mirror of
https://codeberg.org/postscriptum/gemlog.git
synced 2026-02-19 14:32:40 +00:00
81 lines
No EOL
2.7 KiB
Text
81 lines
No EOL
2.7 KiB
Text
# Geminispace speed
|
|
|
|
I really have some curse with connection issues last time.
|
|
|
|
Before, it was just broken ethernet socket, where the reason described in
|
|
|
|
=> tricky-ethernet.gmi Tricky Ethernet
|
|
|
|
By replacing the wires, found also that gemini pages loading much slower, comparing to huge HTML in Firefox.
|
|
|
|
I have nice optical provider, so the issue 99% is local. Much local.
|
|
|
|
Even asked on Station where found nobody have same problem:
|
|
|
|
=> gemini://station.martinrue.com/ps/fbfeda03ab214257bab096e5a0df5d41
|
|
|
|
After some time of thinking, found that I've configured custom Alfis DNS resolver:
|
|
|
|
=> https://github.com/Revertron/Alfis
|
|
|
|
This resolver running on the remote machine and connection there established with encrypted Yggdrasil tunnel:
|
|
|
|
=> https://github.com/yggdrasil-network/yggdrasil-go
|
|
|
|
So well, let's ping current DNS instance first:
|
|
|
|
``` Alfis server over Yggdrasil
|
|
PING 301:5eb5:f061:678e::53(301:5eb5:f061:678e::53) 56 data bytes
|
|
64 bytes from 301:5eb5:f061:678e::53: icmp_seq=1 ttl=64 time=212 ms
|
|
64 bytes from 301:5eb5:f061:678e::53: icmp_seq=2 ttl=64 time=58.6 ms
|
|
64 bytes from 301:5eb5:f061:678e::53: icmp_seq=3 ttl=64 time=55.0 ms
|
|
```
|
|
|
|
An attempt to resolve latency for initial Yggdrasil connection:
|
|
|
|
Q me
|
|
> Is any way to change handshake lifetime if I own both nodes connected?
|
|
|
|
A @neilalexander
|
|
> Not at the moment, no, the session timeout is ~1 minute IIRC
|
|
|
|
Well, let's check DNS points now:
|
|
|
|
``` Google
|
|
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
|
|
64 bytes from 8.8.8.8: icmp_seq=1 ttl=116 time=27.9 ms
|
|
64 bytes from 8.8.8.8: icmp_seq=2 ttl=116 time=27.8 ms
|
|
64 bytes from 8.8.8.8: icmp_seq=3 ttl=116 time=27.8 ms
|
|
```
|
|
|
|
``` CloudFlare
|
|
PING 1.1.1.1 (1.1.1.1) 56(84) bytes of data.
|
|
64 bytes from 1.1.1.1: icmp_seq=1 ttl=60 time=1.13 ms
|
|
64 bytes from 1.1.1.1: icmp_seq=2 ttl=60 time=1.29 ms
|
|
64 bytes from 1.1.1.1: icmp_seq=3 ttl=60 time=1.72 ms
|
|
```
|
|
|
|
``` AdGuard
|
|
PING 94.140.14.14 (94.140.14.14) 56(84) bytes of data.
|
|
64 bytes from 94.140.14.14: icmp_seq=1 ttl=55 time=177 ms
|
|
64 bytes from 94.140.14.14: icmp_seq=2 ttl=55 time=177 ms
|
|
64 bytes from 94.140.14.14: icmp_seq=3 ttl=55 time=177 ms
|
|
```
|
|
|
|
By irony, Alfis resolver has default forward option:
|
|
|
|
=> https://dns.adguard.com/dns-query
|
|
|
|
Including remote connection over Yggdrasil (plus initial routing time),
|
|
I have about 500ms delay in total, just for IP resolve before page starting to load.
|
|
|
|
## Solution
|
|
|
|
* It's better to run Alfis on localhost as canonical
|
|
* Switch to CloudFlare 1.1.1.1:53 on remote /etc/alfis.conf
|
|
* Use Gemini client, that supports DNS cache (e.g. network.dnsCacheExpiration option in about:config) or just delegate this feature to the system-wide environment
|
|
|
|
## Conclusion
|
|
|
|
* Secure connection has it cost
|
|
* Geminispace is good place to find new issues about routing config |