I’m doing some infrastructure work for TADS 3 web servers, and I need some help testing out the geolocation code.
Can you browse to gs.tads.io/test.php and let me know if it matches you to a server that seems way off?
For the purposes of this test, treat this map as the frame of reference. North America west is California and Asia/Pacific east is Australia.
The eventual plan is to have a lot of small servers running behind the gs.tads.io name, which will redirect player sessions to the nearest one. This is different from the approach IFDB takes; it doesn’t (yet) treat location as a factor when picking a server.
I’d also like to make it as easy as possible for authors to add a server to this network, with no knowledge of Linux or web servers required. If that’s something you’re interested in, please let me know. I could really use volunteers to run through the process and give me feedback.
Thanks everyone! Glad that it seems to be getting things right.
Each endpoint server will have its own hostname, and multiplayer will work the same way it does through IFDB. Both players have to connect to the same server; there’s no session sharing or replication between servers.
I got much better response time than George reports from na-east, less than 1/2 second between turns fer sure. I tried going far afield to ap-east and got much slower responses (as expected).
From central Europe the typical timings of each HTTP request is as follows (please note that typical game turn consists of 4 requests so you must multiply by 4 to get an actual response time):
Just noticeable lag in na-east, except when there’s a disambiguation – then I’d say maybe about half a second. (My computer/internet connection is often kind of laggy anyway.)