bind vs djbdns

Jim Reid jim at rfc1035.com
Thu Aug 24 17:49:53 UTC 2000


>>>>> "Bill" == Bill Manning <bmanning at ISI.EDU> writes:

    Bill> The Bind penetration numbers you cite may be a bit high.

Maybe, maybe not. In the absence of any better impartial numbers, what
was the breakdown of the name server implementations you found on your
last in-addr.arpa survey? Could you publish them here or point me at
the URL?

    Bill> Only (relatively) recently has bind had the support that you
    Bill> mention.

True, but even before then the software was supported by the OS
vendors who shipped it. At least notionally.

    Bill> For most of its life it was "open-source".

It still is.

    Bill> A monoculture codebase has its own problems. I am greatful
    Bill> that lucent, microsoft, ultradns, cisco, and even djb have
    Bill> contributed to the DNS "genepool".

Yup. As you've said before, diversity is good. I agree. Having other
implementations and different perspectives on how to run a name server
is a Good Thing. It's a shame for the net that none of the products
you've mentioned has so far found its way on to something important
like a root or TLD server. Maybe that's down to inertia, or lack of
maturity, or robustness or support or some combination of these?

Some of the products you mention are disguised versions of BIND. IIUC
the Lucent and Cisco offerings are just BIND with a slick GUI and some
proprietary extensions, though I hope that's not being unduly glib or
dismissive of those products.

Speaking of monoculture codebases, it would be nice if there was a
viable alternative to IOS and Cisco's implementation of routing
protocols for a significant chunk of the world's routers.



More information about the bind-users mailing list