[discuss] Real world Impact of multiple roots

Brian E Carpenter brian.e.carpenter at gmail.com
Tue Jan 28 03:12:30 UTC 2014


On 28/01/2014 12:29, Louis Pouzin (well) wrote:
> Hi Brian,
> 
> .. and now we have FB, a never debugged Gmail, Google+, LinkedIn, Twitter,
> dozens of wikis, etc. etc.
> Isn't ever more chaotic, and balkanized ?

True, but that is at the level of information and human interaction.
To me it is like trying to listen to all the conversations in a busy
restaurant: a lot of confusion, but food and drink is still delivered
to the correct tables in a systematic way.

    Brian

> 
> Louis
> - - -
> 
> On Mon, Jan 27, 2014 at 11:13 PM, Brian E Carpenter <
> brian.e.carpenter at gmail.com> wrote:
> 
>> On 28/01/2014 05:38, Dr. Ben Fuller wrote:
>>> All,
>>>
>>> Thanks for the replies. The answers seem to confirm my suspicion that an
>> Internet with two or more roots will have major impacts to business and
>> society.
>>
>> Way back in history, there was a time when many different email systems
>> were in use, and there was no global DNS. At that time, some of us had
>> to deal with the situation, since a balkanized email world was no use
>> to us. The consequences were multiple, including at least
>>
>> 1. The need to pay for highly skilled staff and additional equipment
>> to implement and operate multi-protocol mail gateways.
>>
>> 2. The need for end users to understand details of various email
>> addressing schemes, and in some cases to compose ad hoc addresses
>> (which in my case usually included !mcvax, !unido or !seismo
>> as well as !cernvax, not to mention things like %bitnet). In this case
>> it was the lack of a single root for a single namespace that actually
>> forced the end user to understand routes. No pun intended.
>>
>> 3. Frequent long delays and delivery failures.
>>
>> So, if we had a balkanized DNS namespace, I'm sure we'd figure
>> out ways round it, but I'd expect issues like the above to return,
>> for all services, and enormous costs and lost business as a result.
>>
>>     Brian
>>
> 



More information about the discuss mailing list