cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2078
Views
5
Helpful
5
Replies

IP Precedence vs DSCP in MPLS VPN

maher
Level 1
Level 1

Hi there,

From my reading through Cisco website, I noticed that most of implementation suggested to use DSCP values and sometimes use IP Precedence in CE/PE class-map for classfication and markings.

What is the most appropriate or maybe best practise to configure by using IP Precedence or DSCP? Is there any difference in terms of providing QoS especially in MPLS VPN enviroment?

Thanks.

maher

1 Accepted Solution

Accepted Solutions

Hey Maher,

I should add that the PE will receive IP traffic from the CE with DSCP markings set by the CE. Based on the DSCP markings on the packet, the router will determine the EXP value for the MPLS label(s) that are slapped on to the IP packet...

On egress, the PE will strip off the MPLS label and the original DSCP markings on the IP packet will still be present...

Hope that helps - pls rate the post if it does.

Paresh

View solution in original post

5 Replies 5

pkhatri
Level 11
Level 11

Hi Maher,

DiffServ offers you the advantage of a 6-bit field so that you can offer much more granular QoS. IP Precedence only offers you 3 bits so that the QoS offered is fairly coarse. These days, most implementations would be using DSCP for that reason.

Within an MPLS core, the QoS setting is carried within the EXP field, which is a 3-bit field. Therefore, even if you use 6-bits for DSCP at the edge, you will end up having to map those 6-bits into 3-bits for carriage across the MPLS network. That is not a problem, however. You are bound to get more congestion at the edges so you need more granular QoS markings there. The core of a network is unlikely to be congested so 3 bits of QoS marking is sufficient.

Hope that helps - pls rate the post if it does.

Paresh

Hi Paresh,

Thanks for the info. If let say, I want to implement DiffServ, from CE2PE, PE2PE, I need to mark/classify intersting packet into DSCP value right?

thanks again.

maher

In the CE2PE direction, you would mark and classify based on the DSCP values. In the PE2PE direction, you would mark and classify based on the EXP values in the MPLS header.

Paresh

Hey Maher,

I should add that the PE will receive IP traffic from the CE with DSCP markings set by the CE. Based on the DSCP markings on the packet, the router will determine the EXP value for the MPLS label(s) that are slapped on to the IP packet...

On egress, the PE will strip off the MPLS label and the original DSCP markings on the IP packet will still be present...

Hope that helps - pls rate the post if it does.

Paresh

Paresh,

Thanks for the help