In our case I'm referring to originated mail, not forwarding/SPF upgrade
scenarios.
Over roughly the last 12–18 months we've seen a noticeable change at
OVH. More recently, a significant fraction of mail from OVH customers
arrives through their centralized outbound SMTP infrastructure (e.g.
|*.mo574.mail-out.ovh.net|), so it appears they are encouraging or
enforcing the use of managed outbound relays.
Before that, however, we saw very large malware and phishing waves
originating directly from customer VPSes (typically |vps-*.ovh.net|).
Operationally, the Return-Path and visible From: covered a very wide
variety of domains. I can't say what controls OVH applied at account
creation or submission time, only that from the receiver's perspective
these systems generated a substantial amount of abuse. Eventually we
ended up scoring entire hosting ranges because individual IP reputation
couldn't keep up with the churn.
We're a multi-tenant perimeter provider, and many of the organizations
we protect still have relatively weak SPF/DMARC deployments. As a
result, attacks using lookalike or impersonated From: addresses remain
common in our environment, even if they don't always involve a strict
RFC 5322.From / RFC 5321.MailFrom mismatch.
So my earlier comment was more about where the abuse originates (hosting
providers and residential networks with open port 25) than about
confirmed two-address impersonation specifically.
Separately, as a receiver and forwarding service, we've become
increasingly interested in ARC because legitimate aliases and Sieve
redirects can otherwise cause SPF/DMARC failures downstream, which in
turn may negatively affect the reputation of forwarding MTAs. That's a
different problem from the abuse originating on hosting providers, but
it's one we've also encountered operationally.
Carlos Horowicz
Planisys
On 27/07/2026 01:08, Douglas Foster wrote:
Are you saying that ovh and some others ensure a correct Mail From
address but allow arbitrary From addresses? I expect most hosting
services to prevent that. I also expect attackers to avoid
environments that require them to expose their identity.
Are you seeing this on originated mail or is it SPF upgrade trickery
via forwarding though the hosting service?
My 1000 to 1 estimate is based on several years experience requiring
SPF Pass or equivalent in local policy. Only on the last 90 days have
I found a confirmed case of two address impersonation. Once it was
confirmed, I created exceptions for acceptable ESPs and others, then
began requiring From authentication via DMARC or equivalent in local
policy. The tighter block rule catches very little impersonation
DF
On Sun, Jul 26, 2026, 5:42 PM Carlos Horowicz
<[email protected]> wrote:
Hosting (in particular ovh) has been the biggest source of waves
of malicious email in the last 12-18 months ... not necessarily
bare metal
Residential FTTH-like services that have port 25 open are also a
huge source of spam, malware, etc
Last but not least, some sporadic malicious campaigns of mails
containing phishing+malware links via sendgrid ... these are very
expensive to detect
Carlos Horowicz
Planisys
On 26/07/2026 22:09, Bron Gondwana wrote:
On Sat, Jul 25, 2026, at 13:07, Douglas Foster wrote:
Essentially all malicious impersonation come from clients of
bare-metal hosting services, when those hosting services do not
restrict outbound SMTP.
I'd be interested to know if other receivers agree that this is
their main source of impersonated email in 2026.
Bron.
--
Bron Gondwana, CEO, Fastmail Pty Ltd / Fastmail US LLC
[email protected]
_______________________________________________
dmarc mailing list [email protected]
To unsubscribe send an email [email protected]
_______________________________________________
dmarc mailing list -- [email protected]
To unsubscribe send an email to [email protected]
_______________________________________________
dmarc mailing list -- [email protected]
To unsubscribe send an email to [email protected]