From: James (james@towardex.com)
Date: Sun Sep 26 2004 - 06:01:21 GMT-3
On Sun, Sep 26, 2004 at 01:34:40AM -0500, Jonathan R. Charles wrote:
> A friend of a friend said he was at Networkers and Cisco said it would only
> be a very basic configuration on one link, where you would have to apply the
> address and get connectivity going, nothing major, no routing protocols, no
> interlacing of IPv4 and IPv6 for all your BGP speakers.
That's very simple then. But as everyone suggests -- anything is a fair game!
>
> Remember IPv6 implementation is not around the corner, it is not even in
> this time zone for the general networking community.
I wouldn't exactly say that IPv6 ipmlementation is not around the corner.
As far as infrastructure and foundation technology goes, 6bone was the
most successful project in giving networks operational experience with IPv6.
As far as implementations go, Cisco, Juniper, and couple other major network
equipment vendors all support IPv6 solidly now with their software and new
ASIC hardware for wire-speed routing. The support for IPv6 is *there* and
it is strong. People like Verio and Global Crossing now have their entire
AS running full international native IPv6 backbone, without use of any
makeshift tunnels at the core backbone level (although there are still lots
of tunnels to customers and some peers due to cost restrictions).
The problem with IPv6 today however is not the equipment of network vendors.
It is simply that there is no content in IPv6 world right now. Customers
don't see IPv6 useful until they see some serious content to be useful, and
content providers don't see IPv6 useful until there is business model and/or
demand for it by majority of their customers. Hence, chicken and egg type of
problem. So I wouldn't say it is the implementation that is not around the
corner.. There are implementation issues with IPv6 with respect to multihoming
and with upper layer protocols such as support for IPv6 by end-user applications
and what not, but as far as network layer goes, IPv6 I'd say is very well
supported by networking vendors.
However, one thing for sure is that, as observed by implementations at
DREN and couple other IPv6 networks, once you build an infrastructure, some
people will naturally come to make use of it. And it will eventually drive
the demand later -- that seems to be the trend with most ISP's experimenting
with IPv6 on their network.
>
> I suspect highly that the configuration will be simple, basic and probably
> only worth 2 points or so.
>
> While we are on the topic, what is this whole 'address-family' thing for
> mBGP referring to? Would you subconfigure two routing processes inside the
> BGP AS, one for IPv4 and the other for IPv6?
That is effectively correct. The reason why this whole mBGP stuff came about
for IPv6 was that it had been decided that there should be IPv6 support for
existing BGP4 implementation through use of the multi-protocol extensions.
You may find RFC2545 interesting to read just to see where it came about..
If you wish to see a configuration tutorial about IPv6 MBGP used in real
world, in both Cisco* and Juniper formats try this URL:
http://www.occaid.org/tutorial-ipv6bgp.html
For anyone interested, I can also provide real world examples of v6
configurations taken off of production routers if requested off-list.
>
> Also, what ever happened to IPv5? Was it a complete disaster?
Never going to happen ;) Howard can probably fill us in with more details
on this with respect to what actually happened during earlier approaches
in the industry.. My knowledge on history of IPv6 technology since its
days in early 6bone-era (circa 1996) is quite weak.
HTH,
-J
-- James Jun TowardEX Technologies, Inc. Technical Lead Network Design, Consulting, IT Outsourcing james@towardex.com Boston-based Colocation & Bandwidth Services cell: 1(978)-394-2867 web: http://www.towardex.com , noc: www.twdx.net
This archive was generated by hypermail 2.1.4 : Fri Oct 01 2004 - 15:00:49 GMT-3