Intelligent path selection O365 (vMX?)


Intelligent path selection O365 (vMX?)

How to achieve intelligent path selection for Office 365 (or other SaaS services)? Is anyone using the vMX for this purpose?


We're considering purchasing Meraki MX appliances to optimize connectivity to (SaaS) cloud services in many of our Global South offices (Africa, Middle East, Asia).


In our testing we found that the MX doesn't failover to secondary connection if the performance on the primary connection gets degraded by packet drops (+6%) and high latency (+300ms) (while degration is detected by Insight WAN Health).


The Cisco Live Session DGTL-BRKCRS-1579 mentions multiple smart SaaS optimization features are coming (Optimal Path) but for now the only way to achieve performance based routing is using Custom Performance Classes in combination with autoVPN.


Is it advisable to put a vMX in Azure optimize O365 and failover in case of degraded internet connectivity?

2 Replies 2
Kind of a big deal
Kind of a big deal

The only thing the VMX does is provide access to private resources in either Azure or Amazon AWS (depending on where you host it).  If this is your actual case, and your remote on-premise sites have dual WAN circuits, then it can choose optimal paths using those WAN circuits.


It does nothing for accessing resources outside of those environments (such as Office 365).

Thanks Philip,


That's what I was afraid of. I was hoping there could be a way to use the vMX as a breakout for O365 traffic, but then also realise that this would probably not adhere to the Microsoft Connectivity Principles.


It seems Citrix SD-WAN is a bit further advanced in this area, so hope that Cisco Meraki will catch up soon to deliver optimal path selection for SaaS applications. However, I've not heard any ETA on this yet.

Get notified when there are additional replies to this discussion.
Welcome to the Meraki Community!
To start contributing, simply sign in with your Cisco account. If you don't yet have a Cisco account, you can sign up.