Microsoft 365 is a global service that is consumed by many corporations and individuals. Most end users connecting to this service do not realize the great lengths Microsoft went through to ensure you can get to your data. In this blog post I will shed some light on Microsoft’s global architecture and show why monitoring is needed more than ever.

Microsoft has invested heavily in their cloud infrastructure to ensure that your data is readily available. Let us take a deeper look at how Exchange Online has evolved over the years. Exchange Online has many different frontend/backend systems that are crucial to how you connect to the service and consume data. All mailboxes in Microsoft 365 reside in a database availability group (DAG) and are replicated to datacenters in the same region that are geographically separated. Every mailbox has five copies spread out across multiple datacenters for maximum redundancy, Exchange Online Resiliency . This means that you could be accessing your data from any one of these datacenters so essentially, Microsoft brings you to your data.

Microsoft has continued to employ the same distributed Exchange Online architecture to all the services they offer, SharePoint Online, Teams, etc. using a similar frontend/backend configuration. Network reliability and low latency between Microsoft 365 clients and Microsoft 365 service front doors are two factors that determines the quality of the Microsoft 365 end user experience, Microsoft 365 Networking. What we have learned so far, is that Microsoft made great investments in their frontend/backend infrastructure, provided data resiliency, and brings you the end user to your data. This means that networking plays a huge part of in the quality of the end user experience.

Azure Performance


Even though users may share the same tenant and are accessing Microsoft 365 from the same location, their data could be physically located in different datacenters. This means that they are connecting to Microsoft’s network in different ways, using different network paths. Now with more and more people working from home, this adds more complexity for IT departments; end users are now using their own home networks to consume Microsoft 365 services. So, to truly understand the quality of service being delivered to end users, IT departments need to understand where their users are and where their data resides. Perfrax TrueDEM is the only solution on the market today that can provide this kind of in-depth visibility.

Most competitors rely on synthetic transactions only to monitor Microsoft 365. Even though synthetic M365 monitoring can provide some data on the service account’s mailbox, it relies on the use of APIs that an actual user does not use. The mailbox being used by the “robot” is a dedicated mailbox, meaning it is not tied to any real end user and this traffic could be going to another datacenter than the actual end users. This is also a waste of M365 licenses especially if you are performing multiple tests from more than one location.

Graphical user interface

Description automatically generated
Teams call history

Perfrax TrueDEM monitors the end users directly so we are the only solution that can show the actual network path they traversed, what Microsoft 365 frontend/backend servers they hit and call out any latency along the way including the ISP. For every Microsoft 365 outage so far this year, we alerted on the issue sometimes hours before Microsoft even announced it.

Network performance