On Thu, Dec 06, 2007 at 03:22:04PM +0000, Lican Huang wrote:
> My draft is about to handle possible problems when huge amount of domain
> names when Internet is in Ipv6 stage. Because of unlimited amount of Ipv6
> addresses, unlimited amount of hosts ( servers, PC, even mobile phones, etc
> ) will have static Ip addresses in the Internet. So, these hosts may
> require domain names if we use Domain Name Systems as the way today.
not unlimited. there is an upper bound to the number of
IP addresses as well as an upper bound on the number of
domain names. these bounds are mathmatically fixed.
and potentially, yes. each IP address might have a unique
DNS domain name attached. Or there may be more than
one name attached to an IP address. Or there may be many
IP addresses attached to one DNS domain name.
>
> One problem is how to implement the DNS with huge amount domain names.
> I don't think today's DNS implementation can handle successively with
> huge amount domain names in the future. That is why I wrote a
> distributed-dns-implementation draft to try to solve this problem.
Your thinking here seems to be flawed. Why do you think the DNS
can not handle the potentially large numbers of new domain names
which might be registered? Please show the math.
> Another question is when there are so huge amount domain names in the
> future, why we don't give these domain names semantic meaning? Can you
> figure out what's the meaning about "www.u8erbjsdhdfdsdf.com " from bilions
> of domain names? You may say we can use SEARCH by the key words and get
> the link of www.u8erbjsdhdfdsdf.com. But, in this way, domain names are
> useless , because we can totally use IP address or any other handle to
> represent www.u8erbjsdhdfdsdf.com. You may say we use domain names as stable
> name because Ip address may be changed. But , why use these ugly domain
> names? Why not semantic domain names?
dealing with larger numbers of names is a different question
than the one of semantic meaning. your concerns about semantic
meaning are understandable, but flawed. the flaws are along two
axis... a) that the lable www.example.com MEANS there is an HTTP
server. there is no such requirement or even implied assumption
that it is the case. you ask the question; "what's the meaning i
about "www.u8erbjsdhdfdsdf.com " from bilions of domain names?" to
which I reply - the meaning is that is the site where I can download
my favorite licorice recipies. It means something to ME. Just
because it is not apparent to you, does not divorce it from meaning
to others. this leads to the second (and perhaps fatal) flaw;
b) that you seem to assume ASCII encoded labels. In todays DNS,
there are many places that do not use ASCII encoding. DNS lables
take the form xn-?????.xn-????.xn-???. which in and of themselves
have no semantic meaning whatsoever. The beauty of the encoding
is that when the DNS lable is handed to an application, it is able
to decode the lable into a script that is understandable to the
target user... so folks who only read Mandarin can actually have
DNS lables encoded so that they are parsed/displayed as Mandarin.
And the same can be done for Hindi, Farsi, Korean, and even English.
> How to name semantic domain names? We can let specific virtual
> organizations ( or registrar comanies ) to do. That is, ICANN controls top
> level domains. Lower level domain names is controlled by virtual
> organizations ( or registrar comanies) according to the clasification of
> contents. In this way, we can figure out hieararchical classification of
> contents very easily by trace down the heararchical domain names.
> If domain names are named as this way , we can easily add SEARCH power in
> DNS just like my draft.
Perhaps that could be made to work. I remain dubious of the
value, particularly in light of alternate encoding methods and
the ambiguity on the semantic equivalence of various strings
encoded differently. You have not persuaded me of the value
of proceding down this development path.
> Semantic domain names does not takeover the current domain names in
> the first stage. We can use new TLDs to manage semantic domain names, and
> let the old TLDs to be managed as the way today.
Show me where/how you will get enough traction to get this started
and what real problems it solves for people that are not already
solved in other ways.
>
> Lican
>
>
>
>
>
>
>
>
>
> [EMAIL PROTECTED] wrote:
> On Wed, Dec 05, 2007 at 02:10:52AM +0000, Lican Huang wrote:
> > If SEARCH outside DNS were full power, then DNS would disappear soon. And
> > all DNS registrar companies would broken out.
>
> perhaps you are right. at this point we don't have enough data.
>
> > What is the difference between www.microsoft.com and www.jksdfjsdfdfsdf.com
> > if they represent for the same address of http page? We can browser the
> > micorsoft's web page through the link of the SEARCH output easily. But if
> > microsoft company used www.jksdfjsdfdfsdf.com for its domain name, then
> > what useful this kind of DNS would exist?
>
> at what poiint in time did the string "microsoft" gain any sort
> of human memorable meaning? what would have been the result if
> Bill Gates named his new company "jksdfjsdfdfsdf"? 25 years
> later, it would be a globally recognizable mark and you would
> be arguing over other strings.
>
> >
> > My opion is that in the future if DNS would survive, DNS must have some
> > reform.
>
> you are entitled to your opinion. others are entitled to thier
> opinions as well. you seem to have failed, this time, to persuade
> people that adding search to the DNS is a wise & prudent thing for
> the evolution of the protocol. im my own case, having implemented
> rudimentary search in the DNS - i can't recommend it for anyting
> other than as an interesting academic exercise. the pieces
> you have written drafts about fail to include a key, critical
> component of a DNS with Search capability. Still, an interesting
> stab at a perceived problem. It might make more sense if you actually
> had all the required peices documented.
>
> --bill
>
> >
> > [EMAIL PROTECTED] wrote:
> > On Tue, Dec 04, 2007 at 04:27:06AM +0000, Lican Huang wrote:
> > > When Ipv4 addresses will be Exhausted in the near future and the next
> > > generation Intenert( Ipv6) will take over, DNS names will also be
> > > exhausted soon with the increase of hosts and users. Lenny Foner has
> > > pointed other disadvantage in the today's DNS.
> > > Please see the section of "What's broken?" in the article of Lenny Foner
> > > in
> > > http://www.cfp2000.org/workshop/materials/projects-dns.html.
> >
> > Full IPv4 utilization and increasing use of IPv6 is completely
> > orthonginal to DNS label exaustion. Some have argued that all
> > the "good" names are taken; e.g. the DNS is exausted. This was
> > first proposed in 1996 (to my memory) yet more than a decade later,
> > we see that the domain name system is robust and growing.
> > With the inherent hierarchical structure of the DNS lable, the
> > mathmatical upper bound is pretty high and we are no where near
> > DNS name exaustion. If you have actual data indicating otherwise,
> > I'd love to see the studies.
> >
> > >
> > > Domain Names in DNS must have some human-understanding meaning it,
> > > otherwise, we can just use IP addresses or numerials for the names. In
> > > other words, if we use human-not-understanding Names in DNS, the DNS
> > > system can be throwed away.
> > >
> > > The draft namespace is different with the today's DNS namespace. But, due
> > > to the exhaustion of Names in DNS in the near future, The DNS will add
> > > new domains.
> > > Why adding new domain names with semantic meaning in the future?
> >
> > DNS names do not -HAVE- to have human understandable components.
> > In many cases, this is highly desired -BUT- is not required for
> > use. And yes, numeric literals have been used in the past.
> > Use of the IP address instead of the name is one of the failures
> > of application design. The IP address indicates WHERE a node is
> > in the Internet topology, not the identity of the node. The
> > Name is the indicator of the node IDENTITY. the DNS maps names to
> > addresses and makes no assurance as to the human friendliness of the
> > name or the reachability of the address. Your assertion that the
> > "DNS system can be thowed away" is vacuously true. If you find it
> > non-useful, there is no requirement for you to use it. Many people
> > use the DNS to get a lable, memorable or not, and then use other
> > tools to map that lable into something meaningful... e.g. SEARCH.
> > It does not invalidate the use of the DNS in any way.
> >
> > >
> > > This draft can be used for search the locatons of the resources if the
> > > DNS using classified hierarchical Domain Names.
> > >
> >
> > I think I prefer SEARCH to be outside the DNS (having actually
> > built a varient of the DNS which supported regular expression
> > expansion of the "?" and "*" characters...)
> >
> > Your milage will vary.
> >
> > --bill
> >
> >
> > > Mohsen Souissi wrote:
> > > I have read the I-D as well and I second Joe's point of view and his
> > > arguments below.
> > >
> > > Mohsen.
> > >
> > > On 03 Dec, Joe Abley wrote:
> > > | Hi,
> > > |
> > > | I have read your draft, draft-licanhuang-dnsop-urnresolution-00.
> > > |
> > > | The question was raised just now in the dnsop working group meeting in
> > > | Vancouver as to whether the content of this draft was suitable for
> > > | adoption as a working group item. The question was triggered by the
> > > | presence of "dnsop" in the draft name.
> > > |
> > > | I have read your document. I do not believe it is a suitable basis for
> > > | a dnsop working group item. Specifically:
> > > |
> > > | 1. The document describes a namespace which is substantially different
> > > | form what is available in the DNS today. The existing DNS namespace is
> > > | not addressed at all.
> > > |
> > > | 2. The document seems to address an extension to (or an application
> > > | for) the protocol described in draft-licanhuang-dnsop-distributeddns,
> > > | which (to this reader) seems clearly not to be "the DNS", at least any
> > > | conventional meaning of that term.
> > >
> > >
> > >
> > > ---------------------------------
> > > Support the World Aids Awareness campaign this month with Yahoo! for Good
> > > _______________________________________________
> > > DNSOP mailing list
> > > DNSOP@ietf.org
> > > https://www1.ietf.org/mailman/listinfo/dnsop
> >
> >
> >
> >
> > ---------------------------------
> > Sent from Yahoo! - the World's favourite mail.
> > _______________________________________________
> > DNSOP mailing list
> > DNSOP@ietf.org
> > https://www1.ietf.org/mailman/listinfo/dnsop
>
>
>
>
> ---------------------------------
> Sent from Yahoo! - the World's favourite mail.
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www1.ietf.org/mailman/listinfo/dnsop
