Microsoft 365, Dynamics 365, and any other Microsoft SaaS services are hosted in multiple Microsoft datacenters and can be accessed from any geographic location. The Microsoft global network has Microsoft Edge point-of-presence (PoP) locations around the world where it can connect to an end user via their service providers.

Microsoft and partner service providers ensure that the traffic for the prefixes registered with a Peering Service connection enters and exits the nearest Microsoft Edge PoP locations on the Microsoft global network. Microsoft ensures that the networking traffic egressing from the prefixes registered with Peering Service connections takes the nearest Microsoft Edge PoP locations on the Microsoft global network.

Companies looking for internet-first access to the cloud or considering SD-WAN architecture or with high usage of Microsoft SaaS services need robust and high-performing internet connectivity. Customers can make that transition happen by using Peering Service. Microsoft and service providers have partnered to deliver reliable and performance-centric public connectivity to the Microsoft cloud. Some of the key customer features are listed here:

• Best public routing over the internet to Microsoft Azure Cloud Services for optimal performance and reliability.
• Ability to select the preferred service provider to connect to the Microsoft cloud.
• Traffic insights such as latency reporting and prefix monitoring.
• Optimum network hops (AS hops) from the Microsoft cloud.
• Route analytics and statistics: Events for (BGP) route anomalies (leak or hijack detection) and suboptimal routing.

Peering Service uses two types of redundancy:

• Local redundancy:
  • Microsoft and service providers interconnect across multiple Microsoft Edge PoP locations to deliver Peering Service. In each location, interconnection must support failover across two routers.
  • Each peering location is provisioned with redundant and diverse peering links.

• Geo-redundancy:
  • Microsoft has interconnected with service providers at multiple metro locations so that if one of the Edge nodes has degraded performance, the traffic routes to and from Microsoft via alternate sites. Microsoft routes traffic in its global network by using SDN-based routing policies for optimal performance.
  • This type of redundancy uses the shortest routing path by always choosing the nearest Microsoft Edge PoP to the end user and ensures that the customer is one network hop (AS hops) away from Microsoft.

The following routing technique is preferred:

• Cold-potato routing

  The software-defined cold-potato routing technique offers control over network traffic that originates from the Microsoft cloud. It ensures that traffic stays on the high-capacity, low-latency, and highly reliable Microsoft global network until it's as close to the destination as possible.

  Routing that doesn't use the cold-potato technique is referred to as hot-potato routing. With hot-potato routing, traffic that originates from the Microsoft cloud then goes over the internet.

Service monitoring is offered to analyze customer traffic and routing, and it provides the following capabilities:

• Internet BGP route anomalies detection

  This service is used to detect and alert for any route anomaly events like route hijacks to the customer prefixes.

• Customer latency:

  This service monitors the routing performance between the customer's location and Microsoft.

  Routing performance is measured by validating the round-trip time taken from the client to reach the Microsoft Edge PoP. Customers can view the latency reports for different geographic locations.

  Monitoring captures the events in case of any service degradation.

Routing happens only via a preferred path that's defined when the customer is registered with Peering Service.

Microsoft guarantees to route the traffic via preferred paths even if malicious activity is detected.

BGP route anomalies are reported in the Azure portal, if any.