Performance end to end testing and comparison between MPLS VPN and VPLS VPN

[hawkeye244]

Hi,

I am student of MSc Network Security and as for my project which is " Comparison between MPLS L3 VPN and VPLS VPN, performance monitoring by end to end testing " I have heard a lot of buzz about VPLS as becoming NGN, I wanted to exppore that and produce a comparison report of which technology is better. To accomplish this I am using GNS3, with respect to the MPLS L3 VPN lab setup that is not a problem but I am stuck at the VPLS part how to setup that ? I have searched but unable to find any cost effective mean, even it is not possible in the university lab as we dont have 7600 series

I would appreciate any support, guidence, advice.

Thanks

Shahbaz

[rsimoni]

Hi Shahbaz,

I am not completely sure I understand your request.

MPLS VPN and VPLS are 2 technologies meant to address to different needs, L3 VPN as opposed as L2 VPN. Not completely sure how you would compare them in terms of performance. Would you compare the performance of a F1 racing car with a Rally racing car?

From the ISP point of view there is little difference (if we don't want to consider the specific inherent peculiarities of each technology) , as in the very basic scenarios we can boil down to the following basic operations for both:

Ingress PE impose 2 labels (at least)

Core Ps swap top most MPLS label

Egress PE removes last label exposing underlying packet or frame.

So whether the LSRs deal with underlying L2 frames or L3 IP packets there is no real difference in terms of performance (actually the P routers don't even notice any difference).

About simulators, I am not aware of anyone able to simulate a L2 VPN (AtoM or VPLS).

Riccardo

[hawkeye244]

Riccardo,

Thanks for your reply, I have done a lot of read up and things started to make sense a bit (a bit). Your F1 and Ferari comparisson comment sets everything in correct order.

My actual wonderness is about whether - Analyse Ethernet VPLS technology potential to be the NGN - if Yes then how and if NO then Y not ....

And for the reason I was trying to set some type of test methods and compare the results.

Thanks

Shahbaz

[Luc De Ghein]

Hi Shahbaz,

If you are serious about testing performance, then you must use real hardware. There is not much point in doing testing with any kind of simulator.

One must also consider the performance of multicast traffic and not only consider unicast traffic.

Best regards,

Luc

[Aabid Saleem]

Hi Shahbaz,

The Dynamips (GNS3) does not support VPLS in the first place, because it cannot simulate more than 7200 series hardware,

It requires 7600 series hardware to be able to run end-to-end.

[hawkeye244]

Hi Aabid,

Thanks for your reply, I know GNS3 does not support VPLS and its limitation. Though it would have been so nice especially for students like me if it did support.

I am trying to use opnet for my testing, but that is not very friendly.

Thanks

Shahbaz

[hawkeye244]

Hi Luc,

Thanks for the advice, though its a bit difficult for me to arrange hardware but I am searching, thanks

Shahbaz