To:
"Scott Rose" <scottr@nist.gov>
Cc:
"DNSOP WG" <dnsop@cafax.se>
From:
"Jaehoon Jeong" <paul@etri.re.kr>
Date:
Thu, 26 Jun 2003 09:04:48 +0900
Sender:
owner-dnsop@cafax.se
Subject:
Re: Comment on draft-jeong-hmipv6-dns-optimization-01.txt
After thinking about your comments, I became to agree with you. Right, I'll let it a part of network admin to maintain the list of RDNSS. I will modify my draft, reflecting Scott's and Luc's comments. Thanks a lot. Jaehoon ----- Original Message ----- From: "Scott Rose" <scottr@nist.gov> To: "Jaehoon Jeong" <paul@etri.re.kr> Sent: Tuesday, June 24, 2003 9:54 PM Subject: Re: Comment on draft-jeong-hmipv6-dns-optimization-01.txt > > My main concern is that maintaining the list of RDNSS seems to be an > operational concern, and not really a part of the protocol extensions the > draft describes. It is not really important how a network keeps the list > of DNS servers current, there are many options for a network admin to use > depending on how the network is configured. > > It would be enough to say that the MAP should take action to keep the list > of RDNSS up to date, but not mandate how that is done. That can be left as > an operational concern for an individual network. > > Scott > > ----- Original Message ----- > From: "Jaehoon Jeong" <paul@etri.re.kr> > To: "Scott Rose" <scottr@nist.gov> > Cc: "DNSOP WG" <dnsop@cafax.se> > Sent: Tuesday, June 24, 2003 1:02 AM > Subject: Re: Comment on draft-jeong-hmipv6-dns-optimization-01.txt > > > > Hi, Scott. > > Thanks for your good comments. > > Look at the inlines. > > > > ----- Original Message ----- > > From: "Scott Rose" <scottr@nist.gov> > > To: <paul@etri.re.kr> > > Sent: Monday, June 23, 2003 11:00 PM > > Subject: Comment on draft-jeong-hmipv6-dns-optimization-01.txt > > > > > > > > > > I was reading over the draft and had a question regarding section 7 > > > Detection of RDNSS Failure: > > > > > > Why does the MAP need to check the state of the recursive DNS servers? > If > > > they are all part of the same administrative domain, an operator will > make > > > sure the DNS servers are up and running. > > > > > The reason MAP checks recursive DNS servers' state is to let MAP > announce > > only the living or reachable DNS servers. > > As you say, network operator can make sure the DNS servers within the > managed domain > > are working in order. But until some DNS servers are down or out of > order and > > the network administrator handles the problem, MAP will continue to > announce the invalid DNS servers. > > > > > A client can also choose to ignore a recursive DNS server that it > believes > > > to be down. If it sends out a DNS query, but does not recieve a reply, > it > > > can mark the server as "bad" or delete the server from its list of > recursive > > > servers to query. > > > > > Though a client can detect some DNS server, it will take a time to > detect. > > I intended to reducing the detection time by MAP polling DNS servers > regularly. > > > > > I think it adds complexity for the MAP to ping every recursive DNS > server > > > unless there are a number of unstable DNS servers on a network. > > > > > Yes, you are right when MAP often checks the DNS servers. > > If MAP pings every DNS server at appropriate intervals, I think, > > there is not much complexity. > > > > I'd like to listen to other opinions. > > Thanks :-) > > > > Regards, > > Jaehoon > > > > > Scott > > > ================================= > > > Scott Rose > > > Adv. Network Technology Div., NIST > > > http://www.antd.nist.gov/proj/dnssec > > > > > > ph - 301-975-8439 > > > ================================== > > > > #---------------------------------------------------------------------- # To unsubscribe, send a message to <dnsop-request@cafax.se>.