[ale] V6 question

Michael H. Warfield mhw at WittsEnd.com
Sat Feb 5 12:49:32 EST 2011


Minor nit.  I caught myself using incorrect terminology so, before
anyone else catches it and calls me on the table...

On Sat, 2011-02-05 at 12:23 -0500, Michael H. Warfield wrote: 
> Oh, where shall I begin.
> 
> First off...  My slide deck from my ALE presentation is here in several
> formats:
> 
> http://www.wittsend.com/mhw/2011/IPv6-BNW-ALE-2011.odp
> http://www.wittsend.com/mhw/2011/IPv6-BNW-ALE-2011.pdf
> http://www.wittsend.com/mhw/2011/IPv6-BNW-ALE-2011.ppt
> 
> Some of that might help.  Aaron informs me that he's tied up on a paid
> assignment (which absolutely takes priority in my book) for the next
> couple of weeks so it may be awhile before he has the recording ready.
> When he does, either he or I will post it as well.  That will help some
> more.
> 
> Now to try and answer your questions...
> 
> 
> On Sat, 2011-02-05 at 11:31 -0500, Jim Lynch wrote: 
> > I'm truly sorry to have missed the talks on IPV6.  So how is it going to 
> > replace NAT?
> 
> Simple answer.  It won't.  NAT (more specifically and precisely) NAT44
> (NAT IPv4:IPv4) will always be with us as long as IPv4 remains
> supported.  IPv6 does not have or support NAT.  It doesn't need it.
> 
> > I assume all the systems I have behind my router will have 
> > IPV6 addresses.  Is that correct?
> 
> That is correct.  Just as all your systems now have IPv4 addresses.
> Difference is, the v6 addresses of your systems will be global unicast
> addresses (in addition to a variety of multicast and link-local
> addresses) while the v4 systems have "private addresses" (in v6 land we
> call them site-local addresses).
> 
> Example...  On my laptop right now, here are my addresses: 
> 
> eth0      Link encap:Ethernet  HWaddr 00:24:7E:E1:2A:A7  
>           inet addr:130.205.38.43  Bcast:130.205.38.255  Mask:255.255.255.0
>           inet6 addr: 2001:4830:3000:8200:224:7eff:fee1:2aa7/64 Scope:Global
>           inet6 addr: fe80::224:7eff:fee1:2aa7/64 Scope:Link
> 
> The first one is my IPv4 address (I have lots of public addresses so I
> don't need to use NAT at all so that's a global v4 address in this
> case).  The second one is my IPv6 global address.  The third one is my
> link-local address (Note the "Scope:" field).
> 
> > Is DHCP going away?
> 
> It might.  It might not.
> 
> IPv4 has two methods for configuring addreses.  Static, and stateful
> autoconfiguration (dhcp).  IPv6 has those plus it adds stateless
> autoconfiguration.
> 
> Static works the same way in both.  You code the addresses into the
> system.
> 
> Stateful works in a similar way.  On IPv4, it's dhcp.  On IPv6 it's
> dhcp6.  So no, if you want to use stateful autoconfiguration, you just
> run a new dhcp daemon only on the v6 service.  On IPv6, dhcp does work a
> little different than on IPv4 because dhcp6 uses multicast addressing
> (IPv6 has no broadcast addresses at all) but the principle remains the
> same.
> 
> Stateless autoconfiguration involves router advertisements and router
> discovery along with neighbor discovery.  You can think of these sorts
> of things is like "arp on steriods" extended to cover things arp
> doesn't.  If you look at my addresses above, you'll see the numbers
> after the "fe80::" are identical to the numbers after the
> "2001:4830:3000:8200:".  That my EUI (End Unit Identifier) or host
> identifier field.  That's the local part of your address and the system
> automatically computes this from your MAC (HWaddr above).  Compare them.
> You can see the similarity.
> 
> So, where did the "2001:4830:3000:8200" come from?  It came from the
> router.  The router is broadcasting this prefix periodically and when a
> system first comes up, it sends a request to the "all routers" multicast
> address and solicits the routers to send it a prefix.  It takes that
> prefix and combines it with its EUI and voila, you have a global
> address.  So, if you want to go that route, no, you don't need dhcp.

Correction to the above "The router is broadcasting".  As I stated
elsewhere, there is no broadcast address.  The correct precise
terminology would be this:

The router is sending this prefix to the "all nodes multicast address"
periodically and when a system first comes up, it sends a request to the
"all routers" multicast address and solicits the routers to send it a
prefix.  The "all nodes" address is "ff02::1" while the "all routers"
address is "ff02::2" (the dhcp6 multicast address is ff02::1:2 or
ff05::1:3)

> > So is the port 
> > the ISP furnishes me going to be just a connection to the wan without a 
> > IP address?
> 
> No.  They will assign you a prefix.  I now understand that Comcast is
> handing out /64 prefixes (single subnet) to customers participating in
> their beta rollout.  In principle, if you have multiple subnets, you
> were originally suppose to get a /48 (that's 65,536 subnets for you) but
> that's a bit much for residential customers.  Freenet6 hands out
> free /56 networks (256 subnets).  I don't know how Comcast is going to
> deal with that down the road.  Maybe a request or on demand for subnets.
> 
> Once you get a prefix, your router will advertise that prefix and all
> your machines will number themselves.
> 
> > I'm confused.
> 
> > Jim.
> 
> Regards,
> Mike

-- 
Michael H. Warfield (AI4NB) | (770) 985-6132 |  mhw at WittsEnd.com
   /\/\|=mhw=|\/\/          | (678) 463-0932 |  http://www.wittsend.com/mhw/
   NIC whois: MHW9          | An optimist believes we live in the best of all
 PGP Key: 0x674627FF        | possible worlds.  A pessimist is sure of it!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 482 bytes
Desc: This is a digitally signed message part
Url : http://mail.ale.org/pipermail/ale/attachments/20110205/6bc6645f/attachment-0001.bin 


More information about the Ale mailing list