Google Cloud CDN

Google Cloud CDN is Google Cloud’s web acceleration layer that caches and serves HTTP(S) content from Google’s global edge, bringing assets like HTML, CSS, JavaScript, images, and APIs closer to end users to reduce latency and origin load. Architecturally, Cloud CDN sits behind the global external Application Load Balancer, which terminates client connections at the nearest Google Front End (GFE) and routes traffic over Google’s private backbone to your origin if a cache miss occurs. When enabled on a backend service or bucket, Cloud CDN checks its cache and serves a cache hit directly from the edge; otherwise, it forwards the request to the origin and stores the response based on cache policy. Coverage leverages the same edge infrastructure Google uses for its own properties, with caches operating in 100+ metropolitan areas and edge POPs connected to hundreds of interconnection facilities and internet exchanges. This footprint, coupled with Global Anycast addressing, ensures users are routed to the closest edge for consistent, low‑latency delivery—often within one ISP hop—while origin fetches traverse Google’s high‑capacity fiber backbone. Cloud CDN supports modern client protocols out of the box. HTTP/3 over QUIC provides multiplexed, connection‑resilient transport that improves page rendering and video start times, with Google reporting measurable latency and rebuffering gains during internal rollouts; TLS 1.3 is also supported for faster handshakes and stronger security. Enabling HTTP/3 is a one‑click setting on HTTPS load balancers, and non‑supporting clients gracefully negotiate down to HTTP/2 or HTTP/1.1. Caching behavior is highly configurable. You can choose among cache modes such as CACHE_ALL_STATIC (automatic caching of common static types), USE_ORIGIN_HEADERS (respect origin cache directives), or FORCE_CACHE_ALL (override origin directives, suitable only for public content). Custom cache keys allow you to ignore protocol, host, or query strings to improve hit ratio, and features like negative caching let you cache specific error codes (e.g., 404) for short TTLs to shield origins. Byte‑range requests and partial hits are supported for large assets. Access control and security are integrated at the edge. Signed URLs and signed cookies provide time‑limited access to private content without requiring user accounts, while Cloud Armor adds WAF rules, geo/IP controls, rate limiting, and DDoS mitigation tied directly into the load balancer. For some backends (e.g., Cloud Storage buckets), edge policies focus on IP/geo enforcement; for compute origins, you can apply full L7 WAF policies. Operations and observability are first‑class. Cloud Logging and Cloud Monitoring expose request logs, optional TLS fields, cache status, and sampled metrics across backend services, with configurable sampling rates and dashboards to track hit ratios, errors, and performance. Fast cache invalidation—by path patterns, host, or cache tags—propagates globally in ~10 seconds, supporting agile release workflows and emergency purges.

Google Cloud CDN’s most distinctive advantage is deep integration with Google’s global network and load‑balancing stack. Requests terminate at a nearby Google Front End and traverse Google’s private fiber backbone to reach origins, minimizing public Internet hops and variability. This architecture consistently improves latency and throughput versus CDNs that rely more heavily on public transit, especially for long‑haul routes or congested peering paths. The platform’s Global Anycast and wide edge POP footprint—spanning 100+ metro areas and hundreds of interconnection facilities—provides predictable proximity to users worldwide. Anycast reduces DNS‑propagation delays and simplifies operations by presenting a single global IP, while Google’s extensive peering ensures “one‑hop” reach into most ISPs. In practice, this combination yields faster initial connection, lower TTFB, and resilient failover under large traffic spikes.

• Small Businesses (1-99)
• Mid Market (100-500)
• Enterprise (500+)