10
votes

I have around 300 million text files, ranging from 50kb to 100kb that I need to read and serve on a web page.

Remarks:

On Azure Table, I would need to split some of the text files, to make sure it doesn't exceed the maximum of 64kb per column.

On Azure Blob, I would not have the problem of splitting, but would need to store a reference to them on the Azure SQL/Azure Table.

Okay, intro is done! Now, on my Azure Web App:

  1. Would be faster to read the text file from Azure Blob or from Azure Table?
  2. Would the Azure Table be able to process more requests?
2
I did some testing, and, to my surprise, reading the file from the Blob is faster than from Table. Also, I agree, for search, I plan on using Azure Search.Fernando Silva

2 Answers

6
votes

I agree with Andrei M. The best solution is blob storage with a table service or even a sql database to store reference.

Just one more thing, if you plan to search for content in those file, neither table service / blob will be good for that. If that's the case, I recommend you setup a VM with elastic search / solr and index those contents.

PS: there's also Azure Search, for that, you can give it a try too.

6
votes

Blob storage is better then Table storage if you want to work with files. I definitely would not want to deal with file chunks in a table. If you need to keep references to your blobs, how about combining two options and use Table storage for references? I think this would be the best in terms of performance :).