How it reduced from 60ms to 2ms

[AnkitSharma1]

Team,

We have two sites in Mexico; one is a data center and the other is a different site. What I have noticed is that before applying the policy below, I was getting a 60 ms latency from source to destination, but now I see a drastic reduction. I want to understand the reason for this change. If you can provide an answer, I would appreciate it. For your information, we have two WAN connections (WAN1 and WAN2), load balancing enabled, and auto VPN enabled.

[BlakeRichardson]

Was VPN traffic previously using WAN 2? If so have you tried running a trace route to see if there is a bottle neck somewhere? 

[AnkitSharma1]

It was using both WAN because load balance is enabled 

[Ryan_Miles]

Depending on traffic types flows will be kept on the WAN it was initiated on. Load balancing doesn't implicitly mean all traffic will always be spread across both WAN ports.

This is mentioned in the following doc

https://documentation.meraki.com/MX/Firewall_and_Traffic_Shaping/MX_Load_Balancing_and_Flow_Preferen... 

[RWelch]

SD-WAN Internet Policies (SD-Internet) 

I believe this document might help explain it.

[RWelch]

@Ryan_Miles might be able to chime in/shed some light on it?

[Ryan_Miles]

Not much info to go off of. I would ask the same as @BlakeRichardson as it sounds like after the config change traffic is taking a different WAN path.

[RWelch]

Dynamic Path Selection is what I believe reduces the latency.