Blobs are binary large object , traditionally we had to either store them either inside a filesystem or inside a database , but with REST protocol , we now have a new way to access bits that are encapsulated by some metadata. All we need is a http server to answer to the rest api calls and you can get access to your group of bits or BLOB ….so thats it Azure BLOB is microsofts offering of Object storage in the cloud . When you add a layer of filesystem on top of it , you get Azure Data Lake storage …because its a filesystem …you now can have hierarchies or folders with sub folders with some more sub folders and so on and so forth …..the Blob storage is flat names space …you define your storage account name , your container and then pour your blobs into that container .
as you can see from the picture above once you are in the container , you don’t have access to create a second container , all you have is access to upload a file into this container. So if you want to store your objects and organize it neatly and manage it , then ADLS is your answer.
With Blob or object storage , you are not limited by the filesystem limitations – like the inode table etc …you just specify the exact identifier and it gets it back to you …. much easier so a lot more scalable …with ADLS you get the best of both worlds …the ABFS driver makes the rest call to the underlying blob storage and fetches the data and get it you
you cannot access the blob objects that are in data lake through a REST api call to underlying blob layer and access it using the filesystem …this is just way too risky ….this is not allowed.
( this is my understanding , its over simplified …feel free to add more clarity )