rbl-style zones?
Jim Reid
jim at rfc1035.com
Mon May 9 10:45:09 UTC 2005
>>>>> "Stefan" == Stefan Schmidt <zaphodb--bind at zaphods.net> writes:
Stefan> On Mon, May 09, 2005 at 10:56:53AM +0100, Jim Reid wrote:
Stefan> A smaller footprint for DNSBL-Zones would have spared us
Stefan> one extra GB of RAM per machine.
>> Maybe so, but how much is a GB of RAM these days? And is that
>> more expensive than spending disk space, CPU cycles, postmaster
>> resources and bandwidth dealing with spam that the RBL zones
>> block?
Stefan> No, thats why we upgraded the machines in question some
Stefan> time ago. Even so i am not a native english speaker, i
Stefan> did not use the formulation 'would have spared us' on
Stefan> error.
We seem to be confusing each other. You're right. Reducing the memory
footprint of the name server is highly desirable. This isn't always
possible because of huge data sets: zones and/or RRs. Which is where
database back-ends like BIND-DLZ come in. The point I was trying to
make is that the hardware for a big name server for spam prevention is
a small part of those anti-spam costs. For instance the labour cost
alone of someone installing some other DNS implementation on those
servers is probably a lot more than throwing hardware at the problem
by adding an extra GB or two of memory.
More information about the bind-workers
mailing list