[ Let's see how long this thread will continue and if this time some new points are made or simply all the well known(?) facts are repeated. I mostly participate with this message since I feel many "spectators" / lurkers simply don't _know_ much else but BIND -- if at all they know any alternative. That's why I urge everyone with serious interest in the topic to actually have a look at them before repeating the ubiquitous "all the world is BIND" and "if others _say_ it's good it must be, I don't need to check myself" mumble. ] On Mon, Jan 21, 2002 at 04:14 -0700, Kurt Seifried wrote:
This is a lot like Sendmail, older versions sucked, so they did a rewrite/audit and secured it reasonably well. Things change.
http://www.isc.org/products/BIND/bind-security.html
So far (knock on wood) Bind 9.x hasn't had any serious security bugs.
Well, I'm not positive if there haven't been such bugs or if they simply haven't been discussed in public. I remember the recent bugtraq mini thread which started with the MaraDNS announcement (see <20020109123631.A24072@artemas.reachin.com> and <20020110040505.24874.qmail@cr.yp.to>). Especially the message by DJB which wasn't accepted by the moderator (http://cr.yp.to/djbdns/bugtraq/20010201072942-22539-qmail@cr-yp-to) and the links therein caught my interest: As much as some of us might dislike his wording or doggedness(id?) in some respects, he often (always?) has a valid point to make. As long as those urging questions are still left unanswered (are the new BIND developers really different from those drunken monkeys who wrote previous versions? is the code really a new rewrite or just added bloat on top of previous acked crap? has none of the many severe acked bugs been security relevant? who else besides DJB actually *guarantees* the software's functionality and robustness? etc) I somewhat doubt ISC's announcements and put them into the same drawer where I put Bill Gates' latest "we're heading for secure products and put features behind" blurb. Once I can see that things improve, I might change my mind. But until then I'm sceptic when looking at the recent history. In contrast I have yet to see djbdns die on me when it's pushed by the outside world with all the [snip] in it. And frankly speaking setting up djbdns I never had to watch its logs too closely. It simply runs. Reliably and without hogging resources. Plus once you get the concept it's way much easier to handle than "the standard" in any day-to-day scenario, see the "ease of use" comparison on the cr.yp.to site. Admittedly I don't need any of the "missing" features (SecDNS, dynamic updates). But I assume neither do most of the people still running BIND. I can only explain people use BIND because they're too lazy to change or just don't know there's something else out there. Very few will be prevented from changing since the alternative lacks features. And it's still left to further discussion if missing features are better handled by a separate tool (and to return to the most asked for: djbdns *does* support AXFR -- as a server as well as a client, it's just not done in the process which holds the authoritative data and serves record requests). Regarding the license which keeps distributors from providing packages I don't have any problems here. Most people only understand "I don't get a binary, so I stick with what comes with the CD". They don't see the point behind where DJB actually states "employ the software in the way it was designed by me and I will guarantee that it *will* work as designed and announced". To repeat the above point: Who else does this? Plus I'm always free to get the source and modify (patch) it should I wish for a different behaviour. What else could I want? To conclude: djbdns _definitely_ is worth a look. And if it only was to see that things can be done differently. :) But those who don't _need_ BIND probably will stick with djbdns once they get over the "being different" (BTW: what's the point in running any kind of UNIX on a PC when they always come with Windows preinstalled? If "what's usually preinstalled MUST be good for me" is not a valid excuse then why should be "most others run BIND, so do I"?). virtually yours 82D1 9B9C 01DC 4FB4 D7B4 61BE 3F49 4F77 72DE DA76 Gerhard Sittig true | mail -s "get gpg key" Gerhard.Sittig@gmx.net -- If you don't understand or are scared by any of the above ask your parents or an adult to help you.