Design Considerations

Hi,

  I have a few queries. I would appreciate if you could help me in clearing these doubts


1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

"1. On slow links configure LFI.

2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.

3. If voice, use Header Compression.

Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-

mls ip cef load-sharing full

It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
 
I could not understand this concept.

3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?

4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.

Comments

  • Question 1

     

    When deciding on  the compression needed on a slow speed WAN link - you look at the traffic that is the QoS concern for the link. If you are dealing with primaily voice on the circuit - you use RTP Header Compression - this is because the the Voice traffic in the packet is already highly compressed so trying to compress it more will be a waste of resources. Compressing the headers in this case is the only thing of benefit.

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram
    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts


    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.

    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-

    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?

    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx

  • Thanks Anthony!!!

    On Sun, Nov 21, 2010 at 1:15 AM, Anthony Sequeira <[email protected]> wrote:

    Question 1

     

    When deciding on  the compression needed on a slow speed WAN link - you look at the traffic that is the QoS concern for the link. If you are dealing with primaily voice on the circuit - you use RTP Header Compression - this is because the the Voice traffic in the packet is already highly compressed so trying to compress it more will be a waste of resources. Compressing the headers in this case is the only thing of benefit.

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts



    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.


    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-


    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?


    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/124918.aspx#124918



    Vikram
    CCIE#22735

    Phone: 001-520-834-7675
  • Regarding question 2 - this command is recommended to ensure the fastest possible load balancing.

     

    Remember that load balancing is a function of the Route Processor selecting the best paths that can be load balanced against, but then the CEF engine actual performs the load balancing. So this command is ensureing that layer of the mechanics is working as efficiently as possible.

     

    http://www.cisco.com/en/US/products/hw/switches/ps700/products_configuration_example09186a00800ab513.shtml

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram
    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts


    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.

    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-

    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?

    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx

  • Hi Anthony,


      Thanks for the mail. This link was helpful. Also got to know about the other recommendation:

    In order to achieve the best
    CEF load balancing, alternate L3 and L4 hashing on access, distribution and
    core routers, and use this type of configuration:

    • On access and core routers - mls ip cef load-sharing
      simple

    • On distribution routers - mls ip cef load-sharing
      full

    Are these commands still recommended ( since I cannot see them in my distribution/access/core switches configs in the production designed by Cisco engineers) or have become redundant?



    On Mon, Nov 22, 2010 at 4:53 PM, Anthony Sequeira <[email protected]> wrote:

    Regarding question 2 - this command is recommended to ensure the fastest possible load balancing.

     

    Remember that load balancing is a function of the Route Processor selecting the best paths that can be load balanced against, but then the CEF engine actual performs the load balancing. So this command is ensureing that layer of the mechanics is working as efficiently as possible.

     

    http://www.cisco.com/en/US/products/hw/switches/ps700/products_configuration_example09186a00800ab513.shtml

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts



    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.


    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-


    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?


    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/124970.aspx#124970



    Vikram
    CCIE#22735

    Phone: 001-520-834-7675
  • Yeah - great point - these commands are not needed now on a lot of newer platforms and code - the optimizations are built in.

    Sent from my iPhone

    On Nov 22, 2010, at 7:40 PM, "parvikram" <[email protected]> wrote:

    Hi Anthony,


      Thanks for the mail. This link was helpful. Also got to know about the other recommendation:

    In order to achieve the best
    CEF load balancing, alternate L3 and L4 hashing on access, distribution and
    core routers, and use this type of configuration:

    • On access and core routers - mls ip cef load-sharing
      simple

    • On distribution routers - mls ip cef load-sharing
      full

    Are these commands still recommended ( since I cannot see them in my distribution/access/core switches configs in the production designed by Cisco engineers) or have become redundant?



    On Mon, Nov 22, 2010 at 4:53 PM, Anthony Sequeira <[email protected]> wrote:

    Regarding question 2 - this command is recommended to ensure the fastest possible load balancing.

     

    Remember that load balancing is a function of the Route Processor selecting the best paths that can be load balanced against, but then the CEF engine actual performs the load balancing. So this command is ensureing that layer of the mechanics is working as efficiently as possible.

     

    http://www.cisco.com/en/US/products/hw/switches/ps700/products_configuration_example09186a00800ab513.shtml

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts



    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.


    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-


    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?


    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/124970.aspx#124970



    Vikram
    CCIE#22735

    Phone: 001-520-834-7675



    Internetwork Expert - The Industry Leader in CCIE Preparation

    http://www.internetworkexpert.com



    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx
  • Regarding question 3 - this will vary from switch to switch. Unfortunately, the QoS settings vary from switch to switch, so you want to check the commands and capabilities of your switch very carefully.

     

    But - you are on the right track for sure with your logic. A Cisco IP Phone is able to mark the traffic at Layer 2 (Cos) and at Layer 3 (DSCP). So if your switchport is able to trust the DSCP markings, you enable that behaviour and send them on to the next hop router….

     

    Regarding question 4 - yes - we like to do an area boundary between core and distribution that promotes summarization and future scalability.

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram
    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts


    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.

    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-

    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?

    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx

  • Thanks Anthony. This is what I understood. Please correct me if I'm wrong. If we mark at layer 2 (Cos) , then l2 ports have no issues but at l3 ports we can configure the ports to trust the Cos and it will pick the corresponding value from the CosToDscp mappings. And if we mark at layer 3(DSCP), then l3 ports are fine but for the l2 ports,  we can configure them to trust the DSCP markings and will pick the value from the DSCP-cos mappings?


    On Tue, Nov 23, 2010 at 8:42 AM, Anthony Sequeira <[email protected]> wrote:

    Regarding question 3 - this will vary from switch to switch. Unfortunately, the QoS settings vary from switch to switch, so you want to check the commands and capabilities of your switch very carefully.

     

    But - you are on the right track for sure with your logic. A Cisco IP Phone is able to mark the traffic at Layer 2 (Cos) and at Layer 3 (DSCP). So if your switchport is able to trust the DSCP markings, you enable that behaviour and send them on to the next hop router….

     

    Regarding question 4 - yes - we like to do an area boundary between core and distribution that promotes summarization and future scalability.

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts



    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.


    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-


    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?


    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/124993.aspx#124993



    Vikram
    CCIE#22735

    Phone: 001-520-834-7675
  • Yeah - you really want to check the documenation for QoS trusting on your particular platform. Also - check the default map behaviour.

     

    What is nice is the fact that on a Multilayer Switch (like the 3560 for example) - the device can indeed trust L3 markings (DSCP) on a layer 2 port.

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram
    Sent: Tuesday, November 23, 2010 12:02 PM
    To: Anthony Sequeira
    Subject: Re: [ccdp] Design Considerations

     

    Thanks Anthony. This is what I understood. Please correct me if I'm wrong. If we mark at layer 2 (Cos) , then l2 ports have no issues but at l3 ports we can configure the ports to trust the Cos and it will pick the corresponding value from the CosToDscp mappings. And if we mark at layer 3(DSCP), then l3 ports are fine but for the l2 ports,  we can configure them to trust the DSCP markings and will pick the value from the DSCP-cos mappings?

    On Tue, Nov 23, 2010 at 8:42 AM, Anthony Sequeira <[email protected]> wrote:

    Regarding question 3 - this will vary from switch to switch. Unfortunately, the QoS settings vary from switch to switch, so you want to check the commands and capabilities of your switch very carefully.

     

    But - you are on the right track for sure with your logic. A Cisco IP Phone is able to mark the traffic at Layer 2 (Cos) and at Layer 3 (DSCP). So if your switchport is able to trust the DSCP markings, you enable that behaviour and send them on to the next hop router….

     

    Regarding question 4 - yes - we like to do an area boundary between core and distribution that promotes summarization and future scalability.

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram
    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts


    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.

    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-

    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?

    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/124993.aspx#124993


    Vikram
    CCIE#22735
    Phone: 001-520-834-7675




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx

  • Hi Anthony,

      Thanks a lot for taking time out and answering all my doubts!!!

    Thanks ,
    Vikram



    On Tue, Nov 23, 2010 at 12:38 PM, Anthony Sequeira <[email protected]> wrote:

    Yeah - you really want to check the documenation for QoS trusting on your particular platform. Also - check the default map behaviour.

     

    What is nice is the fact that on a Multilayer Switch (like the 3560 for example) - the device can indeed trust L3 markings (DSCP) on a layer 2 port.

     

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Tuesday, November 23, 2010 12:02 PM
    To: Anthony Sequeira
    Subject: Re: [ccdp] Design Considerations

     


    Thanks Anthony. This is what I understood. Please correct me if I'm wrong. If we mark at layer 2 (Cos) , then l2 ports have no issues but at l3 ports we can configure the ports to trust the Cos and it will pick the corresponding value from the CosToDscp mappings. And if we mark at layer 3(DSCP), then l3 ports are fine but for the l2 ports,  we can configure them to trust the DSCP markings and will pick the value from the DSCP-cos mappings?

    On Tue, Nov 23, 2010 at 8:42 AM, Anthony Sequeira <[email protected]> wrote:

    Regarding question 3 - this will vary from switch to switch. Unfortunately, the QoS settings vary from switch to switch, so you want to check the commands and capabilities of your switch very carefully.

     

    But - you are on the right track for sure with your logic. A Cisco IP Phone is able to mark the traffic at Layer 2 (Cos) and at Layer 3 (DSCP). So if your switchport is able to trust the DSCP markings, you enable that behaviour and send them on to the next hop router….

     

    Regarding question 4 - yes - we like to do an area boundary between core and distribution that promotes summarization and future scalability.

    From: [email protected] [mailto:[email protected]] On Behalf Of parvikram

    Sent: Wednesday, November 17, 2010 1:10 PM
    To: Anthony Sequeira
    Subject: [ccdp] Design Considerations

     

    Hi,

      I have a few queries. I would appreciate if you could help me in clearing these doubts



    1. Regarding Link Efficiency Mechanisms. In the books I could find only this in a nutshell  --

    "1. On slow links configure LFI.

    2. If using HDLC or PPP use link payload compression, if using ATM or Frame relay , use layer 2 payload compression.


    3. If voice, use Header Compression.

    Can we use both Layer 2 payload compression and Header compression on WAN links? What is recommended?

    2. On the distribution switches, why is this command recommended as a good design practice( referring to the Net design guide by Cisco from 1995)-


    mls ip cef load-sharing full

    It was talking something about Left/Left , Right/Right path taken,L3/L4 hash and load sharing not taken care of properly.
     
    I could not understand this concept.

    3. On which ports is it recommended to apply mls qos trust cos and on which ports is it recommended to apply mls qos trust DSCP? Does it depend on whether we have marked Cos and DSCP and if the ports are layer 2 or layer 3. I mean after marking on access layer, on the upper layers (dist/core), how do we determine whether to trust cos or dscp?


    4.When desgining OSPF is it recommended to have the area 0 between Distribution and Core. This question stems from a reading which says that to reduce SPF and LSA load to area 0 , summarize on distribution to core.




    Internetwork Expert - The Industry Leader in CCIE Preparation
    http://www.internetworkexpert.com

    Subscription information may be found at:

    http://www.ieoc.com/forums/ForumSubscriptions.aspx




    Vikram
    CCIE#22735
    Phone: 001-520-834-7675




    Internetwork Expert - The Industry Leader in CCIE Preparation

    http://www.internetworkexpert.com

    Subscription information may be found at:
    http://www.ieoc.com/forums/ForumSubscriptions.aspx



    --
    View this message online at: http://127.0.0.1/forums/p/13845/125005.aspx#125005



    Vikram
    CCIE#22735

    Phone: 001-520-834-7675
  • You have provided an nice article, Thank you very much for this one. And i hope this will be useful for many people gmail login

Sign In or Register to comment.