Head end router will chose both the path. Look at this case below.
PE1------P1------PE2
| |
|_____P2----------|
Now say all links b/w PE1 and PE2 are ten gig interface. ospf cost to reach
PE2 via P1 or P2 is same.
We create two tunnels T1 via P1 and T2 via P2.
once tunnels are up, PE1 will load balance packets to PE2 by using both
tunnels.
Thanks,
Sonu
On Sun, Jan 29, 2012 at 11:36 PM, CCIE KID <eliteccie_at_gmail.com> wrote:
> Gaurav
>
> Then wat s the significance of Metric command in TE ?
>
>
> On Sun, Jan 29, 2012 at 11:14 PM, GAURAV MADAN <gauravmadan1177_at_gmail.com>wrote:
>
>> There can be multiple tunnels between a given src and a given dest.
>> If all properties of tunnel are same .... This will result in equal path
>> load balancing
>> If properties like b/w are diff .... Unequal path load balancing will
>> take place ..
>>
>>
>>
>> On Sun, Jan 29, 2012 at 11:02 PM, CCIE KID <eliteccie_at_gmail.com> wrote:
>>
>>> Sonu
>>>
>>> But the Head end Router l will choose only one path to the same Tail End
>>> Router with only one Tunnel. Why it chooses two Tunnels ?
>>>
>>>
>>>
>>> On Sun, Jan 29, 2012 at 10:12 PM, Sonu Khandelwal <sonu.kwl_at_gmail.com>wrote:
>>>
>>>> Yes, for two path to the same destination two tunnels will be needed
>>>> with two different explicit path.
>>>>
>>>> "tunnel mpls traffic-eng autoroute metric" to used to change the igp
>>>> metric of the tunnel interface. By default tunnel's igp metric is the same
>>>> as igp metric b/w source and destination.
>>>>
>>>> for e.g
>>>>
>>>> if tunnel source is 1.1.1.1 and tunnel destination is 3.3.3.3. Assume
>>>> tunnel is not present and cost to reach from 1.1.1.1 to 3.3.3.3 is 10.
>>>>
>>>> Now when tunnel comes up, its metric will be 10.
>>>>
>>>> if you want this metric of 10 to be changed then you can use "tunnel
>>>> mpls traffic-eng autoroute metric" to do the same thing.
>>>>
>>>> Thanks,
>>>> Sonu
>>>>
>>>>
>>>> On Sun, Jan 29, 2012 at 9:27 PM, CCIE KID <eliteccie_at_gmail.com> wrote:
>>>>
>>>>> Gaurav
>>>>>
>>>>> Dude one new doubt man
>>>>>
>>>>> To reach a particular Tail end router, there should be only route.
>>>>>
>>>>> In regular SPF, the result of path computation may be two equal cost
>>>>> multipaths (ECMP) to a destination. However, the result of CSPF is always
>>>>> one path to the tailend router. In case, two paths are equal, CSPF has 3
>>>>> rules followed sequentially to break the tie:
>>>>>
>>>>> *a.* Take the path with largest minimum available bandwidth
>>>>>
>>>>> *b.* Take the path with the lowest hop-count (number of routers in
>>>>> the path)
>>>>>
>>>>> *c.* Take one path at random.
>>>>>
>>>>>
>>>>> So there should be only one path to a TAIL END router.. How does it
>>>>> has two paths for a same TAILEND router. via two different TE TUNNELS.
>>>>>
>>>>>
>>>>>
>>>>> On Sun, Jan 29, 2012 at 8:44 PM, CCIE KID <eliteccie_at_gmail.com> wrote:
>>>>>
>>>>>> Gaurav
>>>>>>
>>>>>> Tats gr8 mate .. I will also check the purpose of it and let u know
>>>>>> buddy..
>>>>>> Till then study hard
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Sun, Jan 29, 2012 at 8:33 PM, GAURAV MADAN <
>>>>>> gauravmadan1177_at_gmail.com> wrote:
>>>>>>
>>>>>>> Kid
>>>>>>>
>>>>>>> Everything you said is exactly what i understand this topic :)
>>>>>>> We are on same page .
>>>>>>>
>>>>>>> I will simulate the usage of metric command and try to post results ,
>>>>>>>
>>>>>>> Till then .. thnx
>>>>>>>
>>>>>>> Gaurav Madan
>>>>>>>
>>>>>>>
>>>>>>> On Sun, Jan 29, 2012 at 7:39 PM, CCIE KID <eliteccie_at_gmail.com>wrote:
>>>>>>>
>>>>>>>> Gaurav
>>>>>>>>
>>>>>>>> TE Auto Route Announce will just make ur OSPF learned LSA's from
>>>>>>>> the Tunnel Tail End to utitlize the Tunnel path to reach all the networks
>>>>>>>> behind the Tunnel Tail End Router. It has nothing to do with priority
>>>>>>>> levels for each tunnel
>>>>>>>> U can assign priorities to each tunnel and then tat take over the
>>>>>>>> appropriate B.W. available in the interface if it has highest priority
>>>>>>>>
>>>>>>>> In ur case both ur tunnels has both set up and hold priority has 7
>>>>>>>> and 7
>>>>>>>> 7 and 7 is the least priority tat u can assign to a tunnel and both
>>>>>>>> the tunnels has the same priority and both r pointing to the same tunnel
>>>>>>>> tail end and also u r running Auto Route Announce on both the tunnels
>>>>>>>>
>>>>>>>> But the only difference between the two tunnels is the Bandwidth
>>>>>>>> and to form the two tunnels it requires 60 and 70 kbps end to end to form
>>>>>>>> the tunnel. So atleast the B.W of the end to end path should be atleast
>>>>>>>> 130kpbs
>>>>>>>>
>>>>>>>> So now OSPF will carry the TE information like the requested B.W
>>>>>>>> and also other TE colors associated with TE in its Opaque LSA Type 9 ,10
>>>>>>>> and 11. depends upon it is a Single Area OSPF or Multi Area and also it is
>>>>>>>> within the Single AS. If the Head end sees that it has enough B.W. then the
>>>>>>>> Path message from Head end to Tail End and Reserve message from the Tail
>>>>>>>> End to head End is succesfull , the tunnel will be up.
>>>>>>>>
>>>>>>>> if there is any change in the available BW in that particular end
>>>>>>>> to end path, it will be conveyed by OSPF and it triggers the Path message
>>>>>>>> and Reserve message
>>>>>>>>
>>>>>>>> In ur setup
>>>>>>>>
>>>>>>>> U r obviously u r going to see two paths to reach the tail-end
>>>>>>>> router via these two tunnels and there will be Unequal cost LOAD balancing
>>>>>>>> . If u want to prioritize one tunnel over another one , just change the
>>>>>>>> metric of the tunnel rather than the BW
>>>>>>>> Because B.W inside the tunnel doesnt reflect ur BEST METRIC to
>>>>>>>> reach a particular TE Tail End
>>>>>>>>
>>>>>>>> So ultimately u have many paths to the Same tail end router ie via
>>>>>>>> different tunnels.So ultimately it all depends upon how much BW is required
>>>>>>>> to form the tunnel . Int his case , there will 70:60 ratio of traffic going
>>>>>>>> from one tunnel to another tunnel.
>>>>>>>>
>>>>>>>> If u dont want to see Unequal Cost Load Balancing, u have to do
>>>>>>>> Policy Routing rather than using AutoRoute Announce.
>>>>>>>>
>>>>>>>> But I am not sure about this command
>>>>>>>>
>>>>>>>> tunnel mpls traffic-eng autoroute metric
>>>>>>>>
>>>>>>>> If u give this command whether one tunnel will be preferred over
>>>>>>>> another tunnel if ur Tail End Router address are same.
>>>>>>>>
>>>>>>>> Post me about this command dude
>>>>>>>>
>>>>>>>> Hope this helps
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Sun, Jan 29, 2012 at 1:24 AM, Naveen <navin.ms_at_gmail.com> wrote:
>>>>>>>>
>>>>>>>>> Yes. What you have is already doing UELB (unequal cost load
>>>>>>>>> balancing in
>>>>>>>>> the ratio 6:7) also across TE tunnels. Try sending some traffic
>>>>>>>>> end to end,
>>>>>>>>> and check the TE counters.
>>>>>>>>>
>>>>>>>>> If your intention is not to load balancing at all, then remove
>>>>>>>>> "autoroute
>>>>>>>>> announce", and then route traffic to the tunnel you want with a
>>>>>>>>> local
>>>>>>>>> policy.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Sat, Jan 28, 2012 at 11:31 AM, GAURAV MADAN <
>>>>>>>>> gauravmadan1177_at_gmail.com>wrote:
>>>>>>>>>
>>>>>>>>> > Naveen ,
>>>>>>>>> >
>>>>>>>>> > Thnx for the reply ..
>>>>>>>>> > Do you mean to suggest some kind of unequal load-balancing
>>>>>>>>> across various
>>>>>>>>> > TE tunnels .
>>>>>>>>> >
>>>>>>>>> > Between the tunnels ; I can choose using following command :
>>>>>>>>> >
>>>>>>>>> > Router(config-if)#tunnel mpls traffic-eng autoroute metric
>>>>>>>>> >
>>>>>>>>> > I can set one tunnel as lower metric to be announced over the
>>>>>>>>> second
>>>>>>>>> > tunnel . and hence play around with this
>>>>>>>>> >
>>>>>>>>> > Is that what u r suggesting ?
>>>>>>>>> >
>>>>>>>>> >
>>>>>>>>> > Gaurav Madan
>>>>>>>>> >
>>>>>>>>> >
>>>>>>>>> >
>>>>>>>>> > On Sun, Jan 29, 2012 at 12:32 AM, Naveen <navin.ms_at_gmail.com>
>>>>>>>>> wrote:
>>>>>>>>> >
>>>>>>>>> >> What you are seeing is Load balancing across multiple TE
>>>>>>>>> tunnels for
>>>>>>>>> >> prefixes reachable further down the tunnel. This is an expected
>>>>>>>>> behavior
>>>>>>>>> >> with "autoroute announce". You can use "auto-bw" to dynamically
>>>>>>>>> control the
>>>>>>>>> >> TE tunnel Bandwidth after tunnel setup.
>>>>>>>>> >>
>>>>>>>>> >> Once tunnels are established (using CSPF), IGP considers the
>>>>>>>>> Tunnels as
>>>>>>>>> >> always 1-hop (and always prefers over any other IGP 1-hop
>>>>>>>>> path). You can
>>>>>>>>> >> load balance upto 8,16, or 32 TE paths as supported by the
>>>>>>>>> platform.
>>>>>>>>> >>
>>>>>>>>> >>
>>>>>>>>> >>
>>>>>>>>> >> On Sat, Jan 28, 2012 at 10:13 AM, Sonu Khandelwal <
>>>>>>>>> sonu.kwl_at_gmail.com>wrote:
>>>>>>>>> >>
>>>>>>>>> >>> [image: Boxbe] <https://www.boxbe.com/overview> Sonu
>>>>>>>>> Khandelwal (
>>>>>>>>> >>> sonu.kwl_at_gmail.com) is not on your Guest List<
>>>>>>>>> https://www.boxbe.com/approved-list>| Approve
>>>>>>>>> >>> sender <https://www.boxbe.com/anno?tc=10557674166_204049186>
>>>>>>>>> | Approve
>>>>>>>>> >>> domain <
>>>>>>>>> https://www.boxbe.com/anno?tc=10557674166_204049186&dom>
>>>>>>>>>
>>>>>>>>> >>>
>>>>>>>>> >>> on interface it means how much bandwidth can be reserved by
>>>>>>>>> RSVP.
>>>>>>>>> >>> on tunnel it means, how much bandwidth is required by tunnel.
>>>>>>>>> >>>
>>>>>>>>> >>> Hope it helps.
>>>>>>>>> >>>
>>>>>>>>> >>> Thanks,
>>>>>>>>> >>> Sonu
>>>>>>>>> >>>
>>>>>>>>> >>> On Sat, Jan 28, 2012 at 11:37 PM, GAURAV MADAN <
>>>>>>>>> >>> gauravmadan1177_at_gmail.com>wrote:
>>>>>>>>> >>>
>>>>>>>>> >>> > Then what does "rsvp bandwidth <> " on interface does ?
>>>>>>>>> >>> > That is also a constraint on interface to establish the
>>>>>>>>> tunnel ..
>>>>>>>>> >>> right ?
>>>>>>>>> >>> >
>>>>>>>>> >>> > How do you differentiate between the BW that we specify on
>>>>>>>>> interface
>>>>>>>>> >>> level
>>>>>>>>> >>> > and the one that we specify on interface tunnel ?
>>>>>>>>> >>> >
>>>>>>>>> >>> > Thnx
>>>>>>>>> >>> > Gaurav Madan
>>>>>>>>> >>> >
>>>>>>>>> >>> > On Sat, Jan 28, 2012 at 9:23 PM, Paul Negron <
>>>>>>>>> negron.paul_at_gmail.com>
>>>>>>>>> >>> > wrote:
>>>>>>>>> >>> >
>>>>>>>>> >>> > > The bandwidth is a constraint. It only says " I need 60K of
>>>>>>>>> >>> bandwidth in
>>>>>>>>> >>> > > order to establish this tunnel". One tunnel requires 60K
>>>>>>>>> and the
>>>>>>>>> >>> other
>>>>>>>>> >>> > > requires 70K. If the bandwidth is available for both, they
>>>>>>>>> will both
>>>>>>>>> >>> be
>>>>>>>>> >>> > > established. That "bandwidth" constraint has nothing to do
>>>>>>>>> with the
>>>>>>>>> >>> cost
>>>>>>>>> >>> > of
>>>>>>>>> >>> > > the tunnel for route selection.
>>>>>>>>> >>> > > --
>>>>>>>>> >>> > > Paul Negron
>>>>>>>>> >>> > > CCIE# 14856 CCSI# 22752
>>>>>>>>> >>> > > Senior Technical Instructor
>>>>>>>>> >>> > >
>>>>>>>>> >>> > >
>>>>>>>>> >>> > >
>>>>>>>>> >>> > > > From: GAURAV MADAN <gauravmadan1177_at_gmail.com>
>>>>>>>>> >>> > > > Reply-To: GAURAV MADAN <gauravmadan1177_at_gmail.com>
>>>>>>>>> >>> > > > Date: Sat, 28 Jan 2012 20:44:46 +0530
>>>>>>>>> >>> > > > To: Cisco certification <ccielab_at_groupstudy.com>
>>>>>>>>> >>> > > > Subject: MPLS TE Tunnel : Tunnel B/W doubt
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Hi All
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > I am running in some issue to understand how Tunnel
>>>>>>>>> Bandwidth
>>>>>>>>> >>> works .
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Topology
>>>>>>>>> >>> > > > ---------
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > R1 R4
>>>>>>>>> >>> > > > | |
>>>>>>>>> >>> > > > | |
>>>>>>>>> >>> > > > R2-----------------R3
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > R1 is head end of tunnel . It has 2 tunnels Tunnel0 and
>>>>>>>>> Tunnel 1 as
>>>>>>>>> >>> > > follows
>>>>>>>>> >>> > > > :
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > R1#sh run int tun 0
>>>>>>>>> >>> > > > Building configuration...
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Current configuration : 284 bytes
>>>>>>>>> >>> > > > !
>>>>>>>>> >>> > > > interface Tunnel0
>>>>>>>>> >>> > > > ip unnumbered Loopback0
>>>>>>>>> >>> > > > tunnel mode mpls traffic-eng
>>>>>>>>> >>> > > > tunnel destination 4.4.4.4
>>>>>>>>> >>> > > > tunnel mpls traffic-eng autoroute announce
>>>>>>>>> >>> > > > tunnel mpls traffic-eng priority 7 7
>>>>>>>>> >>> > > > tunnel mpls traffic-eng bandwidth 70
>>>>>>>>> >>> > > > tunnel mpls traffic-eng path-option 1 dynamic lockdown
>>>>>>>>> >>> > > > end
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > R1#sh run int tun 1
>>>>>>>>> >>> > > > Building configuration...
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Current configuration : 284 bytes
>>>>>>>>> >>> > > > !
>>>>>>>>> >>> > > > interface Tunnel1
>>>>>>>>> >>> > > > ip unnumbered Loopback0
>>>>>>>>> >>> > > > tunnel mode mpls traffic-eng
>>>>>>>>> >>> > > > tunnel destination 4.4.4.4
>>>>>>>>> >>> > > > tunnel mpls traffic-eng autoroute announce
>>>>>>>>> >>> > > > tunnel mpls traffic-eng priority 7 7
>>>>>>>>> >>> > > > tunnel mpls traffic-eng bandwidth 60
>>>>>>>>> >>> > > > tunnel mpls traffic-eng path-option 1 dynamic lockdown
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Now ; one tunnel has bandwidth of 70Kbps and other has
>>>>>>>>> of 60Kbps .
>>>>>>>>> >>> > > > Following is my show ip route ospf output :
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > 4.0.0.0/32 is subnetted, 1 subnets
>>>>>>>>> >>> > > > O 4.4.4.4 [110/4] via 4.4.4.4, 00:05:52, Tunnel1
>>>>>>>>> >>> > > > [110/4] via 4.4.4.4, 00:05:52, Tunnel0
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > I see both entries there ...
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Is there a reason for this ?
>>>>>>>>> >>> > > > Please let me know .. I think that it should be routing
>>>>>>>>> via Tunnel
>>>>>>>>> >>> 0
>>>>>>>>> >>> > > only .
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Thanks
>>>>>>>>> >>> > > > Gaurav Madan
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > > Blogs and organic groups at http://www.ccie.net
>>>>>>>>> >>> > > >
>>>>>>>>> >>> > > >
>>>>>>>>> >>>
>>>>>>>>> _______________________________________________________________________
>>>>>>>>> >>> > > > Subscription information may be found at:
>>>>>>>>> >>> > > > http://www.groupstudy.com/list/CCIELab.html
>>>>>>>>> >>> >
>>>>>>>>> >>> >
>>>>>>>>> >>> > Blogs and organic groups at http://www.ccie.net
>>>>>>>>> >>> >
>>>>>>>>> >>> >
>>>>>>>>> _______________________________________________________________________
>>>>>>>>> >>> > Subscription information may be found at:
>>>>>>>>> >>> > http://www.groupstudy.com/list/CCIELab.html
>>>>>>>>> >>>
>>>>>>>>> >>>
>>>>>>>>> >>> Blogs and organic groups at http://www.ccie.net
>>>>>>>>> >>>
>>>>>>>>> >>>
>>>>>>>>> _______________________________________________________________________
>>>>>>>>> >>> Subscription information may be found at:
>>>>>>>>> >>> http://www.groupstudy.com/list/CCIELab.html
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Blogs and organic groups at http://www.ccie.net
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> _______________________________________________________________________
>>>>>>>>> Subscription information may be found at:
>>>>>>>>> http://www.groupstudy.com/list/CCIELab.html
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> With Warmest Regards,
>>>>>>>>
>>>>>>>> CCIE KID
>>>>>>>> CCIE#29992 (Security)
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> With Warmest Regards,
>>>>>>
>>>>>> CCIE KID
>>>>>> CCIE#29992 (Security)
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> With Warmest Regards,
>>>>>
>>>>> CCIE KID
>>>>> CCIE#29992 (Security)
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> With Warmest Regards,
>>>
>>> CCIE KID
>>> CCIE#29992 (Security)
>>>
>>>
>>>
>>
>
>
> --
> With Warmest Regards,
>
> CCIE KID
> CCIE#29992 (Security)
Blogs and organic groups at http://www.ccie.net
Received on Mon Jan 30 2012 - 10:52:44 ART
This archive was generated by hypermail 2.2.0 : Thu Feb 02 2012 - 11:52:52 ART