Re: DNS based Destination Selection

[David Gethings <davidg@pipex.net>]

On Thu, 2003-06-19 at 07:24, Cedric de Launois wrote:
> Indeed. We thought of this mechanism in the NAROS solution. We perfomed 
> recently a quick evaluation of this solution (use of DNS to influence
> an external source's destination decision making). First results
> clearly show this is too high level : 70 to 90% of the traffic 
> flows we analyzed were NOT preceeded by a DNS request ! (This
> is only an approximation since it is hard to see which DNS request
> is associated with which flow). This is due to the spread of p2p
> applications : lots of flows without any DNS lookup.

I must admit I was initially surprised by this statement. I have no
reason to believe that this is untrue. However having given it a little
thought this result does not really surprise me. When downloading a
webpage there will only be a few DNS lookups required (usually only one
or two), however there will be a number of flows as the web browser
requests all the images and frames (if there are any) associated with
that page.

I do concede that there will be a number of applications that will never
do a DNS lookup. So this obviously puts a question on whether DNS is the
right place to do destination selection.

The point of my suggestion is that it is a simple method of influencing
which ingress path traffic could take for a multihomed site. It is
simple enough for most SOHO and SME's to use and does not require much
networking expertise - much less than using BGP.

Will it work in every situation: no. Does it offer enough functionality
that a reasonalbe proportion of multihomed sites will wish to use it: I
believe so. Is the cost of changing DNS and host stacks worth it: I
honestly do not know.

I honestly don't believe we will find just one multihoming solution that
will suit every multihomed site. I believe this method to a be good
starting point for people new to multihoming in IPv6.

Cheers

Dg