pTLA request SSVL - review closes 3 May 2002

[Bjorn Pehrson]

Pekka,
I would like to stress that we are sponsored by and work with industry 
and other external organizations, as well as with other universities in 
a global context. It is not a matter just of ipv6 connectivity at 
Swedish universities. And there are many other issues than BGP 
behaviour. We believe that we have the competence and resources to do 
this research in a responsible way, given the access to the appropriate 
mechanisms.
Bjorn

Pekka Savola wrote:

>On Sat, 20 Apr 2002, Bjorn Pehrson wrote:
>
>>Yes, SSVL is an independent research network. I know of other such 
>>networks although it is fairly few, not thousands. Since 1996, it is 
>>used by several universities and non-academic organizations to do 
>>various types of research that are impossible or hard to do in regular 
>>networks. Due to the topology of the network, AS8973 was granted us in 
>>1998. Due to the nature of our experiments, we moved from KTH address 
>>space to 192.16.124.0/22 in 2000, not to cause problems with aggregation.
>>
>[...]
>
>>Sunet do provide addresses, altough they do not yet forward packets, 
>>but, as far as I know, they cannot provide a TLA-prefix for us.
>>
>
>I can feel for you -- life is rather difficult w.r.t. IPv6 connectivity in
>Swedish Universities, I imagine.  Which is partially why you seem to have
>chosen to go this route.
>
>This surely would get IPv6 better in the Swedish community.. but.. is 
>opening this can of worms _in the global scope_ worth the cost..
>
>>Since some time, the focus has shifted towards ipv6 interchange and 
>>global  interdomain issues which require access to the TLA level. I do 
>>consider this a valid reason.
>>
>
>Usually these kind of more or less global tests are done either:
> - in co-operation with someone with possibility to do them, in your case 
>SUNET
> - in a test network
>
>Consider: how would you feel about researches applying for ASN and address 
>block so they could experiment, research, etc. BGP behaviour in the 
>global scope?  The approach and the tools would seem to be wrong.
>