Multi-CDN Strategy & Failover Flashcards

Which DNS-based technique is most commonly used to distribute traffic across multiple CDN providers in a multi-CDN setup?

Answer: GeoDNS with weighted records

GeoDNS with weighted records allows traffic to be directed to different CDN providers based on geography and configured weight percentages.

A company uses two CDNs and detects that CDN-A has a 40% cache hit rate while CDN-B has an 85% cache hit rate for the same content. What is the most likely root cause of the disparity?

Answer: CDN-A's TTL settings are configured too low

Excessively low TTL values cause content to expire quickly and force frequent origin fetches, dramatically reducing cache hit rates.

In a multi-CDN architecture, what is the primary purpose of a real-user monitoring (RUM) system?

Answer: To collect performance data from actual end-user sessions and inform CDN routing decisions

RUM collects latency, availability, and throughput data from real users to enable intelligent, performance-based CDN traffic steering.

Which failover trigger threshold is most appropriate for automated CDN switching in a production multi-CDN environment?

Answer: Switch after 3 consecutive failed health checks within a 60-second window

Requiring multiple consecutive failures within a time window prevents false-positive failovers caused by transient network blips.

A media company streams live video using two CDNs. During a failover event, viewers experience a 10-second buffering delay. Which architecture change would best minimize this delay?

Answer: Pre-warm the secondary CDN by pushing manifest and segment requests before failover

Pre-warming the secondary CDN ensures popular content is already cached, so when traffic shifts, the CDN can serve segments immediately without origin fetches.

What distinguishes active-active multi-CDN from active-passive multi-CDN?

Answer: Active-active sends live traffic to multiple CDNs simultaneously; active-passive keeps the secondary CDN idle until failover

In active-active mode, traffic is distributed across all CDNs continuously, while in active-passive mode, the secondary CDN only receives traffic when the primary fails.

A multi-CDN operator needs to ensure consistent cache purge behavior across all providers when content is updated. Which approach is most reliable?

Answer: Use a centralized purge API orchestration layer that calls each CDN provider's purge endpoint simultaneously

A centralized purge orchestration layer ensures that invalidation requests are sent to all CDN providers atomically when content changes.