[Dxspider-support] Add IP address tp PC16

[Brendan Minish]

On Mon, 2009-12-07 at 11:07 -0800, Lee Sawkins wrote:

>  
> Just when I think I have IPv4 figured out, they are going to change to
> IPv6.

Lee, 
IPv6 has been on the radar for well over 10 years. The issue is that the
pool of allocatable IPv4 addresses is running out. In  the next couple
of years it will become impossible for ISP's to acquire new IPv4 address
ranges for us to allocate to our customers. When this happens we have
basically 1 choice and that is IPv6 combined with Carrier grade NAT for
access to IPv4 only sites.

http://www.potaroo.net/tools/ipv4/index.html

http://en.wikipedia.org/wiki/IPv6

IPv6 is well supported by operating systems and applications, For ISP's
the business case for IPv6 is now becoming pressing and most ISP's that
I am aware of are either testing or beginning to roll out IPv6 

> Brendan you say having the Mac Address in IPv6 will help to narrow down
> the abusers.  Then you say tracking users will be more difficult.  This
> doesn't make sense to me.

IPv6 does not support NAT so each IPV6 allocation will be traceable down
to the end users individual machine. 
Dynamic allocation of IPv6 address ranges will be very uncommon for
normal end users, end users will have the same range assigned to them
for the duration of their ISP contract.
Associating an IPv6 range with one end user will be easy, in most cases
down to the individual machine in that allocation that is making the
connection. 
Local subnets are typically made up of either 48 or 64 bits allocated by
the ISP to the end user site. The last 64 to 96 bits of the address are
typically made up on auto-configured networks partially out of the mac
address of each individual machine, this is a globally unique IDent for
the Ethernet card installed within that machine. A typical address looks
like this 
2a02:3d8:1:0:215:17ff:fecd:eaca/64
this expands to 
2a02:3d8:0001:0000:0215:17ff:fecd:eaca/64
the MAC address on this machine is 00:15:17:cd:ea:ca

In an IPv6 world you will want to block a range, typically either a /64
or a /48 since the smallest end user allocation for a single site is
a /64 subnet (18,446,744,073,709,551,616 unique addresses)
Tracking miscreant users in an IPv6 world will be easier than it
currently is on the IPv4 internet. 
 
However carrier grade NAT will make much of what is currently done with
regard to tracking users by IPv4 address much harder since we ISP's will
have to nat traffic for thousands of customers out via a relativity
small pool of shared IPv4 addresses. We will simply have no choice in
the matter if we are to continue to give our users access to the IPv4
Internet. 

> 
> What I am currently doing with IPv4 seems to work fine.  I will continue
> to use it and modify it when I need to.  Hopefully I can continue this
> with IPv6.


ALL I am asking is simply that any new protocol enhancements designed to
share end user IP address assignments will work equally well with IPv6
addresses. It seems unwise to at this stage be creating a protocol that
cannot correctly handle sharing IPv6 addresses.
Any decision to actually support (or indeed not support) IPv6 is up to
individual software authors of course but it seems illogical at this
stage to design a new protocol extension that does not support sharing
IPv6 address information. 
Especially since Dxpsider already supports IPv6 and even has the
occasional IPv6 user 


> If Spider would include IP addresses with spots, these could also be
> checked for matching spotter country and IP country and dropped when
> they don't match. 

This is not a safe thing to do, here in Ireland users of one of our
large ISP's get allocated IP space allocated out of a pool that the same
ISP also uses in the UK.  
Irish (and presumably other EU users) Satellite internet users typically
get IP address space allocated out of pools allocated in Germany, The
Netherlands or Italy. 
Mobile users may either be natted (there are around 200 thousand users
of 3G internet access in Ireland and that majority of these will appear
to be coming from a few relatively small IPv4 ranges due to the
widespread use of NAT by our 3G operators) 
In a couple of years time it is possible we will see the remainder of
IPv4 space traded between operators which will lead to increasing
fragmentation of IPv4 space 



-- 
73
Brendan EI6IZ