08-19-2011 04:01 AM - edited 03-04-2019 01:20 PM
Guys can someone plz tell me abt the conceot of route summarization.....RIPv1 and IGRP does auto summarization......and cant be stopped is it right....secondly EIGRP, BGP, OSPF can do but we have do no auto summary command.....i am really confused abt this concept ....i was reading the summarization has many beneifits but again i have never seen a network (BGP specially) or eigrp where we do auto summarization why is that ......can someone plz shed a light on this topic
really need help understanding it
Solved! Go to Solution.
08-19-2011 05:22 AM
I think you are confusing the terms of auto-summarisation used by routing protocols and route summarisation which is a way of summarising a number of subnets into a larger supernet.
Auto-summarisation is the ability of a router to summarise the classful network across boundaries eg.
172.16.5.0 R1 -> 192.168.5.0 <- R2 -> 192.168.6.0 <- R3 -> 172.16.6.0
in the above all routers are using EIGRP and have auto-summarisation turned on. R3 when it advertises the 172.16.6.0/24 network to R2 will advertise it as 172.16.0.0/16 because that is the classful network. It does this because the interface connected to R2 is in the 192.168.6.x newtwork and this is not part of the 172.16.0.0 classful network.
But R1 will also advertise 172.16.0.0 to R2 as well for the same reasons. So R2 has 2 routes to 172.16.0.0/16 pointing 2 different ways. It has no idea that the actual networks are 172.16.5.x via R1 and 172.16.6.x via R3. If you disable auto-summarisation under EIGRP ie.
router eigrp 10
no auto-summary
then the actual subnets are advertised to R2 ie. 172.16.5.0/24 from R1 and 172.16.6.0/24 from R3. To all intents and purpose auto-summarisation is of little use in your network and you generally turn it off precisely because of the issues outlined above.
Route summarisation however is very useful ie. you have an HQ office with 3 remote sites. At each remote site you have 4 class subnets eg. we will use one site as an example -
site 1
192.168.0.0/24
192.168.1.0/24
192.168.2.0/24
192.168.3.0/24
now you could advertise these individual subnets to the HQ site but you could also advertise one single summarised route ie.
192.168.0.0 255.255.252.0
the above summarised route covers 192.168.0.0 -> 192.168.3.255
the 2 main advantages of this is -
1) the route table on the HQ router is smaller. In this example it doesn't make much difference but when you apply it to large enterprise networks or the internet for example it can make huge difference in terms of routing table size.
2) routing protocol stability. If you advertise the 4 individual subnets and one of the routers interfaces for one of the subnets fails then the route is removed and this then gets communicated to the HQ site. The HQ site and indeed any other sites then need to find another route. In the case of EIGRP it will start sending out queries from HQ to all the other sites.
If you use the summary route and one of the individual routes within that summary is removed from the routing table this is contained within site 1 and not transmitted to HQ. This is fine because the only way to get to the failed network is via site 1 so there is no point in HQ and any other sites sending out queries.
Of the big 2 IGPs in use, EIGRP and OSPF, EIGRP is particularly vulnerable to this when sending out queries can lead to SIAs (Stuck in active) and instability in the routing tables. Note for completeness EIGRP also has stub functionality which helps alleviate these problems but summarisation is still an important tool.
Jon
08-19-2011 05:22 AM
I think you are confusing the terms of auto-summarisation used by routing protocols and route summarisation which is a way of summarising a number of subnets into a larger supernet.
Auto-summarisation is the ability of a router to summarise the classful network across boundaries eg.
172.16.5.0 R1 -> 192.168.5.0 <- R2 -> 192.168.6.0 <- R3 -> 172.16.6.0
in the above all routers are using EIGRP and have auto-summarisation turned on. R3 when it advertises the 172.16.6.0/24 network to R2 will advertise it as 172.16.0.0/16 because that is the classful network. It does this because the interface connected to R2 is in the 192.168.6.x newtwork and this is not part of the 172.16.0.0 classful network.
But R1 will also advertise 172.16.0.0 to R2 as well for the same reasons. So R2 has 2 routes to 172.16.0.0/16 pointing 2 different ways. It has no idea that the actual networks are 172.16.5.x via R1 and 172.16.6.x via R3. If you disable auto-summarisation under EIGRP ie.
router eigrp 10
no auto-summary
then the actual subnets are advertised to R2 ie. 172.16.5.0/24 from R1 and 172.16.6.0/24 from R3. To all intents and purpose auto-summarisation is of little use in your network and you generally turn it off precisely because of the issues outlined above.
Route summarisation however is very useful ie. you have an HQ office with 3 remote sites. At each remote site you have 4 class subnets eg. we will use one site as an example -
site 1
192.168.0.0/24
192.168.1.0/24
192.168.2.0/24
192.168.3.0/24
now you could advertise these individual subnets to the HQ site but you could also advertise one single summarised route ie.
192.168.0.0 255.255.252.0
the above summarised route covers 192.168.0.0 -> 192.168.3.255
the 2 main advantages of this is -
1) the route table on the HQ router is smaller. In this example it doesn't make much difference but when you apply it to large enterprise networks or the internet for example it can make huge difference in terms of routing table size.
2) routing protocol stability. If you advertise the 4 individual subnets and one of the routers interfaces for one of the subnets fails then the route is removed and this then gets communicated to the HQ site. The HQ site and indeed any other sites then need to find another route. In the case of EIGRP it will start sending out queries from HQ to all the other sites.
If you use the summary route and one of the individual routes within that summary is removed from the routing table this is contained within site 1 and not transmitted to HQ. This is fine because the only way to get to the failed network is via site 1 so there is no point in HQ and any other sites sending out queries.
Of the big 2 IGPs in use, EIGRP and OSPF, EIGRP is particularly vulnerable to this when sending out queries can lead to SIAs (Stuck in active) and instability in the routing tables. Note for completeness EIGRP also has stub functionality which helps alleviate these problems but summarisation is still an important tool.
Jon
08-19-2011 06:07 AM
Thanks Jon you are the main.....very nicely explained.....Jon my second question is abt OSPF......in OSPF you can run the OSPF on interface with 0.0.0.0 wild card and you can run ospf with network command for sake of example if i have arouter and its interfac eis 192.168.1.1/24 i can do
router ospf 1
network 192.168.1.0 0.0.0.0.255
and i can also do
router ospf 1
network 192.168.1.1 0.0.0.0
whats the diffrence
Thanks again
08-19-2011 06:44 AM
Hi,
in first config all interfaces which ip address begins with 192.168.1 will be part of the area (which you forgot to mention at the end of the command).
In second config only the interface with ip address 192.168.1.1 will be participating in the area.You could also do the same thing with the interface command ip ospf 1 area 0 under the interface you want to participate in OSPF and be part of area mentioned.
Regards.
Alain.
08-19-2011 06:48 AM
the network 192.168.1.0 0.0.0.255 will include any interface in the IP range 192.168.1.0/24
for example 192.168.1.1 or 192.168.1.200
while 192.168.1.1 0.0.0.0 will include this IP/interface of 192.168.1.1 only if exist
HTH
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide