Does Local Egress Always Provide Best Office 365 User Experience? Our Data Says No.

I have been working with the Microsoft cloud since the early days, when it was called BPOS (the Business Productivity Online Suite). BPOS became Office 365, which we all know and love. And as you may have heard, it’s name change time again, albeit a minor change.

I am a proponent of doing things properly. So much so that I have been accused of being a Microsoft “Fan Boy” many times. First let me say to those who apply this Microsoft “Fan Boy” label to me, of course I am. But being a fan does not mean that I take everything at face value. I like data and I like facts. And I like to build out environments and run test sets to isolate metrics and find the facts.

Local Egress can be easiest described as getting on the internet as fast as possible, and not using corporate networks or VPN tunnels for connectivity. It is a topic that I have covered before. At Microsoft Ignite 2017 I was quoted in BRK3041 as saying that no matter how we structured a scenario, getting to the Internet, and then your local ISPs peering connection to reach Microsoft’s network was the best thing you could do. In 2017 and 2018 and 2019 that was true. But the world is operating a little different right now.

Gary Steere Speaking Ignite in 2017 on end user experience monitoring in office 365
Slide from session BRK3041 at Microsoft Ignite 2017 showing my quote. Microsoft is the creator of the content shown in this image.

At Perfrax we make end user experience monitoring software for Microsoft 365 and Microsoft Teams. In fact, I personally have written the logic that powers our back end artificial intelligence systems. Not to brag (and yet, I am going to :), but the combination of the data we gather and our software warned customers of issues with Microsoft Teams nearly two hours before the well-publicized Teams outage in February 2020 was formally announced.

As users started transitioning to home we started to see a large change inbound data. It caused such a change that our AI engine was reporting potential anomalies at far too high of a rate. As I looked into the data more closely, it became clear that the AI engine was not wrong. While we all expect performance to behave differently from home, I was surprised at the exact places where we saw changes. Local Egress is no longer guaranteed to be best thing you could do as I started in 2017.

This claim runs exactly opposite to Microsoft’s guidance. Including updated guidance Microsoft has released in the last few weeks. Microsoft clarifies how you should configure connectivity, specifically VPN clients. The articles generally promote local egress for most Office 365 / Microsoft 365 based traffic. I can tell you, using data, that the guidance is not correct as it is written. We can prove that local egress is not always the “gold standard” for connectivity.

I expect that my position that local egress and/or split tunnelling is not always best will bring a lot of questions. I’m going to review the material in a technical webinar on May the 4th. (Any Star Wars People out there?). There are several things we see differently in the world of work from home users and you may need to adjust your strategy. We’ll share our learnings and the data we have gathered so that you can optimize the experience of your users currently working from home.

When the world returns to normal it is quite possible that local egress will once again be king. Or the world may remain changed forever. Until such a time there are technical details you need to understand about how Microsoft 365 / Office 365 is working for your work-from-home users right. Join us on May the 4th, 2020.

Register for our webinar on May 4, 2020

Leave a Reply

Your email address will not be published. Required fields are marked *