To:
<Alain.Durand@Sun.COM>, <yasuhiro@nttv6.jp>
Cc:
<jim@rfc1035.com>, <dnsop@cafax.se>
From:
<teemu.savolainen@nokia.com>
Date:
Fri, 14 Nov 2003 01:39:21 +0200
content-class:
urn:content-classes:message
Sender:
owner-dnsop@cafax.se
Thread-Index:
AcOqNOju3usGXnuVS7ie1TdN67BHXAACcYyg
Thread-Topic:
DNS discovery
Subject:
RE: DNS discovery
> -----Original Message-----
> From: ext Alain Durand [mailto:Alain.Durand@Sun.COM]
> teemu.savolainen@nokia.com wrote:
>
> >The wasted time is not much for one user. Perhaps it is not
> much even for an operator. One point I wish to make is that
> when the wasted time is multiplied by possibly hundreds of
> millions of users worldwide - even daily perhaps - isn't it
> worth of serius consideration in system/protocol desing?
> >
>
> I've been trying to understand better this line of reasoning,
> but I honestly
> have a hard time to reconcile the desire to spare one packet exchange
> with the opinion that some Mobile IPv6 people expressed that
> RA should be send very frequently to ease movement detection....
Please correct me if I misunderstood your comment. The desire to spare one packet exchange was based on the fact that the host has to wait idle until packet exchange completes. Thus that idle time, when good transmission capability is not used at all, postpones start, and completion, of the actual communication the host wishes to do (presumed that the host needs DNS to start communication).
It is another issue if Mobile IPv6 people decide to decrease link throughputs by sending more RAs to those 200.000.000 million users (for good reasons I believe). The effect of their change may be bigger, or a lot bigger, than effect of one packet exchange, but still the packet exchage causes idle time that from my opinion should be avoided if possible.
I suppose in point-to-point links they do not need to increase RA frequency.
An answer to earlier question by Shirasaki Yasuhiro: IF an operator charges per byte and IF the operator includes RA size in byte count, increase in RA size increases cost for the user. I realized that I do not have enough knowledge in operator charging policies to really say if this is an issue, so I must step back a bit..
Regards,
Teemu
#----------------------------------------------------------------------
# To unsubscribe, send a message to <dnsop-request@cafax.se>.