Why CDNs Provide Little Acceleration for Mobile Clients

Google performance engineer and author Ilya Grigorik recently published a blog post titled “Why CDN Provides Little Acceleration for Mobile Clients,” describing the special characteristics of mobile (wireless) networks and proposing ideas for a mobile‑friendly CDN.

Grigorik first criticizes current CDNs for their weak impact on mobile client performance, citing monitoring data that shows only marginal improvements, and calls for a CDN architecture that better supports mobile networks.

He identifies two common misconceptions: (1) traditional CDNs improve mobile and broadband clients equally, and (2) the issue is not a lack of a “wireless CDN” but rather the carrier network itself.

He provides reference data to break down wireless network latency components, including coast‑to‑coast latency (50 ms), server response latency (50 ms), and last‑mile latency for various access types (fiber ≈ 18 ms, cable ≈ 26 ms, DSL ≈ 44 ms, 4G ≈ 50 ms, 3G ≈ 200 ms).

Note: “Last‑mile” refers to the connection from the carrier’s exchange to the end‑user device.

The following diagram (omitted) shows the user access flow and latency when using a CDN.

Using a CDN for Content Distribution Acceleration

CDN acceleration deploys cache servers at many peering points worldwide, aiming to place content as close to users as possible. In the ideal case, the CDN server instantly identifies the client’s ISP, and the last‑mile latency consists only of the ISP‑to‑CDN hop and the CDN’s immediate response.

CDNs reduce propagation latency.

When static resources are cached, CDNs also reduce server response time.

Assuming optimized network paths (coast‑to‑coast latency reduced to 5 ms) and a 5 ms client‑to‑CDN hop, a fiber‑connected client’s total latency drops from 186 ms to 51 ms—a 365 % improvement.

The table below compares performance with and without a CDN for different access technologies.

Last‑mile

Coast‑to‑Coast (low)

Server Response

Total (ms)

Improvement

Fiber

18

50

50

186

Cable

26

50

50

202

DSL

44

50

50

238

4G

50

50

50

250

3G

200

50

50

550

CDN + Fiber

18

5

5

51

-135 ms (365 %)

CDN + Cable

26

5

5

67

-135 ms (301 %)

CDN + DSL

44

5

5

103

-135 ms (231 %)

CDN + 4G

50

5

5

115

-135 ms (217 %)

CDN + 3G

200

5

5

415

-135 ms (133 %)

Repeating the calculation for each connection type reveals a clear trend: the higher the last‑mile latency, the less effective the CDN becomes; node placement outside the ISP network matters; and CDNs still provide some improvement for last‑mile latency.

Overall, CDNs reduce propagation and server response delays, but they offer little optimization for mobile clients—e.g., 3G users see only about a 33 % improvement.

Operational and Business Costs at Edge Nodes

A sensible strategy is to move cache servers closer to customers, ideally inside carrier networks rather than outside ISP peering points. Deploying servers inside carriers is possible, but practical challenges include limited peering locations, costly hardware, complex maintenance, security concerns, and the difficulty of negotiating individual agreements with many carriers.

Peering points are few, so CDNs are limited to well‑known global locations; moving servers into carrier networks would require separate contracts for each carrier.

Placing servers near wireless towers would demand massive hardware deployments, leading to operational nightmares and security issues such as managing third‑party TLS termination.

Carriers are reluctant to sign individual agreements with each website, limiting CDN expansion.

Recent news of Verizon acquiring EdgeCast suggests that integrating such assets could help Verizon’s customers address these challenges.

Beyond operational costs, CDNs provide no special optimization for mobile clients because the dominant factor is the poor last‑mile latency of mobile carrier networks. The solution lies in improving the carrier network itself, encouraging competition, and openly optimizing the internet backbone rather than merely pushing edge caches closer to users.

In China, the situation is even more complex with many carriers (e.g., China Telecom, China Mobile, regional providers) and a fragmented CDN market. Large CDN providers (e.g., Wangsu, Blue Cloud, Kuaiwang, DiLian) are partnering with carriers, and some are launching mobile‑specific acceleration solutions, but the underlying network challenges remain.

The complexity of the domestic environment makes user‑experience problems harder to solve; only coordinated reforms and deeper carrier‑company collaborations can substantially improve mobile access quality.