Hacker News Reader

Comment by snurfer | Hacker News Reader

Our Starlink connection (backup WAN) recently switched from a New York ip to one near Chicago. I think we've switched to a downlink station that is geographically closer and we're getting better pings and, possibly, better throughout (needs more testing).

The service is expensive, but it's been a fun way to "participate" with all the exciting stuff SpaceX is working on.

View on HN

for a very brief period of time my starlink connection was exiting the cgnat to the internet in chicago (the first hop that was a public IP was 1-2ms from various ISPs' looking glasses in chicago), and the latency matched for return trip to LEO/back, down to earth station in the pacific northwest, and then transport circuit to chicago and back. Then after a couple of days it went back to Seattle.

Is your impression that the service is improving?

[We are South of the 45th parallel, by the way.] I think service is improving for us; we don't get the regular 5-10 second disconnects that were occurring every 10 or 15 minutes. According to fast.com, our loaded and unloaded latency has improved. Most media consumption (web pages, video, etc) compares favorably to a terrestrial service.

However, Xbox Live (the limit of our online gaming) is still inconsistent. We get disconnected from game servers and from Live itself pretty regularly.

It is definitely much better since mid January with increasing satellite density. Getting the antenna to a really high position on your roof with no tree obstructions around is important. Right now it's averaging about 0.28% loss to Seattle over the past 3 hours. But there are still some very brief periods of 4-5 second hiccups. Based on the monitoring systems I have running, it's more like 0.00% loss for long extended periods of the time now, and then a minute or two with about 10% loss. I haven't tested it a lot for online gaming but realtime voice/video conference type stuff all appears to work normally.

There is presently a very big difference in simultaneous satellite density and lack of gaps at 49N vs 44-45N, however.

Reply on news.ycombinator.com