My question is a bit naive. I use AWS for hosting my backend servers. I have lots of real-time data logs, which needs to used for Anomaly & Pattern Detection. Microsoft Azure ML offers ready-made suites for the same, but the Azure ML documentation mentions about having to upload the CSV files onto Azure Blob.
Is it possible to call Azure Machine Learning services from my servers directly, without having my servers on Azure Cloud/Blob? I couldn't find about this on Azure ML's website. Thanks
Once you deploy your Azure ML model as web service, you should be able to call it from anywhere. It's a managed service, so you don't need your own servers on Azure to use it.
The documentation item you mentioned applies to batch scoring service. It reads in the input csv file from Azure blob on your storage account, scores it, and writes the result to another Azure blob on your storage account. So to use batch scoring you'd need an Azure storage account.