Azure / AWS Comparison Thread

Azure / AWS Comparison Thread

Azure, although being 5 years stepping into the market, is currently following AWS's market share(%33) in second place with %19. Almost all the services even for niche fields like LLM training, Quantum computing or an endless number of different storage types at block, file, and object are also provided by Azure and due to the clever collaborations with Companies like Samsung, OpenAI, Bosch, it is probably not a bad idea to see what is happening on the Azure side, also thanks to IAC tools the complexity is lowered to be understandable without getting lost and possible to let the application run using a combination of services to reduce the cost immensely.

For non-techies, think it as if your physical post is delivered to you using a combination of DHL, Hermes, and UPS are passing your packets to deliver you faster. Well, for the most basic user, this is not important but soon we will grow strawberries on Mars, that's why, now you know.

high five: AWS, Azure, Google, IBM, Oracle. "Cloud Capital" by the definition of Yanis Varoufakis self.

So because of A, B and C. We are comparing AWS with Azure today. Let's start with what we have for each service.

Object StorageAzure BlobSimple Storage
Block StorageAzure DiskElastic Block Storage
File StorageAzure FilesElastic
Relational DatabaseSQL DatabaseRelational Database
Virtual NetworkAzure VNetVirtual Private Cloud
Content DeliveryAzure CDNCloud Front
Authentication/AuthorizationAzure Active DirectoryIAM
Key ManagementAzure Key VaultKMS
Network SecurityApplication GatewayAWS WAF
NoSQL DatabaseCosmos DBDynamo DB

Azure on the other side has no vCPU limitation thus enabling a range of options for load balancing and scaling easier than AWS.

Covering (with the help of Traffic Manager) non-Azure-end-points for load balancing at the application and network layer at Azure is another gem at Azure, while AWS can only support its endpoints.

One other aspect I noticed on the Azure side is the well-designed Load-Balancing capabilities. Inside the region, we have the classic load balancing while between regions you can easily set a Traffic Manager to do the routing between different end-points based on various criteria, such as weight, priority, weight, latency, performance, geographic, multivalue (which should be yet another topic).