- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-09-2009 11:12 AM - edited 03-06-2019 05:06 AM
Hi every body!
Let say we have ipv6 enabled router r1, r1 has two interface so and f 0.
so has ip address 2340:1111:AAAA :2::1/64
fo has ip address 2340: 1111:AAAA:1::/64 eui-64
S0R1 f 0------------------h2
H2 has ip address 2340:1111:AAA:1: 0234: 56FF:FE78:9ABC
r1 receives a ip packet with destination ip address" 2340:1111:AAA:1: 0234: 56FF:FE78:9ABC" . r1 has no entry in arp cache for h2 's mac address. Will r1 note the " 56FF: FE78" in interface field and conclude this is the mac address for h2 and therefore does not need to send arp request to h2?
2) If i have to use the option of eui-64 to configure ipv6 address on interface, ipv6 sets the 7 bit to 1 to get the modified eui-64 format. My question is why? why can we not use the mac address whose 7 bit is 0?
Thanks a lot!
Solved! Go to Solution.
- Labels:
-
Other Switching
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-09-2009 12:03 PM
Hello Sarah,
in ipv6 ICMPv6 includes and extends functions of ICMPv4, ARP, and IGMP.
There are also duplicate address detection and neighbor discovery that are part of ND.
Actually an IPv6 host mantains a table of known gateway, a table of known hosts.
For each host the IPv6 host tracks the reachability: indirectly if a TCP session is active between the two or sending ICMP probes from time to time.
to answer your first question:
an ICMPv6 packet of type NS (type 135) is sent out using a solicited multicast address destination that is a special link local multicast address that includes the less significant 24 bits.
This packet is used to discover the L2 MAC address or R2.
R2 answers provides the necessary information in an ICMPv6 message of type NA neighbor advertisement (type 136).
This reply is sent with a unicast source MAC address and a unicast destination that is that of R1 lan interface.
IPV6 uses no broadcast at all.
At first is difficult to believe that this is possible
see for an explanation:
http://www.cisco.com/en/US/docs/ios/ipv6/configuration/guide/ip6-addrg_bsc_con.html#wp1282543
Once the entry is created the link address is associated to R2:lan interface ipv6 address.
From time to time if not refreshed by routing protocols or TCP sessions R1 can send again messages to confirm that R1 is alive and information is correct
2)
this is somewhat a strange point
see
http://www.ietf.org/rfc/rfc2464.txt
in IEEE 48 bits MAC addresses the second bit of the most significant octet is the U/L bit : 0 means universal, 1 means locally administered.
In creating the eui-64 interface identifier they decided in someway to revert this bit: now if the bit is set means a universal identifier.
see also appendix A
http://www.faqs.org/rfcs/rfc2373.html
Hope to help
Giuseppe
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-09-2009 11:29 AM
I'm no IPv6 expert but I was thinking there is no arp. My understanding is IPv6 uses Neighbor Discovery and Multicast to get L2 info. The L2 info is part of the IPv6 address / Well the last 64 bits of the address are related to the 48 bit mac address right?
Wasn't one of the big goals of IPv6 to not have so much broadcast traffic?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-09-2009 12:05 PM
Thanks Greg!
The last 64 bits may be or may not be the part of mac address. If last 64 bits has XXFF.EFXX in the middle, then yes if not then no.
yes you are right, ipv6 use multicast address .
thanks a lot!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-09-2009 12:03 PM
Hello Sarah,
in ipv6 ICMPv6 includes and extends functions of ICMPv4, ARP, and IGMP.
There are also duplicate address detection and neighbor discovery that are part of ND.
Actually an IPv6 host mantains a table of known gateway, a table of known hosts.
For each host the IPv6 host tracks the reachability: indirectly if a TCP session is active between the two or sending ICMP probes from time to time.
to answer your first question:
an ICMPv6 packet of type NS (type 135) is sent out using a solicited multicast address destination that is a special link local multicast address that includes the less significant 24 bits.
This packet is used to discover the L2 MAC address or R2.
R2 answers provides the necessary information in an ICMPv6 message of type NA neighbor advertisement (type 136).
This reply is sent with a unicast source MAC address and a unicast destination that is that of R1 lan interface.
IPV6 uses no broadcast at all.
At first is difficult to believe that this is possible
see for an explanation:
http://www.cisco.com/en/US/docs/ios/ipv6/configuration/guide/ip6-addrg_bsc_con.html#wp1282543
Once the entry is created the link address is associated to R2:lan interface ipv6 address.
From time to time if not refreshed by routing protocols or TCP sessions R1 can send again messages to confirm that R1 is alive and information is correct
2)
this is somewhat a strange point
see
http://www.ietf.org/rfc/rfc2464.txt
in IEEE 48 bits MAC addresses the second bit of the most significant octet is the U/L bit : 0 means universal, 1 means locally administered.
In creating the eui-64 interface identifier they decided in someway to revert this bit: now if the bit is set means a universal identifier.
see also appendix A
http://www.faqs.org/rfcs/rfc2373.html
Hope to help
Giuseppe
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-08-2010 07:16 AM
Hi
So what is the command that would replace the "show arp" command?
- Andre
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-08-2010 07:57 AM
Hello,
So what is the command that would replace the "show arp" command?
show ipv6 neighbors
Best regards,
Peter
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-08-2010 08:47 AM
What am I doing wrong:
CORE1#ping ipv6 2801:0:40:a::1
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 2801:0:40:A::1, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 0/0/0 ms
CORE1#show ipv6 neighbors
CORE1#
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-08-2010 12:33 PM
Andre,
Is it possible that you are pinging your own IPv6 address?
For me, the command works fine:
Switch#show ipv6 neighbors
IPv6 Address Age Link-layer Addr State Interface
FE80::8965:A9E6:1759:BB4F 23 36b2.ef59.5a6b STALE Vl539
FE80::20C:29FF:FE78:D92A 23 000c.2978.d92a STALE Vl638
FE80::214:5EFF:FEE0:59A2 23 0014.5ee0.59a2 STALE Vl338
3001:4118:300:122:CC7:20FF:FE5D:5F74 23 0ec7.205d.5f74 STALE Vl239
3001:4118:300:100::32 2 001a.6de9.3c9a STALE Vl238
FE80::216:76FF:FE94:A9CE 23 0016.7694.a9ce STALE Vl156
Best regards,
Peter
