Comparing Azure & AWS Enterprise Transit Network Features
Enterprises have realized the fact that the most important pillar of the public cloud architecture is the transit. This is where you are connecting the spokes, hybrid cloud, branches, partners, remote users, multi-regions and multi-clouds. If you choose the right transit then a lot of your public cloud challenges will go away but If you focus on just one aspect of the transit for example performance and ignore everything else then it will be analogous to building a house and focusing on decoration and wall colors and ignoring the location of the land which is in the flood zone.
Performance should be one of the considerations but not the only one, a well-rounded transit architecture should be the focus including the performance.
Performance – This is important and you want to have your applications communicate with each other across regions and clouds without any bottleneck.
Scale-out Repeatable Architecture – A transit that can repeat itself within a region (for example if multiple business units wants to manage their own infrastructure and firewall rules), across regions and across clouds. One should not think twice that how am I going to architect a new cloud environment because often times growing in a different cloud is the business requirement not technical.
End-to-End Network Awareness – Enterprise environment is always dynamic in nature. IP address space or CIDRs gets updated or VPCs/VNets gets added/removed often. Transit should be smart enough to adopt these changes and must provide seamless experience to users and applications.
Simplified Service Chaining – Networking and security services like next generation firewalls (NGFW) should be inserted in the architecture without any complexity, performance or visibility challenges like IPSec, BGP, SNAT and throughput.
Operational Visibility and Troubleshooting – While Day1 is extremely critical for the strong foundation and at the same time you need to make sure architecture must be efficiently operationalize i.e. common control plane and normalized data plane across multi-cloud environment without sacrificing common troubleshooting tools with end-to-end visibility up to the application level.
Below is the high level comparison between Azure, AWS and Aviatrix transit (GCP and OCI don’t have any native transit solution).