<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
Hello,<br>
<br>
<div class="moz-cite-prefix">Peter Gutmann wrote:<br>
</div>
<blockquote type="cite"
cite="mid:SY4PR01MB6251C79A732DEB133417C836EE549@SY4PR01MB6251.ausprd01.prod.outlook.com">
<pre class="moz-quote-pre" wrap="">Hat tip to an anonymous contributor: Go to <a class="moz-txt-link-freetext" href="https://www.dogpoopsigns.com">https://www.dogpoopsigns.com</a> and
look at the certificate. It's actually issued for <a class="moz-txt-link-freetext" href="https://www.smartsign.com">https://www.smartsign.com</a>,
but is also valid for <a class="moz-txt-link-abbreviated" href="http://www.smartsign.com">www.smartsign.com</a>, smartsign.com, myengravedsign.com, [...]
OK, it's CDN certificate, for which a random jumble of unrelated DNs like this
is the norm. The unintended dogpoop comedy is a good indication of the state
of commercial PKI in this area.</pre>
</blockquote>
<br>
It is so true... There is tech, and there is how it is deployed. PKI
is not that awful as a technology, but we did mess it up immensely
in the past 25~ years of its deployment. One big mistake was the
improper alignment of incentives (CAs are paid for issuing certs and
seldom penalized for issuing them to the wrong requestor), and
another big mistake was accepting a model where any CA can sign any
cert with us being equally happy about it.<br>
<br>
The result of the latter is this CDN effect... Once we decided that
it's fine for anyone to sign anything, we only called for such
cost-cuts by CDNs that happily create one cert for serving all their
customers' data.<br>
<br>
(In an optimal world, we wouldn't have to trust the CDN, but this
will require a TLS without the 'T'...)<br>
<br>
Hagai.<br>
<br>
<div class="moz-signature">-- <br>
<i>Hagai Bar-El</i><br>
<a href="https://www.hbarel.com?med=sig" moz-do-not-send="true">www.hbarel.com</a><br>
</div>
</body>
</html>