Thursday, September 1, 2016

CalSPEED testing in Rural Japan

While touring the Kansai region with my daughter last week and the week before, I ran a few speed tests on various muni WiFi networks. I’m surprised to see MOS hitting 4.0 for some of the tests despite pinging servers on the west and east coasts of the U.S.

1. City of Nara
Capital during the Nara Period (before the capital moved to Kyoto in ancient times), Nara claims to have the world's largest wood building, which houses a gigantic Buddha, Todai-ji. The speed test was run on a municipal network from a guesthouse in town. Deer roam freely through the main park, and you can buy "deer biscuits" for the equivalent of $1.50 to feed them.

Nara speed test (municipal WiFi)
City of Nara, south of Kyoto, Japan

Photo with "Sento-kun," the controversially creepy-looking town mascot.

Deer biscuits - "Awful for people"

2. Shirahama
Shirahama is one of Japan's three largest onsen (hot springs) resorts. The name means "white beach," and as you can see from the photo, the sand is white and the sea is turquoise blue. Because of its proximity to Osaka, it's a popular weekend destination. WiFi in Shirahama is available publicly and is paid for by the Wakayama Prefecture government.

Shirahama speed tests (municipal WiFi)

Shirahama is located on the southern coast of the Kii Peninsula


  1. MOS = Mean Opinion Score? That sounds like a poll result, but your description sounds like an objective measure. Can you provide a link to definitions, and comparisons?

  2. Hi Roger, nice to hear from you.

    Historically, mean opinion score was a subjective judgment of voice quality determined by person (usually someone trained to evaluate voice quality) making a telephone call. The range of scores is between 1 and 5 with 5 being the best. Generally, 4 or greater is considered "toll quality." For VoIP, we use call performance measurements and calculate a score.

    Here is an earlier post with the formula we use to estimate over the top VoIP mean opinion score:

  3. Here's a link to the ITU specification:

  4. Wow, I haven't been to Nara since the early EPON deployments began in Japan. Thanks for the post. How many different locations were you able to test? I'm curious how well distributed the performance really is.

    1. Hi Wayne,

      I only ran tests in those two locations. Maybe I'll have more WiFi connections on my next trip to do more diverse testing.