2004-02-04
Revision History | ||
---|---|---|
Revision 1.1 | 2004-02-04 | Revised by: LKS |
Added "Travelling through several foregin LAN's" and "Returning home". Some cleanup and restructuring. | ||
Revision 1.0 | 2003-12-02 | Revised by: TMM |
Reviewed by LDP | ||
Revision 0.5.2 | 2003-11-26 | Revised by: LKS |
A lot of cleanup. Thanks to John Levon levon [at] movementarian.org | ||
Revision 0.5.1 | 2003-11-22 | Revised by: LKS |
Changed the license from GFDL to OPL due to some GFDL problems. | ||
Revision 0.5 | 2003-11-18 | Revised by: LKS |
Converted to XML Docbook. Some cleanup. | ||
Revision 0.4 | 2002-11-07 | Revised by: LKS |
Fixed some errors + update. Thanks to Henrik Petander petander (at) tcs hut fi. | ||
Revision 0.3.1 | 2003-11-03 | Revised by: LKS |
Updated to MIPL relase 1.0 (kernel 2.4.22). | ||
Revision 0.3 | 2003-08-05 | Revised by: LKS |
Initial release. |
This document describes the software and procedures to set up and use mobile IPv6 for Linux. The "Mobility Support in IPv6" draft answers the what and why of mobile IP:
"Each mobile node is always identified by its home address, regardless of its current point of attachment to the Internet. While situated away from its home, a mobile node is also associated with a care-of address, which provides information about the mobile node's current location. IPv6 packets addressed to a mobile node's home address are transparently routed to its care-of address via the mobile nodes Home Agent (HA). The protocol enables IPv6 nodes to cache the binding of a mobile node's home address with its care-of address, and then to send any packets destined for the mobile node directly to it at this care-of address." --- draft-ietf-mipv6-24, page 1-2.
"Without specific support for mobility in IPv6, packets destined to a mobile node (host or router) would not be able to reach it while the mobile node is away from its home link (the link on which its home IPv6 subnet prefix is in use), since routing is based on the subnet prefix in a packet's destination IP address. In order to continue communication in spite of its movement, a mobile node could change its IP address each time it moves to a new link, but the mobile node would then not be able to maintain transport and higher-layer connections when it changes location. Mobility support in IPv6 is particularly important, as mobile computers are likely to account for a majority or at least a substantial fraction of the population of the Internet during the lifetime of IPv6." --- draft-ietf-mipv6-24, page 6.
For all the details, read the "Mobility Support in IPv6" draft
Mobile IP
The Mobile Node (MN) travels to a foreign network and gets a new care-of-address.
The MN performs a binding update to its Home Agent (HA) (the new care-of-address gets registered at HA). HA sends a binding acknowledgement to MN.
A Correspondent Node (CN) wants to contact the MN. The HA intercepts packets destined to the MN.
The HA then tunnels all packets to the MN from the CN using MN's care-of-address.
When the MN answers the CN, it may use its current care-of-address (and perform a binding to the CN) and communicate with the CN directly (optimized routing) or it can tunnel all its packets through the HA.
See figure "Mobile IP" for an explanation.
IP version 6 (IPv6) is a new version of the Internet Protocol, designed as the successor to IP version 4 (IPv4) [RFC-791]. The changes from IPv4 to IPv6 fall primarily into the following categories:
Expanded addressing capabilities
Header format simplification
Improved support for extensions and options
Flow labeling capability
Authentication and privacy capabilities
You should have basic knowledge of IPv6 stateless auto-configuring to fully understand how 'mobile IPv6' (MIPv6) works. You can read up on IPv6 Stateless Address Autoconfiguration in [RFC2462].
For more information on IPv6 in general, visit the IETF's IPv6 Working Group.
There are currently two Mobile IPv6 Linux implementations available. The Lancaster University in the UK has the oldest(?) implementation ( http://www.cs-ipv6.lancs.ac.uk/MobileIP/). The latest kernel supported is 2.1.90, and is compatible with IETF mobile IPv6 draft-v5 (the current revision is v24). The code and website has not been updated since 1998, so it is considered obsolete.
The other implementation, which is up-to-date, is Helsinki University of Technology's MIPL project. The latest supported kernel is 2.4.22, and they have patches for the upcoming 2.6 kernel (see the FAQ). Visit http://www.mipl.mediapoli.com/ for papers, software or to browse the mail archive.
The MIPL MIPv6 implementation requires a kernel patch. The implementation modifies the IPv6 kernel stack, so a kernel recompile is necessary. The installation process is well documented, but I will give a brief step-by-step howto.
Please note! The need for two different kernels, one for MN and one for HA, is obsolete. Just compile support for MN and HA in the same kernel. It is not possible to run as both an MN and an HA at the same time; which mode is chosen depends on which of the modules are loaded.
Download the latest Linux MIPv6 source code from http://www.mipl.mediapoli.com. The latest release today is: mipv6-1.0-v2.4.22. The last four numbers corresponds to the Linux kernel the patch should be applied to:
# cd /usr/local/src # wget http://www.mipl.mediapoli.com/download/mipv6-1.0-v2.4.22.tar.gz # tar zxfv mipv6-1.0-v2.4.22.tar.gz |
Download and unpack the correspondent Linux kernel version from ftp.kernel.org:
# cd /usr/src # wget ftp://ftp.kernel.org/pub/linux/kernel/v2.4/linux-2.4.22.tar.bz2 # tar jxvf linux-2.4.22.tar.bz2 # ln -s linux-2.4.22 linux # cd linux |
Apply the MIPv6 patch:
# patch -p1 --dry-run < /usr/local/src/mipv6-1.0-v2.4.22/mipv6-1.0-v2.4.22.patch |
The --dry-run option checks that the patch will apply correctly. If you get any failed hunks, you should not proceed. If everything went fine do:
# patch -p1 < /usr/local/src/mipv6-1.0-v2.4.22/mipv6-1.0-v2.4.22.patch |
Now your kernel tree is ready for configuration. Run your favorite make *config. The MIPv6 options are under "Networking Options". The following options should be present in ".config":
CONFIG_EXPERIMENTAL=y CONFIG_SYSCTL=y CONFIG_PROC_FS=y CONFIG_MODULES=y CONFIG_NET=y CONFIG_NETFILTER=y CONFIG_UNIX=y CONFIG_INET=y CONFIG_IPV6=m CONFIG_IPV6_SUBTREES=y CONFIG_IPV6_IPV6_TUNNEL=m CONFIG_IPV6_MOBILITY=m CONFIG_IPV6_MOBILITY_MN=m CONFIG_IPV6_MOBILITY_HA=m |
Since MIPL is still a work-in-progress you might want to enable:
CONFIG_IPV6_MOBILITY_DEBUG=y |
With debug messages it is easier to figure out what happened when something goes wrong. Also, when reporting a bug, debug messages are very helpful.
To be sure you have all the correct options, you can run chkconf_kernel.sh, which is a small shell script included in the MIPL tarball.
Next you should compile and install your kernel.
Hint: To easily distinguish this kernel from other kernels, you can change the "EXTRAVERSION" variable in the /usr/src/linux/Makefile to for example "-MIPv6-1".
Read the Linux Kernel HOWTO for detailed instruction on how to patch, compile and install your new kernel.
The userspace tool mipdiag, config files and init scripts must be installed for the module to work correctly:
# cd /usr/local/src/mipv6-1.0-v2.4.22 # ./configure # make && make install |
The MIPv6 module also needs a new device node entry. Issue the command:
# mknod /dev/mipv6_dev c 0xf9 0 |
Red Hat:
All init scripts are located in /etc/init.d/, which are sym-linked to the correct runlevel (/etc/rcX.d/). You can issue the command:
# chkconfig --add mobile-ip6 |
to enable MIPv6 at startup, or
# chkconfig --del mobile-ip6 |
to remove MIPv6 from startup.
Debian:
If you are so lucky to be running Debian, you can issue the command:
# update-rc.d -n mobile-ip6 start 75 3 4 5 . stop 05 1 2 6 . |
to set up all the necessary links.
Slackware:
Slackware users have all their startup/runlevel scripts in /etc/rc.d. Since 'configure' doesn't check for "/etc/rc.d", you can add INIT_SLACK="/etc/rc.d", and then INIT_SLACK to INITDIRS in 'configure' (search for INITDIR in configure). Since you are running Slackware, you probably know this already. The following command should then do the trick:
# echo '/etc/rc.d/mobile-ip6 start' >> /etc/rc.d/rc.local |
If you don't hack the Makefile, the mobile-ip6 script is installed at '/' (you may then move it to /etc/rc.d/).
Now you should have a working MIPL patched kernel, installed userlevel tools and enabled automatic startup at boot. If anything went wrong, go through the above sections carefully.
The addresses we are using in our test-bed are site-local. You may as well use global addresses, but do note that link local addresses won't work! Our test-bed consist of four nodes; see figure "Mobile IPv6 testbed".
HA - Home Agent: The HA is located at the home network with address fec0:106:2700::2, with one wireless interface.
MN - Mobile Node: When MN is on the "home network", it has address fec0:106:2700::4. When MN travels to another network, it generates a new "care-of" address.
R - Router: This is the router from the home network to the internet. It has one wireless interface with address fec0:106:2700::1 and a wired interface with address fec0:106:2300::2.
AR - Access Router: The link between AR and R is our "internet" - but in this testcase only a cross-cable (can be any network). The AR has two interfaces; the wired interface to R has address fec0:106:2300::1, the wireless has address fec0:106:1100::1.
Mobile IPv6 testbed
Before we can start testing mobile IP, we need a fully functional IPv6 network. All the nodes should be able to ping each other. This is a crucial part. If, for example, AR is not able to ping HA, then there will be no binding update.
I will give a brief instruction to get our network up and running using IPv6. For more info on setting up an IPv6 network, you can read Peter Bieringer's excellent Linux IPv6 HOWTO.
I've turned off encryption for simplicity - NOTE that you should ALWAYS use encryption when dealing with wireless networks!
Also note that the different wireless networks have different ESSIDs!
MN: The Mobile Node has one wireless interface. Forwarding should be turned off, but should accept autoconf and ra's:
# iwconfig eth0 mode ad-hoc essid homenet enc off # ifconfig eth0 inet6 add fec0:106:2700::4/64 # echo "0" > /proc/sys/net/ipv6/conf/eth0/forwarding # echo "1" > /proc/sys/net/ipv6/conf/eth0/autoconf # echo "1" > /proc/sys/net/ipv6/conf/eth0/accept_ra # echo "1" > /proc/sys/net/ipv6/conf/eth0/accept_redirects # /etc/init.d/mobile-ip6 start |
HA: The Home Agent has one wireless interface. It should have forwarding turned on because it uses normal routing to deliver packets captured from a physical interface to the virtual tunnel interface. Note: You must add a default route or else HA will have problem contacting the MN on visited LAN's. One possible solution is to use HA as the default router of the home network.
# iwconfig eth0 mode ad-hoc essid homenet enc off # ifconfig eth0 inet6 add fec0:106:2700::2/64 # echo "1" > /proc/sys/net/ipv6/conf/eth0/forwarding # echo "0" > /proc/sys/net/ipv6/conf/eth0/autoconf # echo "0" > /proc/sys/net/ipv6/conf/eth0/accept_ra # echo "0" > /proc/sys/net/ipv6/conf/eth0/accept_redirects # ip route add ::/0 via fec0:106:2700::1 # /etc/init.d/mobile-ip6 start |
R: The (home) Router has two interfaces; one wireless and one line. The Router must have forwarding turned on.
# ifconfig eth0 inet6 add fec0:106:2300::2/64 # iwconfig eth1 mode ad-hoc essid homenet enc off # ifconfig eth1 inet6 add fec0:106:2700::1/64 # echo "1" > /proc/sys/net/ipv6/conf/all/forwarding # echo "0" > /proc/sys/net/ipv6/conf/all/autoconf # echo "0" > /proc/sys/net/ipv6/conf/all/accept_ra # echo "0" > /proc/sys/net/ipv6/conf/all/accept_redirects # ip route add fec0:106:1100::/64 via fec0:106:2300::1 |
AR: The Access Router (on a foreign network) also has two interfaces; one wireless and one line. Forwarding must be turned on.
# ifconfig eth0 inet6 add fec0:106:2300::1/64 # iwconfig eth1 mode ad-hoc essid visitnet enc off # ifconfig eth1 inet6 add fec0:106:1100::1/64 # echo "1" > /proc/sys/net/ipv6/conf/all/forwarding # echo "0" > /proc/sys/net/ipv6/conf/all/autoconf # echo "0" > /proc/sys/net/ipv6/conf/all/accept_ra # echo "0" > /proc/sys/net/ipv6/conf/all/accept_redirects # ip route add fec0:106:2700::/64 via fec0:106:2300::2 |
Instead of modifying proc variables, you can use sysctl.
Note: We are setting static routes on our test-bed. You should now be able to ping all the hosts from every host.
The last configuration is MIPv6 settings in network-mip6.conf. In Debian/Slackware the file is found under /etc/. (RedHat the file is found under /etc/sysconfig/.) The file should be pretty self-explanatory.
HA: The HA config file should contain these settings:
# cat /etc/network-mip6.conf # Home Agent configuration file FUNCTIONALITY=ha DEBUGLEVEL=1 MIN_TUNNEL_NR=1 MAX_TUNNEL_NR=5 TUNNEL_SITELOCAL=yes |
MN: The MN config file should look like this:
# cat /etc/network-mip6.conf # Mobile Node configuration file FUNCTIONALITY=mn DEBUGLEVEL=1 TUNNEL_SITELOCAL=yes MIN_TUNNEL_NR=1 MAX_TUNNEL_NR=3 HOMEDEV=mip6mnha1 HOMEADDRESS=fec0:106:2700::4/64 # MN's home adress HOMEAGENT=fec0:106:2700::2/64 # HA's address |
Next, start mobile-IP:
# /etc/init.d/mobile-ip6 start Starting Mobile IPv6: OK |
You can verify that it started by doing a ifconfig on HA. If the tunnel(s) comes up, ip6tnl1, mobile-ip6 is started:
You will also see the mipv6 kernel modules are loaded (MN):
# lsmod Module Size Used by Not tainted mip6_mn 59888 0 (unused) ipv6_tunnel 11448 1 [mip6_mn] mip6_base 40728 0 [mip6_mn] ipv6 179764 -1 [mip6_mn ipv6_tunnel mip6_base] ... |
When MN comes to a new network, it does a link-local address configuration, going to the next phase if that succeeds. I'll let [RFC2462] (IPv6 Stateless Address Autoconfiguration) describe the next phase:
"The next phase of autoconfiguration involves obtaining a Router Advertisement or determining that no routers are present. If routers are present, they will send Router Advertisements that specify what sort of autoconfiguration a host should do. If no routers are present, stateful autoconfiguration should be invoked."
"Routers send Router Advertisements periodically, but the delay between successive advertisements will generally be longer than a host performing autoconfiguration will want to wait. To obtain an advertisement quickly, a host sends one or more Router Solicitations to the all-routers multicast group." --- page 8
This is where we use RADVD.
Read [RFC2462] more more details concerning IPv6 Stateless Address Autoconfiguration.
We'll configure RADVD on AR's wireless interface. The radvd.conf file should contain this:
# cat /etc/radvd.conf interface eth1 { AdvSendAdvert on; AdvIntervalOpt on; MinRtrAdvInterval 3; MaxRtrAdvInterval 10; AdvHomeAgentFlag off; prefix fec0:106:1100::/64 { AdvOnLink on; AdvAutonomous on; AdvRouterAddr on; }; }; |
We then start it:
# /etc/init.d/radvd start |
You should now be able to use radvdump to see that the radvd messages really are being sent periodically:
# radvdump Router advertisement from fe80::202:2dff:fe54:d1b2 (hoplimit 255) Received by interface eth1 # Note: {Min,Max}RtrAdvInterval cannot be obtained with radvdump AdvCurHopLimit: 64 AdvManagedFlag: off AdvOtherConfigFlag: off AdvHomeAgentFlag: off AdvReachableTime: 0 AdvRetransTimer: 0 Prefix fec0:106:1100::/64 AdvValidLifetime: 2592000 AdvPreferredLifetime: 604800 AdvOnLink: on AdvAutonomous: on AdvRouterAddr: off AdvSourceLLAddress: 00 02 2D 54 D1 B2 |
Note! When using radvd on HA and enabling "autoconf" (in proc), you will also get an autogenerated IPv6 address on MN (which is superfluous) in addition to your static address:
To enable the MN to know when it's home, HA should also be sending out RAs. We should therefore enable RADVD on the HA as well. The /etc/radvd.conf file should contain:
# cat /etc/radvd.conf interface eth0 { AdvSendAdvert on; MaxRtrAdvInterval 3; MinRtrAdvInterval 1; AdvIntervalOpt off; AdvHomeAgentFlag on; HomeAgentLifetime 10000; HomeAgentPreference 20; AdvHomeAgentInfo on; prefix fec0:106:2700::2/64 { AdvRouterAddr on; AdvOnLink on; AdvAutonomous on; AdvPreferredLifetime 10000; AdvValidLifetime 12000; }; }; |
Also do a radvdump on HA to check whether radvd messages are beeing sent:
# radvdump Router advertisement from fe80::202:2dff:fe54:d11e (hoplimit 255) Received by interface eth0 # Note: {Min,Max}RtrAdvInterval cannot be obtained with radvdump AdvCurHopLimit: 64 AdvManagedFlag: off AdvOtherConfigFlag: off AdvHomeAgentFlag: on AdvReachableTime: 0 AdvRetransTimer: 0 Prefix fec0:106:2700::2/64 AdvValidLifetime: 12000 AdvPreferredLifetime: 10000 AdvOnLink: on AdvAutonomous: on AdvRouterAddr: on AdvSourceLLAddress: 00 02 2D 54 D1 1E AdvHomeAgentInfo: HomeAgentPreference: 20 HomeAgentLifetime: 1000 |
Do every configuration as shown above; it's especially important to have a different ESSID on the home net and visited network.
When you start mobile-IPv6 on MN, you will see multicasting router solicitations messages:
# tcpdump -i eth0 -vv ip6 or proto ipv6 ... 13:32:54.681763 fe80::202:a5ff:fe6f:a08a > ff02::2: icmp6: router solicitation \ (src lladdr: 0:2:a5:6f:a0:8a) (len 16, hlim 255) 13:32:55.681763 fe80::202:a5ff:fe6f:a08a > ff02::2: icmp6: router solicitation \ (src lladdr: 0:2:a5:6f:a0:8a) (len 16, hlim 255) 13:32:57.681765 fe80::202:a5ff:fe6f:a08a > ff02::2: icmp6: router solicitation \ (src lladdr: 0:2:a5:6f:a0:8a) (len 16, hlim 255) ... |
Generic movement detection uses Neighbor Unreachability Detection to detect when the default router is no longer bi-directionally reachable, in which case the mobile node must discover a new default router (usually on a new link).
To easily see whats going on, you should have one xterm window for each of these commands:
# watch ifconfig eth0 # watch route -A inet6 # tcpdump -i eth0 -vv ip6 or proto ipv6 |
To "travel" to another net, you can issue the command on MN:
# iwconfig eth1 essid visitnet |
The MN is then on the other wireless network, and since it is sending out "router solicitation" (multicast), our AR will respond with it's prefix. MN will then configure itself with at new IPv6 address with the received prefix and it's own MAC address. If you type ifconfig eth0 you will see the new IPv6 address:
Almost at the same time, the MN will perform a binding update to HA. In your tcpdump window, you will see several packets destined to HA. To verify that the binding update has been sent and acknowledged from MN:
You can also verify the binding with the following command (on MN):
# mipdiag -l Mobile IPv6 Binding update list Recipient CN: fec0:106:2700::2 BINDING home address: fec0:106:2700::4 care-of address: fec0:106:1100:0:290:7dff:fef3:31a expires: 936 sequence: 0 state: 1 delay: 3 max delay 32 callback time: 736 |
You can also verify it on HA with the statistics option (-s) and with the "binding cache" (-c) option:
# mipdiag -c Mobile IPv6 Binding cache Home Address Care-of Address Lifetime Type fec0:106:2700::4 fec0:106:1100:0:290:7dff:fef3:31a 971 2 |
From the MN, you can try to ping AR's eth1 (fec0:106:1100::1):
# ping6 fec0:106:1100::1 PING fec0:106:1100::1(fec0:106:1100::1) from fec0:106:2700::4 : 56 data bytes 64 bytes from fec0:106:1100::1: icmp_seq=1 ttl=62 time=8.01 ms 64 bytes from fec0:106:1100::1: icmp_seq=2 ttl=62 time=8.02 ms ... |
By using tcpdump, you can see how the packets travel:
You can now see the statistics have been updated (on MN):
# mipdiag -s Mobile IPv6 Statistics NEncapsulations : 56 NDecapsulations : 25 ... |
One interesting thing MIPv6 does is change the default route to a tunnel. The new default route becomes:
# route -A inet6 Kernel IPv6 routing table Destination Next Hop Flags Metric Ref Use Iface ::/0 :: UD 64 0 0 ip6tnl1 .... |
If it doesn't add a default route, you may add it manually:
# ip route ::/0 via dev ip6tnl |
To travel to several visited networks, is no different than travel to one network. The only thing you must have in mind is that you will generate a new address for each visited network.
MN travelling through severeal different LANs.
MN first visits 'visitnet', as we have been through above.
MN is then travelling from 'visitnet' to 'visitnet2'.
When at 'visitnet2', MN generates a new IPv6 address and do a new binding update to HA.
MN then travels back home. (Se next section.)
The AR at "visitnet2", is configured exactly as the other AR (at "visitnet"), except using address fec0:106:1000::/64 instead of fec0:106:1100::/64.
To make the mobile node travel from 'visitnet' to 'visitnet2', issue the command (on MN):
# iwconfig eth0 essid visitnet2 |
You will then see the MN configures itself to the new network:
Note! You may have to restart mobile-ipv6 on MN when coming to a new network!
# /etc/init.d/mobile-ip6 restart Stopping Mobile IPv6: OK Starting Mobile IPv6: OK |
The MN will then perform a new binding update to HA. Notice the new "care-of address":
# mipdiag -l Mobile IPv6 Binding update list Recipient CN: fec0:106:2700::2 BINDING home address: fec0:106:2700::4 care-of address: fec0:106:1000:0:290:7dff:fef3:31a expires: 973 sequence: 14 state: 1 delay: 3 max delay 32 callback time: 773 |
You can also see the "binding cache" on HA has been updated:
# mipdiag -c Mobile IPv6 Binding cache Home Address Care-of Address Lifetime Type fec0:106:2700::4 fec0:106:1000:0:290:7dff:fef3:31a 943 2 |
To make the MN return home, you can just issue the command:
# iwconfig eth0 essid homenet |
The MN will know it is back home, since HA is sending out radvd messages with the HA-bit set (AdvHomeAgentFlag), see Section 4.2.4
You can see the MN "is back home", since the binding cache information at HA is flushed (empty):
Mobile IPv6 Binding cache Home Address Care-of Address Lifetime Type |
To really get the feel on how mobile IP works, fire up GnomeMeeting (See the figure GnomeMeeting and start a netmeeting. Note! You must use the latest GnomeMeeting to get support for IPv6! Then do a "travel" and you can see an almost smooth handover.
Using GnomeMeeting with IPv6 to test roaming between two wireless networks
Q: Why do we have to create the /dev/mipv6_dev entry?
A: The dev file is mainly so that the userspace tool, mipdiag, can make modifications to the kernel parameters using ioctl calls through the device file. mknod creates the special device file with paramters recognizable by the mobile-ip6 module.
Q: Is there any support for kernel 2.6.x?
A: Here is the answer from Henrik Petander on the MIPL mailinglist:
"Here is a short overview of the status of MIPL for 2.6 kernel series:"
"We have finished the kernel infrastructure for Mobile IPv6 in cooperation with the USAGI project. The infrastructure does route optimization, tunneling and policy routing."
"We are now working on the userspace daemon which handles the MIPv6 signaling and controls the operation of the kernel part. The userspace part is also progressing nicely. However, the protocol logic is still missing, so there isn't really anything for users to test yet. We should have a well working and tested prototype ready and by the end of March."
Q: Does MIPL support IPSec?
A: There is no support IPSec on 2.4.x. MIPL for 2.6 series will have IPSec support from the start. You may use a third-party IPSec implementation.
Q: How can I control the type of routing used for communication between the MN and a CN (through HA tunnel or by direct communication using binding update/acks)?
A: You can control this through:
/proc/sys/conf/net/ipv6/mobility/accept_return_routability
If you do not want to use return routability and route optimization, set it to 0 with:
# echo 0 > /proc/sys/..../accept_return_routability
Then MN will communicate with CNs only through the home tunnel.
Q: Can different wireless networks have different ESSIDs/WEP keys?
A: Yes, but you must change this upon arrival to the new network. MIPv6 from MIPL can't do this automatically.
Q: If MN has travelled through several visited LAN, and then returning home; the interface still has all the autogenerated IPv6 addresses from all the visited networks! Is there any way to "flush/delete" these addresses?
A: No, I do not know of any automatic way these adresses can be removed, but you can delete them manually:
# ifconfig eth0 inet6 del <ipv6-address>
Q: Host B has two interfaces with two different subnets assigned. When I ping B from host A, it does not answer! Why not? Host A knows where host B (subnets) are!
A: The host B doesn't know where host A is (B doesn't know where A's net is), so you must add a route entry:
# ip route add fec0:106:2700::/64 via fec0:106:2300::1
or
# route -A inet6 add fec0:106:2700::/64 gw fec0:106:2300::1 dev eth0
Q: How do I set a default gateway in IPv6?
A: You do that using the traditional "route":
# route -A inet6 add default gw <ipv6-host>
or the newer "ip" command:
# ip route ::/0 via <ipv6-host>
Q: Why does the host send a multicast address rather than an anycast address, requesting for router solicitation?
A: Because the host wants an answer from every router, not from just any router. The idea is to be able to get all parameters and to choose the "best" default router.
Q: Why doesn't MN notice that it has moved?
A: It thinks that it's previous router is still reachable. This may result from very large lifetimes in router advertisements. Check the configuration of the program sending router advertisements in the router. If the program supports router advertisement intervals, you can use this to help MN in movement detection by setting the use of interval to on. See man radvd.conf for details.
Mobile IPv6 for Linux http://www.mipl.mediapoli.com/
Mobile IP Working Group (IETF) http://www.ietf.org/html.charters/mobileip-charter.html
Mobile IPv6 draft http://www.ietf.org/internet-drafts/draft-ietf-mobileip-ipv6-24.txt
IPv6 Working Group (IETF) http://www.ietf.org/html.charters/ipv6-charter.html
RFC2460 Internet Protocol, Version 6 (IPv6) Specification http://www.ietf.org/rfc/rfc2460.txt
RFC2461 Neighbor Discovery for IP Version 6 (IPv6) http://www.ietf.org/rfc/rfc2461.txt
RFC2462 IPv6 Stateless Address Autoconfiguration http://www.ietf.org/rfc/rfc2462.txt
Peter Bieringer's Linux IPv6 HOWTO (en) http://ldp.linux.no/HOWTO/Linux+IPv6-HOWTO/
Current Status of IPv6 Support for Networking Applications http://www.deepspace6.net/docs/ipv6_status_page_apps.html
Linux Kernel HOWTO http://www.ibiblio.org/pub/Linux/docs/HOWTO/other-formats/html_single/Kernel-HOWTO.html
Copyright © 2003,2004 by Lars Strand. This material may be distributed only subject to the terms and conditions set forth in the Open Publication License, v1.0 or later (the latest version is presently available at http://www.opencontent.org/openpub/ ).
Linux is a registered trademark of Linus Torvalds.
No liability for the contents of this document can be accepted. Use the concepts, examples and information at your own risk. There may be errors and inaccuracies, that could be damaging to your system. Proceed with caution, and although this is highly unlikely, the author(s) do not take any responsibility.
All copyrights are held by their by their respective owners, unless specifically noted otherwise. Use of a term in this document should not be regarded as affecting the validity of any trademark or service mark. Naming of particular products or brands should not be seen as endorsements.
This document was originally written in LaTeX using Emacs. HTML version created with latex2html. Later it was converted to DocBook XML.
An up-to-date version of this document can be found at:
Suggestions, corrections, additions wanted. Contributors wanted and acknowledged. Flames not wanted.
I can always be reached at <lars at unik no>
Homepage: http://www.gnist.org/~lars
This document was produced as a part of Interoperable Networks for Secure Communications (INSC task 6)
Thanks to Andreas Hafslund (andreha [at] unik.no) for initial support. Also thanks to UniK (University Graduate Center) http://www.unik.no and FFI (Norwegian Defence Research Establishment) http://www.ffi.mil.no for hardware support.
Thanks also to the other HOWTO authors whose works I have referenced:
Linux IPv6 HOWTO (en) by Peter Bieringer