Why CDNs Struggle to Accelerate Mobile Clients – Key Insights

Google performance engineer Ilya Grigorik recently published a blog titled “Why CDN Acceleration ‘Doesn’t’ Work for Mobile Clients,” describing the peculiarities of mobile (wireless) networks and proposing concepts for a mobile‑friendly CDN.

Ilya first criticizes current CDNs for their weak performance on mobile clients. Monitoring data shows that traditional CDN optimizations have minimal impact, prompting a call for a CDN architecture that better supports mobile networks.

He identifies two common misconceptions: (1) traditional CDNs offer similar absolute performance gains for mobile and broadband users; (2) the issue isn’t about a “wireless CDN” but rather the carrier network itself.

Using reference data, he breaks down wireless network latency components. Assuming a client on the West Coast and a server on the East Coast, the coast‑to‑coast propagation delay is 50 ms, and server response adds another 50 ms. Last‑mile latency varies:

Fiber: ~18 ms

Cable: ~26 ms

DSL: ~44 ms

4G: ~50 ms

3G: ~200 ms

He explains how a CDN works: by deploying cache servers at many points worldwide, a CDN can locate the client’s ISP and serve content from a nearby node, reducing propagation latency and, when caching static resources, also reducing server response time.

For example, if a CDN reduces the East‑West latency from 50 ms to 5 ms and the client‑to‑CDN last‑mile is 5 ms, a fiber‑connected client’s total round‑trip becomes 18 + 5 + 5 + 5 + 18 = 51 ms, a 365 % improvement over the 186 ms without CDN.

Last‑mile   Coast‑to‑Coast (low)   Server Response   Total (ms)   Improvement
Fiber       185                     50               185          365 % (−135 ms)
Cable       265                     50               265          301 % (−135 ms)
DSL         445                     50               445          231 % (−135 ms)
4G          505                     50               505          217 % (−135 ms)
3G          200                     50               550          133 % (−135 ms)

The analysis shows that last‑mile latency dominates performance, and CDN effectiveness diminishes as that latency grows. Placing CDN nodes outside ISP networks limits their impact on mobile users, especially on slower connections like 3G.

Operationally, moving edge caches closer to users—potentially inside carrier networks—could improve latency, but it introduces significant hardware, maintenance, and security challenges, such as deploying third‑party TLS termination points.

Carrier‑level agreements are complex; CDNs typically operate at shared peering points because deploying inside each ISP requires separate contracts and settlement. Even when agreements exist (e.g., Verizon’s acquisition of EdgeCast), the cost, security, and logistics remain daunting.

In China, the CDN landscape is even more fragmented, with many regional operators and a mix of large and small providers. Competitive pressures have led to collaborations between CDNs and carriers, but the fundamental issue—high last‑mile latency on mobile networks—remains unresolved.

Source: Internet