At its annual gathering of technology leaders and practitioners this week, Microsoft announced a new public performance dashboard for the Azure network, which is powered by ThousandEyes active monitoring. The dashboard serves as a provider-neutral, ongoing benchmark to guide enterprises in their cloud architecture and operations decisions. By providing a granular level of visibility into Azure cloud performance, Microsoft illustrates its commitment to delivering superior digital experience for enterprises that leverage the service.
How Does It Work?
The Azure performance dashboard displays monthly averages of inter-region latencies across Azure’s global network backbone based on ThousandEyes round-trip time (RTT) metrics collected between Azure region pairs every sixty seconds. The dashboard will also feature latency ranges (min/max) for intra-geographical region connectivity. This increased visibility into regional connectivity helps enterprises make more effective cloud architecture decisions.
Microsoft unveiled the dashboard at its annual Ignite conference. As part of Azure Networking announcements, Yousef Khalidi, Corporate Vice President of Azure Networking, cited a significant performance finding from the ThousandEyes 2018 Public Cloud Performance Benchmark Report, which showed that Azure demonstrated the most network stability when compared to other major cloud providers.
Khalidi’s update on the Azure network underscored Microsoft's architectural approach to exclusively use their global backbone for inter-region connectivity, which may contribute to its increased performance predictability. The preference to leverage its extensive backbone differs from some of the other major public cloud providers, which rely more heavily on Internet connectivity—another finding discussed at length in the 2018 Public Cloud Performance Benchmark Report.
Why Does This Matter?
When enterprises move to the cloud, they are often unprepared for the loss of visibility into the external network infrastructure and services they rely on to support their business critical and customer facing applications.
Network latency tolerance varies greatly depending on application and inter-region use case. For example, disaster recovery may be more or less sensitive depending on whether state must be maintained between workloads in different cloud regions. Network latency is also a significant factor in inter-service communications when building distributed, microservice-based applications. Understanding network performance between cloud regions using this dashboard enables enterprises to make better cloud architecture decisions, as well as understand and manage against performance baselines.
A Step Forward for Cloud Transparency
While Microsoft takes a step forward in providing transparency into its network performance, another cloud provider, Google Cloud Platform (GCP), recently took steps to limit visibility into its network. Google’s action to change the way it handles TTL, effectively pulls the covers over its network, inhibiting the ability of cloud and network operations teams to fully monitor and troubleshoot connectivity to their GCP-hosted workloads.
There's no doubt that enterprises running revenue-impacting workloads need and expect their providers to communicate quickly and fully on issues that impact their business. Having transparency into their cloud providers’ network performance is critical to ensure these high-stakes investments work as anticipated.