Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.arch > #118524

Relentless DDoS out there (was: google usenet archives - up or down?)

From Stefan Monnier <monnier@iro.umontreal.ca>
Newsgroups comp.arch
Subject Relentless DDoS out there (was: google usenet archives - up or down?)
Date 2026-10-07 16:02 -0400
Organization A noiseless patient Spider
Message-ID <jwvece12szr.fsf-monnier+comp.arch@gnu.org> (permalink)
References (1 earlier) <119c1aa$1okot$1@dont-email.me> <2026Sep28.075333@mips.complang.tuwien.ac.at> <119d8mm$23bo0$3@dont-email.me> <jwvbj9hfaqd.fsf-monnier+comp.arch@gnu.org> <11a4p5e$2jpr4$3@dont-email.me>

Show all headers | View raw


Lawrence D’Oliveiro [2026-10-07 06:34:22] wrote:
> On Mon, 28 Sep 2026 15:18:36 -0400, Stefan Monnier wrote:
>> It also means that DDoS is now a normal everyday occurrence for any
>> site whatsoever. It makes a qualitative change in the sense that it
>> takes a professional sysadmin to manage even the most innocuous
>> web-server.
> I don’t think having a “professional sysadmin” in charge is any
> assurance against DDOS attacks. The only remedy seems to be seriously
> grunty hardware acting as a filter frontend to get rid of as much of
> the rubbish as possible before letting anything through to the actual
> content/application server.

AFAIK hardware won't save you: there's an actual arms race going on, and
you need wetware in there to adapt the response to the ever-evolving
forms of the incoming DDoS.

That's because the attackers work hard to make their incoming "rubbish"
requests look as much as possible as the legitimate requests.

BTW, this whole arms race is not just a problem for administrators of
servers, but also for clients.  More and more, you can't use `curl` or
`wget` to download PDF articles, because they get hidden behind "are
you human?" checks.  Similarly the response to the arms race tends to
prevent legitimate scraping, such as when you want to maintain a mirror
for use behind a firewall, or say the Internet Archive.

Currently, the effort goes into technical means to guess which requests
are kosher, but I think it's ultimately impossible to do that without
imposing too many restrictions.  I'd much prefer an approach based on
governance, with ways to impose sanctions on abusers.


=== Stefan

Back to comp.arch | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

google usenet archives - up or down? Michael S <already5chosen@yahoo.com> - 2026-09-27 18:23 +0300
  Re: google usenet archives - up or down? Joseph Seigh <jseigh_es00@xemaps.com> - 2026-09-27 17:20 -0400
    Re: google usenet archives - up or down? anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2026-09-28 05:53 +0000
      Re: google usenet archives - up or down? David Brown <david.brown@hesbynett.no> - 2026-09-28 10:32 +0200
        Re: google usenet archives - up or down? Stefan Monnier <monnier@iro.umontreal.ca> - 2026-09-28 15:18 -0400
          Re: google usenet archives - up or down? Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-10-07 06:34 +0000
            Relentless DDoS out there (was: google usenet archives - up or down?) Stefan Monnier <monnier@iro.umontreal.ca> - 2026-10-07 16:02 -0400
      Re: google usenet archives - up or down? Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-09-30 03:38 +0000
    Re: google usenet archives - up or down? Michael S <already5chosen@yahoo.com> - 2026-09-30 12:53 +0300
  Re: google usenet archives - up or down? "Chris M. Thomasson" <chris.m.thomasson.1@gmail.com> - 2026-10-08 14:33 -0700

csiph-web