Book Image

Learning Microsoft Azure Storage

By : Mohamed Waly
Book Image

Learning Microsoft Azure Storage

By: Mohamed Waly

Overview of this book

Microsoft Azure Storage is the bedrock of Microsoft's core storage solution offering in Azure. No matter what solution you are building for the cloud, you'll find a compelling use for Azure Storage. This book will help you get up-to-speed quickly on Microsoft Azure Storage by teaching you how to use the different storage services. You will be able to leverage secure design patterns based on real-world scenarios and develop a strong storage foundation for Azure virtual machines. You will start this book with an introduction to Microsoft Azure storage and how it can be used to build large-scale, real-world applications using Azure storage services such as blob, table, queue, and file. This book will also teach you about the different types of Azure Storage. You will then find out the best practices for designing your Azure VM storage, whether it is Windows based or Linux based and how to migrate your storage in different scenarios. You will learn also how to plan and implement your SQL database solution on Azure, and how to build your SQL database in different service models (IaaS/PaaS). Moreover, you will also learn best practices for dealing with Azure Backup, the role of Azure Site Recovery, and the solutions they offer in different scenarios. Finally, you will use StorSimple to unlock the maximum value from all of your data, and some tools that will help you manage your storage on-and off-premise (Azure), such as AzCopy and Storage Explorer. Monitoring your storage in no laughing matter, so you will learn to monitor, diagnose and troubleshoot it.
Table of Contents (9 chapters)

Azure Storage types

Azure Storage has many types and even subtypes of those types in order to satisfy Azure services' consumer needs and to fit most scenarios.

The most common types can be classified based on the following factors:

  • Durability (replication)
  • Performance (Standard versus Premium)
  • Persistency (persistent versus non-persistent)


One of the most buzzing questions about the cloud generally is:

What if, for some reason, the SAN/servers that store my data are completely damaged? How can I restore my data?

The answer is very simple because Microsoft Azure Storage is durable and supports data replication, therefore you can make sure your storage is highly available.

Replication ensures that your data is copied somewhere else, whether in the same data center, another data center, or even another region.

For more info about the SLA of Azure Storage, you can access it via the following link:

Replication types

Microsoft Azure supports multiple options for data replication. You can use whatever you feel suits your business, especially as every type has its own price.

In order to calculate your solution's cost, you can use the Azure Pricing Calculator, which can be reached via the following URL:

Locally redundant storage

Locally redundant storage (LRS) replicates three copies of your data within the same data center you have your data in. The write requests you do with your storage are not committed until they are replicated to all three copies, which means it replicates synchronously. Not only this, it also makes sure that these three copies exist in different update domains and fault domains. You can revise the terms guide at the beginning of the chapter to understand what the update domain and the fault domain are.


  • The least durable option, as it replicates only within the same data center
  • Your data will be lost if a catastrophic event, such as a volcanic eruption or flood, affects the data center


  • It is the cheapest type compared to the other types
  • It is the fastest type of data replication, offering the highest throughput since it replicates within the same data center, mitigating the risk of data loss that would occur during data replication caused by a failure having occurred on the original data host
  • It is the only available replication type that can be used with Premium Storage at the time of writing

Zone Redundant Storage

Zone Redundant Storage (ZRS) replicates three copies of data across two or three data centers within one of two regions asynchronously, plus the three copies of data stored within the same data center of the original source of the data.


  • This type can only be used for Block Blobs (one of the Azure services covered in the next chapter), and a Standard Storage account (general purpose Standard Storage accounts will be covered later in this chapter)
  • Does not support metrics or logging
  • Does not support conversion for other replication types, such as LRS, GRS, and vice versa
  • If a disaster occurs, some data might be lost, because the data replicates to the other data center asynchronously
  • If a disaster occurs, there will be some delay in accessing your data until Microsoft failover to the secondary zone

Advantage: It provides higher durability and availability for data than LRS, as it not only replicates in the same data center but also in other data centers.

Geo-redundant storage

Geo-redundant storage (GRS) replicates data not only within the same region but also to another region. Firstly, it replicates three copies of data within the same region synchronously, then it replicates another three copies of data to other regions asynchronously.


  • If a disaster occurs, some data might be lost, because the data replicates to the other regions asynchronously
  • If a disaster occurs, there will be some delay in accessing your data until Microsoft initiates failover to the secondary region


  • It provides the highest durability and availability, even if a disaster occurs in an entire region
  • Unlike ZRS, if the original source of data faces an outage, there will be no possibility of data loss if the other three copies that exist within the same region don't face an outage too, as it replicates synchronously within the same region.

Read-access geo-redundant storage

Read-access geo-redundant storage (RA-GRS) follows the same replication mechanism of GRS, in addition, to read access on your replicated data in the other regions.

Drawback: If a disaster occurs, some data might be lost, because the data replicates to the other region asynchronously.


  • It provides the highest durability and availability, even if a disaster occurs in a whole region
  • If a disaster occurs, you still only have read access to the storage, but no write access until Microsoft initiates failover to the secondary region
  • The region with the read access can be used for data retrieval by the nearest offices to it without the need to go to another region to access the data; as a result, the data latency will decrease
Regarding replication between different regions, it will not work with just any two regions; the regions must be paired.
For example the West Europe region can replicate with North Europe, and not any other region.
For more information about paired regions, check the following article:


As mentioned earlier, Azure provides services for all business types and needs. There are two types based on Azure Storage performance--Standard and Premium.

Standard Storage

Standard Storage is the most common type for all the VMs sizes available in Azure. The Standard Storage type stores its data on non-SSD disks. It is commonly used with workloads within which the latency is not critical. Plus, it is low cost and has support for all Azure Storage services (which will be covered in the next chapter). It is also available in all regions.

Premium Storage

Premium Storage is designed for low latency applications, such as SQL server, which needs intensive IOPs. Premium Storage is stored on SSD disks, that is why it costs more than Standard Storage. Microsoft recommends using Premium Storage for better performance and higher availability.

More details about Standard and Premium Storage will be covered throughout the book.


Another type of Azure Storage depends on data persistence, which means whether data will be there or not after stopping and starting the VM within which your data exists.

Persistent storage

Persistent storage means that the data will be there after stopping and restarting the VM within which your data exists.

Non-persistent storage

Non-persistent storage means that the data will be gone after restarting the VM within which your data exists.

Further details about storage persistency will be covered in Chapter 3, Azure Storage for VMs.