CDN Multi-CDN Strategy & Failover 4 — Questions and Answers
Question 1: In a multi-CDN deployment, which HTTP header should the origin always include to enable downstream CDN nodes to cache responses correctly and consistently?
- X-Cache-Status
- Cache-Control with explicit directives (max-age, s-maxage) (Correct answer)
- Vary: User-Agent
- X-Forwarded-For
Correct answer: Cache-Control with explicit directives (max-age, s-maxage)
Cache-Control headers with explicit s-maxage or max-age directives tell CDN edge nodes precisely how long to cache responses, ensuring consistent behavior across all providers.
Question 2: A multi-CDN operator notices that failover to CDN-B causes a surge in origin traffic. What is the most likely cause?
- CDN-B has a higher base cost per GB than CDN-A
- CDN-B's cache is cold because it receives little traffic in normal active-passive operation (Correct answer)
- CDN-B is configured with a stricter firewall that blocks cache fill requests
- The origin server does not support multiple CDN IP ranges
Correct answer: CDN-B's cache is cold because it receives little traffic in normal active-passive operation
In active-passive setups, the standby CDN receives minimal traffic and therefore has an empty or poorly populated cache, causing cache misses and origin surges during failover.
Question 3: Which approach allows a multi-CDN control plane to make routing decisions based on both performance AND cost simultaneously?
- Always route to the lowest-cost CDN regardless of performance
- Use a weighted scoring function that combines RUM latency data with CDN cost-per-GB pricing (Correct answer)
- Use a separate system for cost management and a separate system for performance routing
- Route all traffic to CDN-A until monthly spend reaches the budget cap, then switch to CDN-B
Correct answer: Use a weighted scoring function that combines RUM latency data with CDN cost-per-GB pricing
A weighted scoring function allows operators to define the trade-off between performance and cost, enabling nuanced traffic distribution that optimizes both dimensions.
Question 4: What is the role of BGP anycast in multi-CDN architecture compared to DNS-based steering?
- BGP anycast is used to route traffic to specific CDN providers; DNS steering routes to specific PoPs within one CDN
- BGP anycast allows network-layer routing to the nearest PoP within a single CDN; DNS steering directs traffic between CDN providers (Correct answer)
- BGP anycast requires manual configuration per user; DNS steering is automated
- BGP anycast only works for UDP traffic; DNS steering works for HTTP
Correct answer: BGP anycast allows network-layer routing to the nearest PoP within a single CDN; DNS steering directs traffic between CDN providers
BGP anycast operates at the network layer to find the nearest PoP within one provider's network, while DNS-based steering selects which CDN provider serves a given request.
Question 5: A content publisher updates a video file and triggers a purge across two CDN providers. CDN-A confirms purge success but CDN-B returns a 207 partial-success response. What is the safest next step?
- Accept the partial success and wait for TTL expiration on CDN-B
- Immediately change the video file URL to a new versioned path to bypass the stale CDN-B cache
- Disable CDN-B until the purge issue is investigated and resolved
- Re-issue the purge request to CDN-B for the failed objects and verify completion before serving traffic (Correct answer)
Correct answer: Re-issue the purge request to CDN-B for the failed objects and verify completion before serving traffic
Re-issuing the purge for failed objects and verifying completion ensures no stale content is served from CDN-B without disrupting normal service.
Question 6: Which protocol or standard enables a multi-CDN control plane to programmatically configure caching rules across different CDN providers through a unified interface?
- CDNI (CDN Interconnect) as defined by IETF RFC 8804 (Correct answer)
- OpenAPI 3.0 specification
- SNMP MIB-II extensions
- CORS preflight headers
Correct answer: CDNI (CDN Interconnect) as defined by IETF RFC 8804
The IETF CDNI framework provides standardized interfaces for CDN interconnection and inter-provider content delivery coordination.
Question 7: A retail site uses CDN-A for static assets and CDN-B for API acceleration. During Black Friday, CDN-B's dynamic caching is overwhelmed. Which multi-CDN design flaw does this expose?
- Using two CDNs instead of one provider is inherently unreliable
- The architecture lacks redundancy for the API acceleration tier; both CDNs should handle all traffic types (Correct answer)
- CDN-B is not sized correctly for peak traffic volumes
- Static and dynamic content should never be separated across CDN providers
Correct answer: The architecture lacks redundancy for the API acceleration tier; both CDNs should handle all traffic types
Assigning CDN responsibilities by content type creates single points of failure for each tier; both CDNs should be capable of serving all content types for true redundancy.
CDN Multi-CDN Strategy & Failover 4
In a multi-CDN deployment, which HTTP header should the origin always include to enable downstream CDN nodes to cache responses correctly and consistently?