Board:   [fsxNet] fsxNet Network Discussion
Subject: Re: Hub 3 rate limiting
From:    fusion(21:1/616)
Date:    Wed Mar 25, 2026 11:16

On 25 Mar 2026, The Wanderer said the following...

 TW>   Re: Re: Hub 3 rate limiting
 TW>   By: fusion to Vorlon on Mon Mar 23 2026 09:47 pm
 TW>
 TW>  Vo>> There is no need what so ever to poll a hub/uplink every minute. If
 TW>  fu> on the contrary, i think that type of polling could be just fine. you
 TW>
 TW> Multiple people have just said this isn't fine. Just because systems are
 TW> modern and/or on the internet doesn't mean being wasteful is OK.

concensus (especially here) does not equate to reality. while i agree this might matter on a 486 running some extremely crusty software, the majority of people don't fit in that category anymore.

my node1.log file on mystic here is currently at 263k lines long. 99% bot traffic and 1% a few calls every day. the bbs is set to just hang up on non-ANSI users.

could i set up a fancy firewall? mess around with fail2ban and a myriad of ip blacklists? limit my bbs only to certain regions? sure

or i could do absolutely nothing whatsoever. let them connect all day long. doesn't affect me. not my ability to stream video, or download, or upload max speed.

the result is the same. the computational resources would be basically the same too. effectively nothing. in the case of an ip blacklist it might be *more* because i'm connecting outward to check an ip, or searching a file for whether an ip is from russia or not. something a *real person* could perceive and be annoyed by.



as an aside:

where does this connection rate limit data get stored? i'm guessing it has to be cached somewhere. if it's in the database too we're talking:

check associated account (done for both rate limited and valid anyways)

check if the connection is too recent (only for rate limited)

scan for new messages in subscribed echos. this would be expensive. EXCEPT that
it already has to be done for every single crash message. if the account is one that polls just set a "has new mail" flag to 1. if the flag is 1 you hand over messages, if it's 0 you just close out as usual

they can poll all day long every minute which they do anyways, the transaction and cpu effort is likely the same (gotta ask for the user data table anyways), and they can get their mail "seemingly" as frequently as a crash user would.


 TW> Being wasteful isn't OK - there's limited resources, and it's wise to
 TW> manage things accordingly.

i can't say i agree that there are "limited resources" the way you think
there are.

--- Mystic BBS v1.12 A47 2021/12/25 (Windows/32)
 * Origin: cold fusion - cfbbs.net - grand rapids, mi (21:1/616)
.