On 08/15/2011 10:31 AM, Steven Bellovin wrote:
On Aug 15, 2011, at 10:12 21AM, Randy Bush wrote:
I've always wondered if the next cisco/juniper 0 day will be delivered via a set of exploits delivered via a link posted to NANOG. :) Maybe I'll do a talk at DEFCON next year about that. more likely a 'shortened' url. how anyone can click those is beyond me.
I'm curious what your objection is.
Mine is privacy -- the owner of the shortening site gets to see every place you visit using one of those.
That's why I have my own url shortening service using yourls. (http://yourls.org/)
I don't think there's a significant incremental security risk, because the URL you click on doesn't tell you what you'll receive in any event. Exactly.
Case in point: https://www.cs.columbia.edu/~smb/SMBlog-in-PDF.pdf does *not* yield a PDF. (As far as I know, it's a completely safe URL to click on, but I can't guarantee that someone else didn't hack my site. I, at least, haven't put any nasties there.)
Or so you claim! :) And a PDF file is a particularly potent infection vector. It would be interesting to put up a PDF (say OSPFvsISIS.pdf or WhyAnyoneWhoIsn'tNamedOwenHasRottenv6Ideas.pdf) with an exploit. This exploit could be a toe hold, which grabs other malware, opens reverse remote shell etc. If one is targeting very long term exploitation at mass scale, sitting in the network control plane for a long period of time is a large factor. And if one entices operators to download malware , the first step of most attacks (elevating privileges) is often much easier (certainly faster, as operators doing something privileged is a regular occurrence).
Given the rate of hacking -- is anyone really safe from a determined amateur attack, Maybe.
let alone state-sponsored nastiness? -- and given the amount of third-party content served up by virtually all ad-containing site, you really have no idea what you're going to receive when you click on any link.
Yep. I see hacked ad content every single day.