[Chapter-delegates] Draft ISOC comments on WGIG report

[Steve Crocker]

Caching secondary servers solve the problem quickly.  Anycast  
replication is also available.  Both are quicker and cheaper than  
operating a lettered root server.

Is there any data on countries which are not being served well enough  
by the current scheme?  It would be very useful to have some data.

Thanks,

Steve



Steve Crocker
steve at shinkuro.com


On Aug 10, 2005, at 1:04 AM, Franck Martin wrote:

> Some countries do need a root server, because they have only one  
> link to
> the Internet. Until redundancy is feasible, a root server would be
> ideal. Unfortunately in practice it does not work that way.
>
> I guess there are some bandwidth and latency savings by having your  
> own
> root server nearby.
>
> Cheers
>
> Steve Crocker wrote:
>
>
>> Patrick,
>>
>> I'd like to take issue with your comments.  See below.
>>
>> Steve
>>
>>
>> Steve Crocker
>> steve at shinkuro.com
>>
>>
>> On Aug 10, 2005, at 12:30 AM, Patrick Vande Walle wrote:
>>
>>
>>> On Wed, August 10, 2005 2:14, Franck Martin said:
>>>
>>>
>>>> Steve,
>>>>
>>>> I guess it predates ICANN, but putting such a statement in the WGIG
>>>> report implicates that there can be only 13 root servers and   
>>>> that's
>>>> it.
>>>> So many countries cannot have a root server. I find this statement
>>>> highly misleading and in defavor of ICANN.
>>>>
>>>>
>>>
>>> Franck,
>>>
>>> The statement is technically correct. I think it would be much more
>>> positive for ISOC to point out how and when the technical community
>>> will
>>> address that weakness.
>>>
>>
>>
>> In what way is this a weakness?  The domain name system is designed
>> to provide fast, reliable translation of domain names into IP
>> addresses.  It is designed to do so for all Internet users around the
>> globe and to give the same answers to everyone.  The amount of
>> redundancy needed at each level is fundamentally an engineering
>> question.  The limitation of 13 root servers addresses comes from the
>> interaction of the sizes of packets and the amount of information
>> needed for each address.  From a design point of view, half that
>> number would probably be more than adequate.  With anycast, the
>> actual replication of servers is now around 100 and can be extended
>> considerably further.
>>
>> Where's the weakness?
>>
>> (I don't want to suggest the domain name system is perfect.  There
>> are indeed some weaknesses in DNS, but they're not related to this
>> discussion.  DNSSEC addresses some of the weaknesses.)
>>
>>
>>> There may not be a technical reason to go beyond 13 root servers,  
>>> but
>>> there are plenty of political reasons. The debate here is whether
>>> technologists are at the service of the society or if the society
>>> should
>>> use what the technologists think is good for them.
>>>
>>
>>
>>
>> Let me push back very hard on this point.  I agree technology should
>> serve society, but there is an equal requirement in the other
>> direction that "society" choose sensible problems for technologists
>> to solve.  The idea that each country needs its own root server is an
>> entirely artificial and meaningless "requirement."  It has no
>> relationship to providing accurate, reliable, responsive Internet
>> service.  It is being driven solely by political forces for
>> appearances and ego.  I don't know whether such forces will
>> ultimately play a role in determining the architecture of the
>> Internet in the future, but I think we in the Internet Society have
>> an obligation to help the rest of the world understand the difference
>> between real and meaningful issues and those that don't serve any
>> larger socially useful purpose.
>>
>> Steve
>>
>
>
> -- 
> Franck Martin
> ICT Specialist
> franck at sopac.org
> SOPAC, Fiji
> GPG Key fingerprint = 44A4 8AE4 392A 3B92 FDF9  D9C6 BE79 9E60 81D9  
> 1320
> "Toute connaissance est une reponse a une question" G.Bachelard
>
>