Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 8 Mar 2011 20:56:56 +0100
From: Andrea Barisani <>
Subject: Re: Vendor-sec hosting and future of closed lists

On Tue, Mar 08, 2011 at 10:59:57AM -0500, Josh Bressers wrote:
> ----- Original Message -----
> > 
> > As suggested by Josh Bressers oCERT would be favourable to providing a
> > system that would accept user submission and allow selection of security
> > contacts from our existing member database as well as other verified
> > contacts.
> > 
> > As Josh pointed out we do this already (even if manually and not with a
> > web selection thing or whatever) and I am open to explore ways to create
> > more cooperation.
> > 
> > We would also be willing to host and maintain a closed vendor-sec style
> > mailing list like the previous one with the only condition for member
> > list to be public (not necessarily the individual contact names but at
> > least the entities represented).
> > 
> I've been thinking about this a bit, and here are my thoughts.
> I think oCERT could be a good fit here. They already have contacts, and
> such a setup would likely have a formal process of sorts for vetting
> recipients of issues. My current fears are:
> 1) Is oCERT in a position to increase its current workload by several
>    magnitudes? I suspect you're going to have to expand your team by a fair
>    amount. I also imagine this will result in changes to the way oCERT
>    currently exists, perhaps not though, I can't see behind the curtain.
> 2) Will dealing with oCERT in this manner generate extra process?
>    vendor-sec was quite process free, a little doesn't hurt, but a lot can
>    be bad.

It all depends on how this process is going to be handled. I can see oCERT
helping in routing reports to the proper contacts via email to our trusted
member contacts as well as external ones that we can seek on a report basis.

Once the report is routed to the contact list I don't expect the handling
process to be different from what was happening on vendor-sec already. (if
you think about it, it would just be like creating ad-hoc aliases list where
we just make sure only affected parties are included, unlike vendor-sec used
to work).

Of course we would monitor handling and try to be a pain :) in case we see
things progressing slowly and, more importantly for us, we would eventually
disclose the information like we do already once the report is handled.

So I don't think 2) is going to be too much of a concern also because putting
a lot of extra process would become a problem re 1).

It is true that we would need to expand our team and we would be happy to do
that, it just requires an evaluation process as clearly only trusted contacts
can be accepted (there are several from the old vendor-sec that we would be
really happy to work with actually).

This is a simplified approach compared to what we do now, which is contacting
project maintainers first, negotiating embargo as well as requesting a patch
that we can verify and relay to vendors, spread the patch to affected parties
along with embargo information and coordinate disclosure. This is a
time-consuming process that worked very well with oCERT because the number of
reports was limited, it helped solving complex bugs or vulnerabilities that
affected a large number of projects.

I think we can continue to do this for selected cases but it would be hard to
scale enough to support this "tailored" approach for every single report if
we expand our visibility, additionally I think everyone would favour less
extra process as pointed out already.

> 3) Are we going to annoy other CERTs? Will they even care?

I don't think this is an issue. We positively worked with other CERTs when that
was applicable anyway.

> 4) oCERT already exists, there are going to be disagreements about how to
>    do things, both sides of all issues will need to be open to ideas and
>    compromise.
> There is also the option of recreating an old style list. This is a bit
> more ad-hoc and Openwall has already offered to host such a thing (Solar
> has quite a bit already in place). I do favor this a bit, as it would make
> a nice compliment to oss-security. It also puts our destiny squarely in our
> own hands. It is more work for the involved parties though (And a lot more
> work for Openwall)
> The disadvantages I recall from the old list are:
> 1) Membership management is a pain. Adding new people is annoying and
>    nobody ever leaves.
> 2) Nobody is in charge, which means sometimes issues can get ignored or
>    forgotten (also see #1)
> 3) The potential for leaks is probably a bit higher than using something
>    like oCERT (downstream recipients are monitored a bit more closely I
>    would hope). Perhaps a benevolent dictator type approach could help
>    prevent this.

I agree, all of these are some of the reasons that made us creating oCERT in
the first place.

Above all things we value having a published member list and having an open
disclosure process of what happened once the reports have been handled
opposed to the "opacity" that vendor-sec had in the past.

If eventually recreating the old list is going to be the favoured option I
hope that we can find a way to address these concerns.


> Whatever is decided should be done so by the groups most affected. Here is
> a collection of the top members that have contributed to the old
> vendor-sec since mid 2008 (my historic archive isn't as easy to get at, I
> can crunch it if someone wishes, I don't expect it to change much though)
> There were a handful of other people that contributed a fair amount but
> were not list members, or not part of one of the above orgs (Tavis Ormandy,
> Chris Evans, Alan Cox oCERT, and Samba for example).
> Once we have a vision for the future, we should try to let various groups
> know who they can contact in the future. I imagine some of them still don't
> know what happened to vendor-sec.
> Thanks.
> -- 
>     JB

Andrea Barisani |                Founder & Project Coordinator
          oCERT | Open Source Computer Emergency Response Team

 0x864C9B9E 0A76 074A 02CD E989 CE7F AC3F DA47 578E 864C 9B9E
        "Pluralitas non est ponenda sine necessitate"

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.