Storage accounts and virtual machines
1. Storage accounts and virtual machines
In this chapter, we'll set up the necessary resources for the ReadRadar project: a storage account and a virtual machine.2. ReadRadar workflow in Azure
Let's take a look again at the workflow for the ReadRadar app. We'll need a storage account to help with data ingestion: we'll need to store and process book data. We'll also need a virtual machine for the recommendation model. The VM will host and run the model, and provide a secure environment for the data science team to further test and develop the model.3. Storage accounts
Storage accounts are used to store data in the cloud securely. You can see them as a sort of virtual container. With the help of a storage account, we'll be able to manage and share the data for the ReadRadar app in Azure.4. Storage types in Azure
Let's now consider the possible storage options in Azure and which of these would be most suitable for the ReadRadar case. Blob storage is the most versatile and general-purpose solution, it can store vast amounts of unstructured data like images, videos, and text files.5. Storage types in Azure
File storage is the best solution if there is a need for file sharing and collaboration across platforms.6. Storage types in Azure
Table storage is specifically suited for structured data where the data needs to be queried.7. Storage types in Azure
Finally, Queue storage is a special type of storage used for cases where requests or tasks need to be managed in sequence, like queuing sales orders. For the ReadRadar case, we have a text file containing book data available, which we'll store in Azure using Blob storage, as it is best suited for cost-efficient storage of text data.8. Virtual machines
Virtual machines are like a digital version of a physical computer. In the case of ReadRadar, a VM will be used to host and run the recommendation model. The data science team would like to test and further develop the recommendation model, so we'll be using an individual VM in this phase of the ReadRadar app development. In a later phase, we can start using Azure Machine Learning instead of the VM for more extensive machine learning capabilities.9. Data Science virtual machine (DSVM)
To meet the needs of the data science team, we'll be using a specific version of a virtual machine: the data science virtual machine, or DSVM. This is a custom VM image in Azure, specifically for data science. It comes pre-configured with common data science tools, like PyTorch for machine learning.10. Keeping track of resources
In this case study, we're only creating and using a couple of resources. But Azure is typically used for multiple projects, each with their own set of resources. AppSculpt, the company of ReadRadar, also wants to move other projects to Azure, while keeping track of costs and performance per project. For this reason, naming conventions or the use of suffixes or prefixes is common practice when naming resources. Azure also has its own features to properly track resources: resource tags and resource groups. Resource groups can be used to manage resources collectively, for example defining and applying access controls for the group once instead of having to repeat the same process for each resource individually. Resource tags are metadata that can be applied manually or automatically. They can be used in other services like Cost Management or Monitor, to filter for resources belonging to a specific project or team.11. Let's practice!
Let's get to work creating the necessary resources for the ReadRadar project!Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.