Re: [RRG] Moving forward... IPv4 now, IPv6 less urgent and perhaps more ambitious

[Brian E Carpenter <brian.e.carpenter@gmail.com>]

On 2008-06-11 02:44, Marshall Eubanks wrote:
> 
> On Jun 10, 2008, at 10:14 AM, William Herrin wrote:
> 
>> On Tue, Jun 10, 2008 at 9:41 AM, David Conrad <drc@virtualized.org>
>> wrote:
>>> On Jun 9, 2008, at 8:07 PM, William Herrin wrote:
>>>> The identified problem is:
>>>
>>>> find a way to drastically reduce the $8000/year cost of each IPv4
>>>> prefix in the core.
>>>
>>> I'd be curious as to how you derived this number.
>>
>> Hi David,
>>
>> http://bill.herrin.us/network/bgpcost.html
>>
>> The source numbers change with time. Feel free to plug any realistic
>> source numbers into the formula that you please. The end result will
>> have the same impact: the cost is nontrivial.
>>
>> If you want to challenge the methodology, I respectfully ask that you
>> find someone who does cost analysis for a living to assist you with
>> the particulars. I had a professional cost analyst review and validate
>> it. Same guy who got the V22 Osprey canceled before Congress put it
>> back.
> 
> Note that another hard to estimate number is the economic benefit of the
> Internet, but it
> is reasonable to estimate this as at least $ 200 billion / year, but
> probably 2 or 3 times that.
> The direct revenue estimates I could find were all > $ 200 billion / year.
> 
> So, the $ 2 billion / year to run BGP that you estimate is a cost of
> doing business no more than 1% of revenue,
> and probably more like 0.3 to 0.5% of Internet revenue, aggregated
> globally.  Non-trivial, to be sure, but that level does  not seem
> unsustainable or even unreasonable.
> 
> To put it another way, that prefix is worth the money.

Of course it is, or the capitalist nature of the Internet would have
caused that prefix to be filtered out. There is an incentive for
all ISPs to share the externality costs of BGP advertisements from
other ISPs - the fact that connectivity is its own reward.

But common sense suggests that the externality cost *per customer*
of an aggregated prefix is lower than for a deaggregated prefix.
So I think Bill's average rate of $8k/prefix (whether it's correct
or not) isn't very helpful, because it conceals the different values
of aggregated and deaggregated prefixes. We need to look at
cost per customer.

    Brian

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg