system
- Use a canary traffic split to route a small percentage of synthetic monitoring traffic through the failover path
- Simulate failover in a staging environment with identical CDN configurations and traffic patterns mirrored from production (Correct answer)
- Disable health checks temporarily and observe natural failover behavior
Correct answer: Simulate failover in a staging environment with identical CDN configurations and traffic patterns mirrored from production
A staging environment with mirrored production CDN configurations allows realistic failover testing without risking real user sessions.
Question 6: Which CDN configuration error most frequently causes request loops in a multi-CDN setup where CDN-A pulls from CDN-B as its origin?
- Mismatched SSL certificate SANs between CDN-A and CDN-B
- CDN-B is configured to pull from CDN-A as its origin, creating a circular dependency (Correct answer)
- Both CDNs using the same DNS CNAME record
- CDN-A and CDN-B sharing the same cache key namespace
Correct answer: CDN-B is configured to pull from CDN-A as its origin, creating a circular dependency
A circular origin-pull configuration where CDN-A pulls from CDN-B and CDN-B pulls from CDN-A results in infinite request loops and total cache miss.
Question 7: An organization needs to attribute CDN cost accurately when using two providers. CDN-A bills per GB transferred and CDN-B bills per request. Which monitoring data is essential for cost forecasting?
- Only total user session count and average page size
- Per-CDN byte transfer volume AND per-CDN request count, broken down by traffic type and region (Correct answer)
- Total origin server egress bandwidth only
- DNS query count per CDN provider per month
Correct answer: Per-CDN byte transfer volume AND per-CDN request count, broken down by traffic type and region
Accurate cost forecasting requires per-CDN breakdowns of both bytes transferred and request counts, since each provider uses a different billing dimension.
CDN Multi-CDN Strategy & Failover 5
A financial services firm requires that its multi-CDN failover completes within 30 seconds of detecting a CDN failure.
Which combination of settings best achieves this?