An office worker, a person surnamed Park (43), recently left for a trip to Japan with a friend, a person surnamed Kim (43). The two connected to the same Japanese carrier's mobile network with the same model of smartphone but used different travel eSIM products. Their smartphones picked up the local mobile network normally. However, according to Park, video calls cut off every three seconds and web pages took about 5–6 seconds to open, while Kim used the internet without particular inconvenience.
Why did the quality differ even though they used the same carrier network and the same smartphone model? The answer lies in the data handling path after the base station. The carrier name and "5G" shown on a smartphone indicate the connection status between the device and the local base station. But which country the data passes through to reach the internet after arriving at the base station varies by eSIM product. This is why it is hard to judge actual network quality based on ad phrases like "local 5G support" or "unlimited data."
According to a July 9 report by global internet speed measurement firm Ookla (The Travel eSIM Performance Penalty), two out of five speed tests conducted on major hub-style travel eSIM brands selected overseas servers rather than servers in the user's country of stay. That is because a widely used "hub method" aggregates data from multiple countries at a specific overseas hub before connecting to the internet.
Ookla analyzed about 190 travel destinations and more than 50 eSIM brands based on data measured on Android devices in the second quarter of this year. Speed tests automatically select the measurement server closest to the network. By cross-checking the location of the selected server with actual latency and public interconnection facilities, Ookla identified the hubs where eSIM data connects to the internet.
◇ Same 5G network, but latency up to 30 times
It is costly for travel eSIM providers to build their own data processing facilities in countries around the world. As a result, many lease global roaming and wholesale carrier networks and send data to one or a few hubs such as Madrid, London or Singapore. Providers can reduce build-out and operating expenses, but users must tolerate latency caused by data traveling long distances.
In Ookla's analysis, the median latency of travel eSIMs ranged from 1.5 times to as much as 30 times that of local mobile users, depending on the brand and destination pairing. A product that sent Australian users' data to Amsterdam in the Netherlands had latency 30 times longer than for local users. Products using the Madrid hub also showed longer latency than local users by 25 times in Singapore, 18 times in Thailand and 15 times in Japan.
On au, the mobile brand of Japanese carrier KDDI, differences by data handling path were clear. Local users' latency was 36 milliseconds (ms). A travel eSIM that connected directly to the internet in Japan recorded 54 ms, but a product that sent data to Madrid reached 543 ms. The radio conditions between smartphone and base station were the same, but latency widened tenfold depending on the post–base station path.
Latency is the time from when a user requests data to when a response is received. Even if download speeds are fast, long latency can cause delayed reactions or choppy screens during video calls or conferences. For example, according to Ookla, travel eSIM provider A had a relatively fast median download speed of 56.87 Mbps measured across countries. However, because it primarily sent data to a Madrid hub, its median latency reached 320 ms. In contrast, providers that placed connection hubs across multiple regions showed relatively short latency. eSIM provider B, which distributed connection hubs in Germany, the Netherlands, the United States, Singapore and Japan, recorded a low median latency of 117 ms.
◇ Data paths hard to verify before purchase
The problem is that consumers find it hard to verify data handling paths before purchasing an eSIM. Many products provide only supported countries, data allowances, validity periods and local partner carriers. They rarely disclose the country of internet connection, expected latency and whether speeds are throttled, which determine the quality users actually feel.
A telecom industry official said, "The 5G icon on a smartphone only means it is connected to the base station in front of you; it does not tell you which city in the world your data traveled to and from."
After purchase, you can roughly infer the data path. According to the telecom industry, turn off local Wi-Fi and virtual private networks (VPNs), turn on only the travel eSIM's mobile data, then check the country and latency of the server automatically selected by the speed test. You can also refer to which country the IP address shows. If you are in Japan but a Spain server is selected or the IP address shows Spain, the data is likely passing through a Spain hub. However, even with this method, you cannot confirm the entire actual route, and it is hard to use for product comparison before purchase.
According to Ookla, travel eSIMs accounted for 3.4% of all international roaming activity in June this year, up 0.7 percentage points (P) from 2.7% in June last year. As eSIM use grows quickly, eSIM providers need to disclose, before purchase, the country of internet connection by destination, expected latency and whether speeds are throttled, rather than emphasizing only "5G" and "unlimited."