modifying bind to do weighted round-robin

Jay C Austad JCA at
Thu Mar 23 05:05:52 UTC 2000

>no-one groks them. Moreover, even if you implement advanced
>weighting/ordering on the master, what are you going to do about slaves,
>or, even worse, about intermediate caching servers,
>which are going to eliminate duplicate records and non-weightedly
>round-robin their cached answers? Reducing TTL values is a blecherous
>way of dealing with the caching servers, I
>suppose, but at what cost?

Well, right now, the commercial load balancing dns servers only hand out one ip at a time, instead of all of them.  If a large ISP, like AOL, caches just that one ip, it's going to generate a significantly larger amount of traffic just on that one server (or cluster).  At least with multiple ip's being handed out, it will round-robin (not weighted though) and distribute the load a little better.  Right now, we'll see one cluster get pounded for several minutes, and then another will get pounded for several minutes right after, even though the ratio is being set the same to each cluster.  The TTL is set very low, so it reduces this sort of behavior, but it still happens for a couple of minutes at a time.  By setting the TTL low, we are increasing the load on the nameserver and increasing our traffic, but the amount of dns traffic is insignificant compared to the amount of web traffic going out.  

As for my secondary,  initially, the zone file can be manually copied across and reloaded when modifications happen to it.

I'd like to get something up and working so I can do some testing on it to see exactly how well it will work.


More information about the bind-users mailing list