From: ccie2be (ccie2be@nyc.rr.com)
Date: Fri Jun 03 2005 - 10:52:39 GMT-3
Hi guys,
I'm making some progress on understanding the config in IE Vol 2, lab 8.
Near the bottom of the post below, you'll see that I didn't understand why
the following config commands were entered on the FA:
ip irdp maxadvertinterval 0
ip irdp minadvertinterval 0
ip irdp holdtime 0
Well, I can now tell you what they do although I'm still not sure they're
needed to fulfill the requirements of this task. By setting all these
values to 0, the FA will wait until it hears a solicitation from the MN
before it sends any FA's advertisements.
This is according to the Cisco Press book on Mobile IP which I highly
recommend. For more info, see Chapter 7, page 161.
BTW, this chapter covers another aspect of the mobile ip tasks in this lab.
Until reading this chapter, I knew that a Cisco router could be configured
to behave as a mobile node but I thought that was the extent of it. And,
the purpose of this was only for testing Mobile IP which I saw could be
especially useful in the lab.
Well, little did I know.
Besides testing Mobile IP, a cisco router can be configured as a MN to allow
connectivity to stub networks attached to it from the other side of the FA
connection. When used this way, the MN becomes a Mobile Router (MR) which
connects mobile networks. I'll try to create a picture of this.
mobile net's <----> MR <----> FA
This topology provides some very interesting possibilities. Imagine a
Boeing 747 flying around or an aircraft carrier sailing the seas. On board
is a MR that connects to a number of stub subnets. This connection can be
wired or wireless - it doesn't matter. However, hosts connected to this MR
don't have any idea that their default-gateway is a MR and therefore they
don't need to have any mobile IP client s/w installed - this mobile ip
business is completely transparent to these hosts. Cool, huh?
As before with MN's, for the MR to connect to the rest of the network, it
needs to attach to a FA.
To make this all work, no special configuration is required on the FA
although some additional options exist. And, it looks like additional
configuration is needed on the HA but I haven't finished reading that part
yet so I can't say.
As always, feel free to comment or correct any mistakes I've made.
Tim
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
ccie2be
Sent: Thursday, June 02, 2005 6:22 PM
To: Group Study
Cc: 'John Matus'
Subject: Mobile IP
Hi guys,
For anyone trying to get up to speed with this new topic, in the new IE Vol
2 workbook, there's an excellent set of tasks that deal with this.
John,
This lab also has a good example of how and why to use virtual networks.
Seeing this example also helped me to see why virtual networks are redist
into the IGP. If you recall, I said it might be possible to have the IGP
include the virtual network in its routing process. I was thinking along
the lines of using network x.x.x.x m.m.m.m under the routing process for
this.
I now think that's not possible for this reason. A virtual network is
analogous to a logical interface like a loopback int in the sense that
neither has physical existence. They both just exist in router memory.
But, still a network isn't the same as an interface. And so, if your
virtual network were say, 12.0.0.0/24, and you added the command network
12.0.0.0 0.0.0.255 under the eigrp process, the eigrp process wouldn't find
any interface, logical or physical, on which it finds 12.0.0.0 running.
And, so, that's why, I think, virtual networks have to always be redist into
the IGP process.
Now, I'm still trying to fully understand the reason and function of every
command used in this set of tasks and I'm sure I'll get there but I'm not
there yet.
For example, for those that have been following the mobile ip thread here on
GS, you all may recall that IRDP must be configured on the router configured
as a FA but not on the HA. (Scott Morris explains why this is.)
In this task, the requirement is to minimize convergence time and configure
the router acting as a MN to solicit roaming requests every second.
So, as you would expect, there's a command that can be configured on the MN
to do this. The interface command is:
ip mobile router-service solicit interval 1
Now, here's the part I didn't expect.
On the FA's, not only is IRDP enable, but there are a bunch of other IRDP
commands used as well.
ip irdp maxadvertinterval 0
ip irdp minadvertinterval 0
ip irdp holdtime 0
After checking with the Doc-CD, these values aren't the default values.
See,
http://www.cisco.com/univercd/cc/td/doc/product/software/ios123/123cgcr/ipra
s_r/ip1_i1g.htm#wp1081946
So, my guess is that these values were put there to meet the "minimize
convergence" requirement but I don't as yet understand this interaction.
And, I'm not at all sure these commands are really needed.
I imagine the MN host will send out it's RRQ's (solicitation requests) every
second regardless of the irdp settings and it's that interval, not the
intervals used by IRDP that determine how long convergence takes. But,
that's just my guess.
One other thing I found very interesting in this example was that the HA
home address can be explicitly defined. I didn't know that. Another person
on GS had pointed out to me that the home address used by the MN can be any
ip address on the HA.. So, here's a little twist on that.
Anyway, if anyone has any comments or corrections, please don't hesitate to
speak up.
This archive was generated by hypermail 2.1.4 : Wed Jul 06 2005 - 14:43:40 GMT-3