Streaming & media
What Is Origin Shielding and How Does It Protect Your Infrastructure?
Learn how a shared caching tier reduces duplicate origin requests, how to measure the benefit, and where caching needs additional protection.
Add a shared cache before the origin
A CDN edge serves viewers near the audience. If an edge lacks an object, it needs to fetch it upstream. With many edges, the origin may receive several requests for the same video segment or image. An origin shield places another caching tier between those edges and the origin.
The request path becomes viewer → CDN edge → shield → origin. A shield hit answers without reaching the origin. Request coalescing can also combine concurrent misses for the same cache key, reducing duplicate work while a fresh object is fetched.
Measure origin offload with a clear denominator
Suppose an illustrative test sends 10,000 requests per second to a shield and only 200 reach the origin. The shield’s request offload is 1 − 200 / 10,000 = 98%. This is arithmetic for an example workload, not a measured Gloman result or a guarantee during an attack.
State whether the baseline is viewer requests, edge misses or origin traffic before shielding. Compare equal workloads over the same period. Request reduction, bandwidth reduction and CPU reduction are different metrics: avoiding a small image and avoiding an expensive dynamic render have different effects.
Cache policy determines the result
Repeated public objects are strong candidates for shielding. Personalised pages, authenticated APIs and requests with unique cache keys may still reach the origin. Include every response-varying dimension in the key and respect private or no-store responses.
Plan TTLs, purge propagation and any permitted stale delivery. Live playlists need freshness; completed media segments often permit a different policy. Test signed URLs and cookies so access checks cannot be bypassed by a shared cached response.
A shield is another dependency. Define capacity, health checks and failover behaviour, including whether a failure bypasses the shield and suddenly sends all misses to the origin.
Combine shielding with access and traffic controls
A popular cached object can be served without repeatedly invoking the application. An HTTP flood targeting uncached URLs can still consume origin resources. A volumetric attack against a directly reachable origin can bypass the caching path entirely.
Restrict origin access to approved delivery paths where the design permits it, and authenticate upstream requests. Combine cache rules with network filtering, rate limits and application protection. Test administrative access and monitoring before tightening origin rules.
Apply shielding to a measured workload
Gloman’s Web Acceleration & CDN service includes origin shielding and rate limiting, built around its caching infrastructure. For a publisher, start with the pages and assets that dominate peak traffic and review which can be shared safely.
Compare origin requests, response time, error rate and cache behaviour before and after deployment. Rehearse a purge, a shield failure and a cold-cache traffic spike. Those tests reveal whether the origin has enough capacity when the cache offers less protection.