Request for information/comment on default-free zone router properties

[Ivo Vachkov]

On Tue, Aug 30, 2011 at 10:05 AM, Andre Oppermann <andre at freebsd.org> wrote:

> On 29.08.2011 15:59, sthaug at nethelp.no wrote:
>
>> Is there any (no need to be official) information what is the number
>>> of different routes (for IPv4 and IPv6) on a default-free zone (DFZ)
>>> router in the Internet? I vaguely remember the number 450 000+
>>> distinct routes for IPv4? But what about IPv6?
>>>
>>
>> See http://www.cidr-report.org - it has all the information you need.
>>
>> Extract from the weekly post to Nanog:
>>
>> Recent Table History
>>         Date      Prefixes    CIDR Agg
>>         19-08-11    371450      219002
>>         20-08-11    371427      219147
>>         21-08-11    371547      219346
>>         22-08-11    371326      218957
>>         23-08-11    371090      219346
>>         24-08-11    371769      219465
>>         25-08-11    372189      219508
>>         26-08-11    372363      219490
>>
>
> For IPv6 the current number of prefixes is about 12451. Both IPv4
> and IPv6 numbers are expected to rise significantly in the future
> as more people migrate to IPv6, and the exhausted IPv4 pool gets
> fragmented more and more.


Yes, I believe the real challenge for effective routing will be the wide
adoption of IPv6. Even with prefixlen /32 the possible number of IPv6
routeable prefixes is greater than the number of routable IPv4 addresses
now.


>


>  On a related note, what is the number of the network interfaces on a
>>> DFZ router? No need for exact number, educated guess/estimation would
>>> be greatly appreciated too.
>>>
>>
>> How long is a piece of string? "It depends."  And the same is the case
>> for number of interfaces. You obviously need a minimum of three for a
>> router to do anything "interesting" with the packets. Also, it depends
>> on whether you're talking about physical interfaces or logical (sub)
>> interfaces. I'd say anything from 3 to 20 is fairly typical.
>>
>
> I'd say that range is about right for pure core/backbone routers.
> What happens more and more is that access concentrators (xDSL) run
> BGP as well. In that case the number of interfaces is 10k and more.
> Only a few of them 'run' BGP though.


My interest is purely academic. I would like to experiment with different
data structures to see if there is a way to increase routing performance
with large number of routes and interfaces.


>

-- 
> Andre
>

-- 
Ivo Vachkov