First and foremost, I am a big fan of Amazon as a company. I can't say enough good things about the online shopping experience Amazon provides. However, when it comes to cloud computing Amazon is in need of a serious competitive advantage; the early movers' advantage can only carry you so far. What can Amazon AWS do to remain relevant in cloud computing?
Some of Amazon's services are hosted versions of commercially available products from its own competitors, such as SQL Server from Microsoft, Oracle (yes... from Oracle), Windows on EC2... and so forth. Some of them are offered as services "on top" of Amazon's AWS by other vendors, such as MongoDB Atlas (although Atlas is also available on top of Microsoft Azure). And some services are Amazon's own, such as DynamoDB (https://aws.amazon.com/dynamodb/). And of course, Amazon needs to build and maintain a large array of infrastructure services to support its own existence, such as S3 storage and DNS services.
However, after removing the underlying infrastructure services, and the hosted products from its own competitors, Amazon has just a few "core" data products left of its own. Sure, if you look at Amazon's list of Solutions (https://aws.amazon.com/products/) the list is rather impressive, covering databases, storage, servers, data analysis, mobile, machine learning, IoT, serverless infrastructure and more. And obviously Amazon can boast many top tier customers such as Zillow and Nordstrom. However Microsoft Azure also provides a large array of services that can largely map to the Amazon service offering. So what will the ultimate selection criteria be when choosing a cloud vendor? How will organizations choose between Amazon AWS and Microsoft Azure? And what about other cloud vendors that came late into the picture, such as Google and Oracle?
The answer is simple: data.
When it comes down to it, it is all about data. Yes... data is king since it is the primary asset that everything else depends on. And the core data services that Amazon offers will largely dictate its future in the long run; that's why Amazon is investing in its no-sql data offering, DynamoDB. Considering that no-sql is the newly accepted data storage engine for new applications, it makes sense for Amazon to promote no-sql, and limit its investment in relational databases. Microsoft knows this to be true as well, which is why it created a brand new no-sql service from the ground up: Cosmos DB (https://docs.microsoft.com/en-us/azure/cosmos-db/introduction) (previously DocumentDB).
Cosmos DB and DynamoDB compete directly with each other; and because no-sql is the future of cloud application development I believe that the ability to provide the best no-sql service will ultimately dictate which cloud provider will be the market leader for the foreseeable future. Although DynamoDB was published about 2 years before Cosmos DB, DynamoDB is starting to show serious limitations that seem to indicate it is falling behind the leaders in no-sql. And you don't have to take my word for it; a quick search on DynamoDB limitations will quickly bring this fact to the surface.
I seriously do hope that Amazon finds it mojo again in cloud computing, as it is the only serious contender to the Microsoft Azure cloud at this point in time. If it wasn't thanks to Amazon, we would perhaps never have had the Microsoft Azure platform available. To keep things interesting, and provide consumers the best services at the best possible price point, healthy competition is required.
But all hope is not lost. Amazon can remain relevant by addressing its no-sql offering. Even better... considering that MongoDB is the leading no-sql platform today, Amazon could simply acquire MongoDB in order to remain a worthy competitor to Microsoft Azure and keep the playing field even.
In my opinion, without a drastic move on its core data platforms, Amazon will slowly lose ground.