FW: [AFRINIC-rpd] Academic IPv4 Allocation Policy Second Draft (AFPUB-2013-GEN-001-DRAFT-02)

[Andrew Alston]

Hi All,

A post from Ben at Liquid...

Thanks

Andrew


-----Original Message-----
From: Ben Roberts [mailto:Ben.Roberts at liquidtelecom.com] 
Sent: Friday, January 25, 2013 5:28 PM
To: alston.networks at gmail.com
Subject: FW: [AFRINIC-rpd] Academic IPv4 Allocation Policy Second Draft
(AFPUB-2013-GEN-001-DRAFT-02)

Andrew,
Can you post this for me dude.  Having an offline discussion with SM.  I
cant seem to post on the list and my mails go to nowhere but I would
appreciate if you can post it up there on my behalf Chrs, Ben



Hi Ben,
At 00:43 25-01-2013, Ben Roberts wrote:
>But what has national/regional peering strategy got to do with Handing 
>out IP addresses? I just don't see the connection.

It is more about creating incentives and benefiting from the network effect.
Let's say that you get IP addresses for (example) one-tenth of the price on
the condition that you invest four-tenth of the price in peering.  You end
up with a 50% reduction and there is more traffic exchange within the
region.  Everyone gets more customers and less latency.

As I mentioned previously it is a bad idea for a lot of reasons, e.g.
making it more difficult to get IP addresses.

Regards,
-sm



SM,
Unfortunately I have problems posting to the list so can only reply to you
privately.

I totally disagree with this.  Afrinics role as I see it is to administer
the allocation of IP addresses to those African entities that need it.

Whether one of those African service providers join and participate in IXPs
in the territories where they operate is their own right to choose if they
want to or not as part of their business strategy.  It has nothing to do
with how many IP addresses they use or how much they pay Afrinic.

I don't understand this concept of 'price of IP addresses'.  Fees paid to
Afrinic are not to 'purchase' IP address space but to contribute to the
running of the agency that looks after IP address space in Africa.

When pricing up say a customer broadband service I do not even consider or
think about the cost of the IP addresses the customer will be using when
calculating what I must charge the customer to break even and the 'cost' of
the Ip address is one I consider trivial in the grand scheme of things.  I
say this coming from the point of view of a 'large' African service
provider.

I personally run a network that choses to use as many IXPs in Africa as
possible and fully optimise how much traffic I keep on continent, but that
is my choice and I believe it's a strategy that will win in the end over
those that chose not to peer.  But I would hate anyone telling me I must
peer at an IXP or even worse with all members of an IXP, and the costs of IP
addresses are so trivial to the the amount I spend on Opex that I wouldn't
even pay attention or think twice about offsetting Afrinic costs vs my
peering costs/gain.

The concept you suggest here is a) impossible to administer by Afrinic and
b) totally pointless as it will do nothing to influence operators peering
strategy.  My disagreement could not be any stronger.

But for end users of all kinds (large enterprises, universities, government
networks) they don't get money back from services sold with the IP addresses
they allocate, so I can see the point of view they should pay less than me.

Cheers
Ben

Regards,

Ben Roberts
Chief Technical Officer
Liquid Telecommunications

Liquid Telecommunications Limited, 9 Kingsway, London, WC2B 6XF
T: +44 (0) 20 7101 6100   -   D: +44 (0) 20 7101 6203   -   M: +44 (0) 7880
730 279   -   E: ben.roberts at liquidtelecom.com
W: www.liquidtelecom.com




-----Original Message-----
From: SM [mailto:sm at resistor.net]
Sent: 25 January 2013 14:55
To: Ben Roberts
Subject: Re: [AFRINIC-rpd] Academic IPv4 Allocation Policy Second Draft
(AFPUB-2013-GEN-001-DRAFT-02)

Hi Ben,
At 00:43 25-01-2013, Ben Roberts wrote:
>But what has national/regional peering strategy got to do with Handing 
>out IP addresses? I just don't see the connection.

It is more about creating incentives and benefiting from the network effect.
Let's say that you get IP addresses for (example) one-tenth of the price on
the condition that you invest four-tenth of the price in peering.  You end
up with a 50% reduction and there is more traffic exchange within the
region.  Everyone gets more customers and less latency.

As I mentioned previously it is a bad idea for a lot of reasons, e.g.
making it more difficult to get IP addresses.

Regards,
-sm


DISCLAIMER:  This email contains proprietary information some or all of
which may be legally privileged. It is for the intended recipient only. If
an addressing or transmission error has misdirected this email, please
notify the author by replying to this email. If you are not the intended
recipient, you must not use, disclose, copy, print, or rely on this email.
We cannot accept liability for any statements made which are clearly the
sender's own and not expressly made on behalf of this company or one of its
agents.