[rpd] Statistics on IPV4 allocation in Africa as of 2016

[Nishal Goburdhan]

On 29 Aug 2016, at 18:32, sm+afrinic at elandsys.com wrote:

> The article does not discuss about why the network was down for two 
> days as that happened afterwards.  I would not describe an IX handling 
> the level of traffic of one FTTH user as functional.

i would describe the IX with at least three peers, that’s happily 
passing the frames of willing participants as functional.
if you don’t like your ISPs policies, vote with your wallet..


> I would also not say that for an IX which is not keeping local traffic 
> local as I do not wish to be the one explaining about that if there is 
> an inquiry about the funds which were spent on the IX.

to be clear, you do realise that it takes *ISPs* that are willing to 
peer, to “keep traffic local”.  if a particular ISP, does not want 
to peer, the IXP isn’t going to magically fix your packet’s 
trajectory.

i’m happy to argue the specifics of how peering works on a more 
appropriate list.


> Should a /22 (IPv4) be reserved for a white elephant? :-)

https://lg.mixp.org tells me that there’s network prefixes being 
announced at the IXP.
statistics from 	http://www.mixp.org/#statistics tell me that there’s 
live traffic going across the IXP.
the members list at http://www.mixp.org/#members shows me what looks 
like likely real peers.

your complaint seems to be about one ISP’s peering policy.  it’s 
probably best to take that up with the ISPs directly.  i don’t think 
RPD can assist you, though.

i wasn’t aware that the MIXP had a /22 reserved.
i wouldn’t call the MIXP a white elephant.

—n.