Jump to content


EIGRP


  • Please log in to reply
71 replies to this topic

#1 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 13 April 2009 - 10:51 PM

EIGRP


EIGRP Summary


The characteristics of EIGRP follow:

 Hybrid routing protocol (distance vector that has link-state protocol characteristics).

 Use DUAL, first proposed by E. W. Dijkstra and C. S. Scholten, to perform distributed shortest-path routing while maintaining freedom from loops at every instant. Although many researchers have contributed to the development of DUAL, the most prominent work is that of J. J. Garcia-Luna-Aceves.

 Cisco Proprietary created in 1994.

 First released in IOS 9.21

 Uses IP protocol 88.

 Makes Automatic summarization on network Class boundary.

 Classless protocol (supports VLSM).

 Have the power to shut the Auto-summarization And make a configured manual Summarization.

 Default composite metric of bandwidth and delay.

 You can factor load, MTU and reliability into the metric.

 Eigrp metric is the same as IGRP*256, It uses the smallest B.W,Reliablity,Load & MTU with the Comulative delay upon the path…..The MTU doesn’t actually used in the Metric calculations,
But is included in the EIGRP Routing updates.

 Sends route updates to multicast address 224.0.0.10, and nei. Reply’s back with Unicast Address.

 Sends non-periodic, partial, and bounded updates.

 Send Hello packets every 5 sec. and Hold down timer is 15 sec.

 For Low speed Hello is every 60 sec. with hold down time 180 sec.

 By default, EIGRP uses no more than 50 percent of the bandwidth of a link.

 Support for authentication via MD5 Only.

 Uses DUAL for loop prevention, and generating Succ./Fesible Succ.

 Maximum paths for Load-balancing are 6 & default is 4 , maximum are 16 in IOS 12.3(2)T and later IOS releases

 By default, Equal-Metric load balancing. If Unequal-Metric load sharing is used the router will load share inversely proportional to the metrics of the paths.

 Administrative distance is 90 for EIGRP internal routes, 170 for EIGRP external routes, and 5 for EIGRP summary routes.

 Potential routing protocol for the core of a network; used in large networks.

 For neighbor relation to be established, both routers must send and receive Hello or Ack packets from each other, they must have the same AS #, and the same Metric K values.

 Eigrp doesn’t restrict that neighbors must have the same Hello & dead interval timers, Unlike OSPF.

 Has a Maximum hop count of 255, the default is 100 in the last IOS releases.



The composite metric for each EIGRP route is calculated as

EIGRP metric = IGRP metric * 256
IGRP metric = [k1*BWIGRP(min) + (k2* BWIGRP(min))/(256-LOAD) + k3*DLYIGRP(sum)]
x [k5/(RELIABILITY + k4)]

If k5 is set to zero, the [k5/(RELIABILITY+k4)] term is not used.

Given the default values for k1 through k5, the composite metric calculation used by EIGRP reduces to the default metric:

IGRP metric = BWIGRP(min) + DLYIGRP(sum)
BWIGRP(min) = 107/BW(min)
DLYIGRP(sum) = DLY(sum) /10




Check the attached document for more information & configurations.
Hope it might help :)

Attached Files


  • 1

#2 raaki_88

raaki_88

    BOW DOWN TO THE KING

  • Members
  • PipPipPipPipPip
  • 1442 posts
  • Gender:Male

Posted 13 April 2009 - 11:49 PM

thank you ... good for revision ...
  • 0

#3 talent pk

talent pk

    Cisco Newbie

  • Technical Experts
  • PipPipPipPipPip
  • 4028 posts
  • Gender:Male
  • Location:Pakistan
  • Interests:Oracle

Posted 15 April 2009 - 05:27 PM

well thats a good stuff for quick reference.

thanks once again
  • 0

#4 tigerplug

tigerplug

    Newbie

  • Members
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Wherever I may Roam!

Posted 15 April 2009 - 10:54 PM

Excellent!
Thanks for this, it'll come in really useful over the next week or so :)
  • 0

#5 Mo'men

Mo'men

    Advanced Member

  • Members
  • PipPipPip
  • 245 posts
  • Gender:Male
  • Location:Egypt
  • Interests:Reading, Computer games (Multiplayer Medal of Honor)
    Jogging, Matlab Programming!

Posted 16 April 2009 - 02:23 AM

Great stuff, thanks for sharing knowledge...

Proud to see an Egyptian CCIE around! :)

Regards

Mo'men
  • 0

#6 martinlo

martinlo

    V.I.P. Member

  • Global Moderators
  • PipPipPipPipPip
  • 2697 posts
  • Gender:Male
  • Location:Land Of 10,000 Lakes

Posted 16 April 2009 - 02:49 AM

Another major benefit is Feasible Successor route or Backup Route. EIGRP calculates Backup route if a condition is met. Backup route is stored in EIGRP topology (router knows about it) and is ready "to go" in case of main route failure. Backup route is not displayed in routing table (sh ip route). that is why EIGRP convergence is considered to be faster then OSPF.

EIGRP could route (support) AppleTalk and Novell IXP protocols. in the past when those protocols were used in networks, that was another benefit for EIGRP. Today, there is no need for those "faded out" protocols.
  • 0

#7 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 16 April 2009 - 07:30 AM

Another major benefit is Feasible Successor route or Backup Route. EIGRP calculates Backup route if a condition is met. Backup route is stored in EIGRP topology (router knows about it) and is ready "to go" in case of main route failure. Backup route is not displayed in routing table (sh ip route). that is why EIGRP convergence is considered to be faster then OSPF.

EIGRP could route (support) AppleTalk and Novell IXP protocols. in the past when those protocols were used in networks, that was another benefit for EIGRP. Today, there is no need for those "faded out" protocols.


Yes that's what cisco say, but in OSPF the Alternate routes will be Also in the topology Table (LSA Data-base) !
  • 0

#8 martinlo

martinlo

    V.I.P. Member

  • Global Moderators
  • PipPipPipPipPip
  • 2697 posts
  • Gender:Male
  • Location:Land Of 10,000 Lakes

Posted 16 April 2009 - 01:32 PM

Yes that's what cisco say, but in OSPF the Alternate routes will be Also in the topology Table (LSA Data-base) !


EIGRP calculates both routes, Feasible and Feasible Successor at the very beginning, at the same time. So, the backup route is ready when needed. My understanding is OSPF calculates "the best" route first, and then, when backup route is needed, OSPF will have to recalculate next best route from what it stores in database.

But, I could be wrong about OSPF backup route. some more digging maybe needed ?
  • 0

#9 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 16 April 2009 - 07:09 PM

EIGRP calculates both routes, Feasible and Feasible Successor at the very beginning, at the same time. So, the backup route is ready when needed. My understanding is OSPF calculates "the best" route first, and then, when backup route is needed, OSPF will have to recalculate next best route from what it stores in database.

But, I could be wrong about OSPF backup route. some more digging maybe needed ?


You may check the LSA database for OSPF via the cmd sh ip ospf databse, check the output if there is any field for the cost/metric

Edited by DarkFiber, 16 April 2009 - 07:11 PM.

  • 0

#10 martinlo

martinlo

    V.I.P. Member

  • Global Moderators
  • PipPipPipPipPip
  • 2697 posts
  • Gender:Male
  • Location:Land Of 10,000 Lakes

Posted 18 April 2009 - 12:19 PM

You may check the LSA database for OSPF via the cmd sh ip ospf databse, check the output if there is any field for the cost/metric


Do you have an example? snapshot of database with alternative route.

I checked Sybex CCIE, CCNP, CCNA, and CiscoPress ICDN2. none of those books have mentioned Alternative route!
  • 0

#11 vijayjuan

vijayjuan

    Member

  • Members
  • PipPip
  • 50 posts
  • Gender:Male
  • Location:Pudukkottai

Posted 18 April 2009 - 03:22 PM

Dear Darkfibre,
SPF has to recalcuate to find out the next best path..
  • 0

#12 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 18 April 2009 - 06:59 PM

Do you have an example? snapshot of database with alternative route.

I checked Sybex CCIE, CCNP, CCNA, and CiscoPress ICDN2. none of those books have mentioned Alternative route!


Dear Martinlo, just becoz a fact is not listed in those books doesnt mean that it doesnt exist. And a friendly advice, i have gone throught a lot of books except one which i have just heard of, never seen anywhere(it explained how to manage 100 areas in ospf, my instructor told me in my ccna class 3 years ago). If you want to understand ospf or any open protocol, shut these books down, get your routers up, open the RFC, open cisco docs and start experimenting. Trust me 1 hour spent in this will be a lot better then reading those books for years ;-).

These books are just for those who are just starting to play with these protocols, if you are not a newbie and experienced professional then you can take my advice :-)

But you were right ospf does have a alternate route in its database but it doesnt calculates it !!!, the moment there is a change in topology the SPF initializes and run agains. Atleast thats how interpreted the debug, if i am wrong then kindly correct me anyone ;-)
  • 0

#13 martinlo

martinlo

    V.I.P. Member

  • Global Moderators
  • PipPipPipPipPip
  • 2697 posts
  • Gender:Male
  • Location:Land Of 10,000 Lakes

Posted 18 April 2009 - 07:34 PM

But you were right ospf does have a alternate route in its database but it doesnt calculates it !!!, the moment there is a change in topology the SPF initializes and run agains. At least thats how interpreted the debug, if i am wrong then kindly correct me anyone ;-)

that is my point. there is no show ospf topology and show ospf database is not the same. so, there is NO Alternative route "ready to go" in OSPF like in EIGRP.
  • 0

#14 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 18 April 2009 - 10:10 PM

that is my point. there is no show ospf topology and show ospf database is not the same. so, there is NO Alternative route "ready to go" in OSPF like in EIGRP.


hmmm.........so how in the 1st place OSPF router picked the best LSA in Topology and inserted it in routing table. !

ofcorse i know what exactly OSPF is doing , im just brain storming with you :)
  • 0

#15 moekad

moekad

    Advanced Member

  • Members
  • PipPipPip
  • 270 posts

Posted 18 April 2009 - 10:42 PM

Greeting all
while i was studying i didn't got this:
P 192.168.1.0/24, 1 successors, FD is 3014400
via 192.168.10.10 (3014400/28160), Serial0/0/1
via 172.16.3.1 (41026560/2172416), Serial0/0/1

ok 3014400 i know this is the Feasible Distance to reach network 192.168.1.0
and i know 28160 is Report distance which it report the routers about the FD
and 192.168.10.10 is the next hop
and i know in order EIGRP use DUAL loop-free backup the RD(28160) must be less than FD (3014400)
so to reach network 192.168.1.0/24 using this next hope 192.168.10.10 ok
then now we will go to next hope 172.16.3.1 (2172416 which is RD) is less than FD (3014400)
why if next hop(192.168.10.10) got down the Passive 192.168.1.0/24 removed from the topology table
and the network will be timed out? altough Feasible condition say RD must be less than FD
and 2172416 is less than FD = 3014400
and here another example:
P 172.16.2.0/24, 1 successors, FD is 2172416
via 172.16.3.2 (2172416/28160), Serial0/0/0
via 192.168.10.6 (4294967295/2172416), Serial0/0/1
here if 172.16.3.2 is go down the 192.168.10.6 will reserve it and the ping to network 172.16.2.0/24 will still work why! what difference between the both!
+ what the difference between Passive and Active!!
can anyone explain for me this:
i'll be thankfull and thanks for your help.
  • 0

#16 tunerX

tunerX

    Advanced Member

  • Members
  • PipPipPip
  • 478 posts
  • Gender:Male
  • Location:Back in the USA...

Posted 18 April 2009 - 10:52 PM

Moekad,

The FD is always the lowest cost... FC is used when choosing a feasible successor. The FD is compared to the RD off all other routes in the topology table. The FD is 3014400, that value is compared to all other RDs in the topology table to meet the FC requirement. In this case the neighbor 172.16.3.1 has an RD of 2172416. Since FD is greater than the 172.16.3.1, it is not considered a feasible successor.

via 192.168.10.10 (3014400/28160), Serial0/0/1
via 172.16.3.1 (41026560/2172416), Serial0/0/1
  • 0

#17 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 19 April 2009 - 08:12 AM

OK . a Quick Summary

STEPs:

1- Selecting the Successor , which will be having the smallest FD (Metric)

2- Selecting all eligible Feasible Successors, where each Feasible Successor must pass the feasible condition/Rule :

FD ( successor of step 1 ) > RD ( Feasible Successor Candidate)

3- After having a group of eligible Feasible Successors, Choose the Lowest FD/Metric in this group to be the FS1 , then FS2......ect


Hope this might help
  • 0

#18 moekad

moekad

    Advanced Member

  • Members
  • PipPipPip
  • 270 posts

Posted 19 April 2009 - 08:38 AM

Moekad,

The FD is always the lowest cost... FC is used when choosing a feasible successor. The FD is compared to the RD off all other routes in the topology table. The FD is 3014400, that value is compared to all other RDs in the topology table to meet the FC requirement. In this case the neighbor 172.16.3.1 has an RD of 2172416. Since FD is greater than the 172.16.3.1, it is not considered a feasible successor.

via 192.168.10.10 (3014400/28160), Serial0/0/1
via 172.16.3.1 (41026560/2172416), Serial0/0/1


hey thanks for replying
i understood that in this first next hop
via 192.168.10.10 (3014400/28160), Serial0/0/1 RD(28160 is less than FD 3014400) ok it work
next hop via 172.16.3.1 (41026560/2172416), Serial0/0/1 also ( RD 2172416 is less than FD 41026560)
this is the FC so it must work because i know if RD is less than FD then it work the backup path...
thanks
  • 0

#19 martinlo

martinlo

    V.I.P. Member

  • Global Moderators
  • PipPipPipPipPip
  • 2697 posts
  • Gender:Male
  • Location:Land Of 10,000 Lakes

Posted 19 April 2009 - 03:33 PM

hey thanks for replying
i understood that in this first next hop
via 192.168.10.10 (3014400/28160), Serial0/0/1 RD(28160 is less than FD 3014400) ok it work
next hop via 172.16.3.1 (41026560/2172416), Serial0/0/1 also ( RD 2172416 is less than FD 41026560)
this is the FC so it must work because i know if RD is less than FD then it work the backup path...
thanks


NO, I think, it should be FD 3014400 is the lowest metric from all routes = the best route to router X
RD 2172416 is less than FD 3014400

Another Example:

P 10.1.3.0/24, 1 successors, FD is 2172416
via 10.1.6.3 (2172416/28160), Serial0/1
via 10.1.4.2 (2684416/1794560), Serial0/0

The route to 10.1.3.0 through 10.1.6.3 (Seville) is the successor route, because the
calculated metric (2,172,416), shown as the first of the two numbers in parentheses, is
the best calculated metric.

The route to 10.1.3.0 through 10.1.4.2 (Yosemite) is a feasible successor route, because
the neighbor’s Reported Distance (1,794,560, shown as the second number in
parentheses) is lower than Albuquerque’s FD.

Although both the successor and feasible successor routes are in the EIGRP topology
table, only the successor route is added to the IP routing table.
  • 0

#20 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 19 April 2009 - 06:11 PM

hmmm.........so how in the 1st place OSPF router picked the best LSA in Topology and inserted it in routing table. !

ofcorse i know what exactly OSPF is doing , im just brain storming with you :)


Dear Darkfiber, a very interesting statement you have given here. BEST LSA. Even if we take a very general view of this, it will be correct for type 3 lsa, where every inter area Route corresponds to ONE type3 lsa. But in case of type 1 and type 2, all LSAs are used to form a SPF tree and after that through algorithm best path is selected. So i think this statement is not correct. What do you think ?
  • 0

#21 raaki_88

raaki_88

    BOW DOWN TO THE KING

  • Members
  • PipPipPipPipPip
  • 1442 posts
  • Gender:Male

Posted 19 April 2009 - 08:33 PM

i have forumlated a two direct connected router topology and few loopbacks to test whats going on .. interesting enough .. i have just started ospf process on one router and advertised it networks and i have not even touched the second router here is the output

here are the interfaces

rb(config-router)#do show ip int brief
Interface				  IP-Address	  OK? Method Status				Protocol
FastEthernet0/0			10.0.0.2		YES manual up					up	  
FastEthernet0/1			unassigned	  YES unset  administratively down down	
Loopback1				  4.4.4.4		 YES manual up					up	  
Loopback6				  6.6.6.6		 YES manual up					up

rb(config-router)#
*Mar  1 00:14:35.447: OSPF: Schedule SPF in area 0
	  Change in LS ID 6.6.6.6, LSA type R, , spf-type Full
rb(config-router)#
*Mar  1 00:14:40.447: OSPF: running SPF for area 0, SPF-type Full
*Mar  1 00:14:40.447: OSPF: Initializing to run spf
*Mar  1 00:14:40.447:  OSPF - spf_intra() - rebuilding the tree
*Mar  1 00:14:40.447:  It is a router LSA 6.6.6.6. Link Count 3
*Mar  1 00:14:40.447:   Processing link 0, id 6.6.6.6, link data 255.255.255.255, type 3
*Mar  1 00:14:40.451:	Add better path to LSA ID 6.6.6.6, gateway 6.6.6.6, dist 1
*Mar  1 00:14:40.451:	Add path: next-hop 6.6.6.6, interface Loopback6
*Mar  1 00:14:40.451:   Processing link 1, id 4.4.4.4, link data 255.255.255.255, type 3
*Mar  1 00:14:40.455:	Add better path to LSA ID 4.4.4.4, gateway 4.4.4.4, dist 1
*Mar  1 00:14:40.455:	Add path: next-hop 4.4.4.4, interface Loopback1
*Mar  1 00:14:40.455:   Processing link 2, id 10.0.0.0, link data 255.255.255.0, type 3
*Mar  1 00:14:40.455:	Add better path to LSA ID 10.0.0.255, gateway 10.0.0.0, dist 10
*Mar  1 00:14:40.459:	Add path: next-hop 10.0.0.2, interface FastEt
rb(config-router)#hernet0/0
*Mar  1 00:14:40.459: OSPF: Adding Stub nets
*Mar  1 00:14:40.459: OSPF: Added Network Route to 4.4.4.4 Mask /32. Metric: 1, Next Hop: 4.4.4.4
*Mar  1 00:14:40.459: OSPF: Added Network Route to 6.6.6.6 Mask /32. Metric: 1, Next Hop: 6.6.6.6
*Mar  1 00:14:40.463: OSPF: Added Network Route to 10.0.0.0 Mask /24. Metric: 10, Next Hop: 10.0.0.2
*Mar  1 00:14:40.463: OSPF: Entered intra-area route sync - area 0
*Mar  1 00:14:40.463: OSPF: ospf_gen_asbr_sum_all_areas
*Mar  1 00:14:40.467: OSPF: running spf for summaries area 0
*Mar  1 00:14:40.467: OSPF: Entered inter-area route sync - area 0
*Mar  1 00:14:40.467: OSPF: Started Building Type 5 External Routes
*Mar  1 00:14:40.467: OSPF: Started Building Type 7 External Routes
*Mar  1 00:14:40.467: OSPF: Entered External route sync - area dummy area
*Mar  1 00:14:40.471: OSPF: Entered NSSA route sync - area 0
rb(config-router)#

my question is without any other router in area 0 why router is selecting best path to 6.6.6.6 ?

"Add better path to LSA ID 6.6.6.6, gateway 6.6.6.6, dist 1"

next this is the rouput after intiating ospf in other router

*Mar  1 00:20:10.303: OSPF: Adding Stub nets
*Mar  1 00:20:10.303: OSPF: Added Network Route to 1.1.1.1 Mask /32. Metric: 11, Next Hop: 10.0.0.1
*Mar  1 00:20:10.307: OSPF: Added Network Route to 2.2.2.2 Mask /32. Metric: 11, Next Hop: 10.0.0.1
*Mar  1 00:20:10.307: OSPF: Added Network Route to 4.4.4.4 Mask /32. Metric: 1, Next H
rb(config-router)#op: 4.4.4.4
*Mar  1 00:20:10.307: OSPF: Added Network Route to 6.6.6.6 Mask /32. Metric: 1, Next Hop: 6.6.6.6
*Mar  1 00:20:10.307: OSPF: Entered intra-area route sync - area 0
*Mar  1 00:20:10.311: OSPF: ospf_gen_asbr_sum_all_areas
*Mar  1 00:20:10.311: OSPF: running spf for summaries area 0
*Mar  1 00:20:10.311: OSPF: Entered inter-area route sync - area 0
*Mar  1 00:20:10.315: OSPF: Started Building Type 5 External Routes
*Mar  1 00:20:10.315: OSPF: Started Building Type 7 External Routes
*Mar  1 00:20:10.315: OSPF: Entered External route sync - area dummy area
*Mar  1 00:20:10.315: OSPF: Entered NSSA route sync - area 0

OSPF: running SPF for area 0, SPF-type Full

spf - type = full ok i understand this as spf algorithm execution .. do we have even partial spf algorithms running .. that is they might run when a route goes down and you have a backup path may not be in topology table but i think it wont be running algorithm 100% ...

does spf states resemble these two bgp commands "clear ip bgp *" and "clear ip bgp * soft"
  • 0

#22 DarkFiber

DarkFiber

    Cisco Expert

  • Members
  • PipPipPipPip
  • 927 posts
  • Gender:Male
  • Location:Cairo-Egypt
  • Interests:Networks

Posted 19 April 2009 - 10:41 PM

Dear Darkfiber, a very interesting statement you have given here. BEST LSA. Even if we take a very general view of this, it will be correct for type 3 lsa, where every inter area Route corresponds to ONE type3 lsa. But in case of type 1 and type 2, all LSAs are used to form a SPF tree and after that through algorithm best path is selected. So i think this statement is not correct. What do you think ?



That's not 100 % how it works, may you Kindly have a look on my OSPF post , I,ve added more information for the LSA types & cmds.

http://www.sadikhov.com/forum/index.php?showtopic=166538&st=0&gopid=831361&#entry831361
  • 0

#23 moekad

moekad

    Advanced Member

  • Members
  • PipPipPip
  • 270 posts

Posted 20 April 2009 - 12:04 AM

NO, I think, it should be FD 3014400 is the lowest metric from all routes = the best route to router X
RD 2172416 is less than FD 3014400

Another Example:

P 10.1.3.0/24, 1 successors, FD is 2172416
via 10.1.6.3 (2172416/28160), Serial0/1
via 10.1.4.2 (2684416/1794560), Serial0/0

The route to 10.1.3.0 through 10.1.6.3 (Seville) is the successor route, because the
calculated metric (2,172,416), shown as the first of the two numbers in parentheses, is
the best calculated metric.

The route to 10.1.3.0 through 10.1.4.2 (Yosemite) is a feasible successor route, because
the neighbor’s Reported Distance (1,794,560, shown as the second number in
parentheses) is lower than Albuquerque’s FD.

Although both the successor and feasible successor routes are in the EIGRP topology
table, only the successor route is added to the IP routing table.


hey again
mean any route RD is less than FD = feasible successor
and any route FD= to the FD mean 10.1.6.3 FD = to 2172416 FD is the successor route?
Right?
one more thing the feasible successor won't be also the backup as the successor route?
Thanks
  • 0

#24 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 20 April 2009 - 02:09 AM

That's not 100 % how it works, may you Kindly have a look on my OSPF post , I,ve added more information for the LSA types & cmds.

http://www.sadikhov.com/forum/index.php?showtopic=166538&st=0&gopid=831361&#entry831361


Dear Darkfiber, i didnt get which part is not correct ? LSA 3 or LSA 1/2 ? I have seen the post and nodoubt wonderful for review but i missed what i have to look for :-). I know about the types and their related commands :-)

Kindly correct me if i am wrong somewhere
  • 0

#25 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 20 April 2009 - 02:38 AM

Dear Raaki, you cam up with yet another wonderful question :-)

Yes there is another type of spf, not partial but called incremental spf. This is used when only partial network is changed, in that case most of the topology is same so intializing the complete tree is not recommended specially if you have large number of routers in a single area coz in real life networks are flapping a lot (atleast where i am doing job ;-) ), so if at each change spf is formed from scratch, it will introduce a lot of load on cpu. So its better to used incremental spf to only calculate the part that is affected. Its enabled by this command

R1(config-router)#ispf

In that case you will see spf type to be incremental instead of full.

Hope this helps
  • 0

#26 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 20 April 2009 - 02:54 AM

i have forumlated a two direct connected router topology and few loopbacks to test whats going on .. interesting enough .. i have just started ospf process on one router and advertised it networks and i have not even touched the second router here is the output

here are the interfaces

rb(config-router)#do show ip int brief
Interface				  IP-Address	  OK? Method Status				Protocol
FastEthernet0/0			10.0.0.2		YES manual up					up	  
FastEthernet0/1			unassigned	  YES unset  administratively down down	
Loopback1				  4.4.4.4		 YES manual up					up	  
Loopback6				  6.6.6.6		 YES manual up					up

rb(config-router)#
*Mar  1 00:14:35.447: OSPF: Schedule SPF in area 0
	  Change in LS ID 6.6.6.6, LSA type R, , spf-type Full
rb(config-router)#
*Mar  1 00:14:40.447: OSPF: running SPF for area 0, SPF-type Full
*Mar  1 00:14:40.447: OSPF: Initializing to run spf
*Mar  1 00:14:40.447:  OSPF - spf_intra() - rebuilding the tree
*Mar  1 00:14:40.447:  It is a router LSA 6.6.6.6. Link Count 3
*Mar  1 00:14:40.447:   Processing link 0, id 6.6.6.6, link data 255.255.255.255, type 3
*Mar  1 00:14:40.451:	Add better path to LSA ID 6.6.6.6, gateway 6.6.6.6, dist 1
*Mar  1 00:14:40.451:	Add path: next-hop 6.6.6.6, interface Loopback6
*Mar  1 00:14:40.451:   Processing link 1, id 4.4.4.4, link data 255.255.255.255, type 3
*Mar  1 00:14:40.455:	Add better path to LSA ID 4.4.4.4, gateway 4.4.4.4, dist 1
*Mar  1 00:14:40.455:	Add path: next-hop 4.4.4.4, interface Loopback1
*Mar  1 00:14:40.455:   Processing link 2, id 10.0.0.0, link data 255.255.255.0, type 3
*Mar  1 00:14:40.455:	Add better path to LSA ID 10.0.0.255, gateway 10.0.0.0, dist 10
*Mar  1 00:14:40.459:	Add path: next-hop 10.0.0.2, interface FastEt
rb(config-router)#hernet0/0
*Mar  1 00:14:40.459: OSPF: Adding Stub nets
*Mar  1 00:14:40.459: OSPF: Added Network Route to 4.4.4.4 Mask /32. Metric: 1, Next Hop: 4.4.4.4
*Mar  1 00:14:40.459: OSPF: Added Network Route to 6.6.6.6 Mask /32. Metric: 1, Next Hop: 6.6.6.6
*Mar  1 00:14:40.463: OSPF: Added Network Route to 10.0.0.0 Mask /24. Metric: 10, Next Hop: 10.0.0.2
*Mar  1 00:14:40.463: OSPF: Entered intra-area route sync - area 0
*Mar  1 00:14:40.463: OSPF: ospf_gen_asbr_sum_all_areas
*Mar  1 00:14:40.467: OSPF: running spf for summaries area 0
*Mar  1 00:14:40.467: OSPF: Entered inter-area route sync - area 0
*Mar  1 00:14:40.467: OSPF: Started Building Type 5 External Routes
*Mar  1 00:14:40.467: OSPF: Started Building Type 7 External Routes
*Mar  1 00:14:40.467: OSPF: Entered External route sync - area dummy area
*Mar  1 00:14:40.471: OSPF: Entered NSSA route sync - area 0
rb(config-router)#

my question is without any other router in area 0 why router is selecting best path to 6.6.6.6 ?

"Add better path to LSA ID 6.6.6.6, gateway 6.6.6.6, dist 1"

next this is the rouput after intiating ospf in other router

*Mar  1 00:20:10.303: OSPF: Adding Stub nets
*Mar  1 00:20:10.303: OSPF: Added Network Route to 1.1.1.1 Mask /32. Metric: 11, Next Hop: 10.0.0.1
*Mar  1 00:20:10.307: OSPF: Added Network Route to 2.2.2.2 Mask /32. Metric: 11, Next Hop: 10.0.0.1
*Mar  1 00:20:10.307: OSPF: Added Network Route to 4.4.4.4 Mask /32. Metric: 1, Next H
rb(config-router)#op: 4.4.4.4
*Mar  1 00:20:10.307: OSPF: Added Network Route to 6.6.6.6 Mask /32. Metric: 1, Next Hop: 6.6.6.6
*Mar  1 00:20:10.307: OSPF: Entered intra-area route sync - area 0
*Mar  1 00:20:10.311: OSPF: ospf_gen_asbr_sum_all_areas
*Mar  1 00:20:10.311: OSPF: running spf for summaries area 0
*Mar  1 00:20:10.311: OSPF: Entered inter-area route sync - area 0
*Mar  1 00:20:10.315: OSPF: Started Building Type 5 External Routes
*Mar  1 00:20:10.315: OSPF: Started Building Type 7 External Routes
*Mar  1 00:20:10.315: OSPF: Entered External route sync - area dummy area
*Mar  1 00:20:10.315: OSPF: Entered NSSA route sync - area 0

OSPF: running SPF for area 0, SPF-type Full

spf - type = full ok i understand this as spf algorithm execution .. do we have even partial spf algorithms running .. that is they might run when a route goes down and you have a backup path may not be in topology table but i think it wont be running algorithm 100% ...

does spf states resemble these two bgp commands "clear ip bgp *" and "clear ip bgp * soft"


Dear Raaki, a really nice question. But if you look carefully debugs give you the answer to your question ;-)

Ok now see, When spf initializes it takes LSDB and starts its calculation right ? i hope everyone will agree with that, now if you look carefully at router LSDB even when there are no other neighbors, the moment you run ospf it will place its own LSA (router lsa) in its LSDB. Now at the time of running spf whether full or incremental(if it involves its own LSA) router own links will be used in conjucntion with others to form a full spf tree. Now imagine i have a hub with 50 spokes connected, and a network fails on 1 spoke will cause the entire topology to form spf tree from scratch including their own LSAs !!! as you can imagine the biggest local list of LSA would be that of HUB, so if any network flaps anywhere, HUB will be suffering a lot ;-)

I hope this helps, let me know in case of any confusion
  • 0

#27 raaki_88

raaki_88

    BOW DOWN TO THE KING

  • Members
  • PipPipPipPipPip
  • 1442 posts
  • Gender:Male

Posted 20 April 2009 - 03:24 AM

Dear Raaki, a really nice question. But if you look carefully debugs give you the answer to your question ;-)

Ok now see, When spf initializes it takes LSDB and starts its calculation right ? i hope everyone will agree with that, now if you look carefully at router LSDB even when there are no other neighbors, the moment you run ospf it will place its own LSA (router lsa) in its LSDB. Now at the time of running spf whether full or incremental(if it involves its own LSA) router own links will be used in conjucntion with others to form a full spf tree. Now imagine i have a hub with 50 spokes connected, and a network fails on 1 spoke will cause the entire topology to form spf tree from scratch including their own LSAs !!! as you can imagine the biggest local list of LSA would be that of HUB, so if any network flaps anywhere, HUB will be suffering a lot ;-)

I hope this helps, let me know in case of any confusion


thank you bro it helped .. i still have a strange feeling why a concept like confederations which is in bgp is not designed for ospf .. strange thought as i see your hub and spoke topology any ways got cleared ..

regards
  • 0

#28 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 20 April 2009 - 03:44 AM

thank you bro it helped .. i still have a strange feeling why a concept like confederations which is in bgp is not designed for ospf .. strange thought as i see your hub and spoke topology any ways got cleared ..

regards


Dear Raaki, this is where a desighner is born ;-). What you want to achieve through a confederation in ospf ? how do you think it will be helpful ? what problem do you see which can be solved through confederation ?
  • 0

#29 raaki_88

raaki_88

    BOW DOWN TO THE KING

  • Members
  • PipPipPipPipPip
  • 1442 posts
  • Gender:Male

Posted 20 April 2009 - 04:37 AM

Dear Raaki, this is where a desighner is born ;-). What you want to achieve through a confederation in ospf ? how do you think it will be helpful ? what problem do you see which can be solved through confederation ?


well not quite sure though i was playing with bgp for past few weeks ... for each spoke i thought i could split each into one internal and one external area ... in that case if even a network goes down in one spoke it need not run spf on entire of the area outside .. although it can be done with stub and tsa area types just thought over it .. haha might scare you or might look funny just sharing my thoughts
  • 0

#30 pappyaar

pappyaar

    Cisco Routing/EEM/TCL

  • Technical Experts
  • PipPipPipPip
  • 959 posts
  • Gender:Male

Posted 20 April 2009 - 05:29 AM

well not quite sure though i was playing with bgp for past few weeks ... for each spoke i thought i could split each into one internal and one external area ... in that case if even a network goes down in one spoke it need not run spf on entire of the area outside .. although it can be done with stub and tsa area types just thought over it .. haha might scare you or might look funny just sharing my thoughts


Dont worry frnd, nothing scares me and nothing is funny enough to make me laugh ;-) i have already gone through the funniest days of my life.

Anyways its good to have these thoughts :-)
  • 0





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users