BlueXP Blog

Why Should You Migrate Databases to Azure?

Written by Jeff Whitaker, Cloud Data Services | Dec 5, 2019 5:30:00 AM

In adopting a cloud-first strategy, one of the first items on the list is: How do I get there? Application migration to Azure is a top priority for organizations, whether moving from the data center to the cloud, or migrating across clouds. Of all of an organizations’ applications, database workloads are often mission critical, and migrating databases to Azure successfully requires a well-defined plan. There are multiple factors to consider during the planning process, among them the need for always-on high availability, consistently high performance, data durability, security, and data migration capabilities for hybrid architectures.

Azure NetApp Files (ANF) meets all these needs. It is also quick and easy to deploy, making it ideal for demanding enterprise database workloads such as Oracle. In this blog, we’ll explore ANF’s capabilities to understand why it’s the best choice for deploying enterprise databases in Azure.

This is part of a series of articles about Azure files

Cloud Databases: What Are Enterprises Looking For?

Data forms the backbone of all line of business (LoB) applications, and database systems must be capable of processing data efficiently while also ensuring data security and integrity. But the ability to do so largely depends upon the database storage solution used. For enterprise databases, the underlying storage must offer specific capabilities and features supportive of large-scale production environments.

High Availability

Database services require a storage solution that’s always available, without interruptions. Disruption to data access can bring your entire business operation to a standstill, often resulting in lost SLAs and financial penalties. Ensuring highly available storage with multiple access paths to the underlying data is therefore critical. In order to avoid the added overhead that can come with designing and configuring storage level redundancy, organizations should opt for solutions where high availability is built in by design.

Performance

Database performance levels are directly correlated with the speed at which the data is accessed. However, performance requirements vary by use case. For example, database backups can reside in lower performing storage since the data is not accessed frequently, but production databases require faster storage for a better user experience. Cost optimization is, in this regard, top of mind, because higher performance storage can be costly. Having the flexibility to choose among multiple service levels for cloud database use cases is very valuable for enterprises.

Durability

Databases are often the backbone of mission-critical applications, and businesses cannot risk these environments going down or becoming corrupted. The data integrity of any committed database transaction must be maintained at all times. To that end, database storage must be capable of overcoming failure to ensure durability. Storage-level data redundancy used by databases plays a crucial role in establishing durability.

Rapid Deployment

Database deployment issues can often be traced back to complicated storage design as well as poor capacity planning and a lengthy provisioning process. In a traditional on-premises environment, deployment could take anywhere from a day (if the capacity is available) to months (for example, if additional storage boxes need to be procured to meet the demand). Faster time to market is a key success indicator, and these organizations would lose their competitive edge if production slowed. Quick deployment and automation are thus highly desirable features for database storage platforms.

Azure Migration: Manage Cloud Databases in Azure

Deploying solutions that require a database, such as data and log files residing in NFS file shares, poses a number of challenges for organizations. The challenge stems from the need for fast response times, which are difficult to achieve with cloud file resources. The sheer complexity and the effort involved with this type of cloud database deployment are major drawbacks.

While setting up and configuring a complicated NFS file share server can be achieved manually using Azure IaaS storage and compute components, the process is cumbersome and is not scalable. But:

Azure Netapp Files was created by Microsoft and NetApp to address the many challenges of deploying Linux file share-dependent workloads in Azure. Available as a first-party service in Azure, ANF was purpose-built to help customers to deploy and run Linux file share workloads. Here’s what makes ANF the best solution for your database workloads in Azure:

Meet varying database performance needs: Migrate your database workloads to Azure to achieve sub-millisecond latency at multiple throughput levels. The choice of service levels—Standard, Premium, and Ultra—offer throughput of 16 MiB/s, 64 Mib/s, and 128 Mib/s, respectively, giving customers the flexibility to right-size performance to be best suited for their databases.

Built-in high availability: ANF is highly available by design, powered in the backend by trusted NetApp technology and Azure infrastructure with network redundancy, transparent failovers, and advanced data protection.

Durable and stable: The service offers the “nine 9s” (99.9999999%) of durability desired for database workloads. Your data remains protected from potential issues—such as hardware errors and disk failures—that can negatively impact data durability.

Data and management plane security: Data stored in ANF is secured through a FIPS-140-2-compliant encryption mechanism for best-in-class data security at rest. The ANF mount points are presented in a secure Azure VNET—rather than a public IP—to enable secure access to the data. Export policies for Network ACL, AD integration, and Azure RBAC ensure management plane security.

Deploy in less than 30 seconds: ANF helps your database deployments jump the queue when it comes to storage provisioning. Because the service is directly integrated with the Azure portal, file volumes can be deployed and made available for your databases in a matter of seconds. ANF also supports all automation options available for cloud-native services like Rest API, PowerShell, or Azure CLI.

Deployment of High-Performance Database Workloads in Azure

Two absolute must-haves for any production database deployed in the cloud are high availability and consistent performance. ANF gives customers the flexibility and power to choose from multiple deployment architectures to achieve the performance benchmarks required by Azure database workloads.

Volume Layout for Databases

Customers are able to choose either a dedicated or shared volume layout for their Azure databases. In a shared volume layout, all files—database data files, log files, redo logs—are placed in the same volume. This layout is more suitable for dev/test environments, which are lenient when it comes to performance requirements but still require that the data be highly available. With production workloads, it’s best to go with a dedicated volume layout, in which the files are placed in independent volumes that offer optimal IOPS for workloads.

Reference Architecture: Single VM

Next, let’s explore a sample reference architecture for an Oracle deployment in Azure using a Single VM. Oracle data files and log files are placed on separate volumes. The volume hosting the data files can be configured to use the Ultra service level for the highest storage throughput, while the log files can be placed in the Premium service level for cost optimization. As this architecture uses single VM deployment, NetAppⓇ SnapshotTM technology can be used to increase the resilience of the solution by creating a point-in-time copy of data as backup copies, without hurting performance.

Image 1: Single VM reference architecture

Reference Architecture: Multiple VMs for High Availability

The image below shows an active-standby architecture for Oracle deployment in Azure using Azure NetApp Files. The data and log files are placed in different volumes and replicated to the secondary region. The VMs and networks are placed in different availability zones to enable high availability for compute. The Ultra service level is used for volume-hosting datafiles, while the Premium tier is used for log file volumes. This architecture ensures high availability and performance of the database system at both the compute and storage levels and is recommended for mission-critical production workloads.

Image 2: Multi VM architecture for high availability

Meet Your Cloud-First Mandate for Database Migration to Azure

Digital transformation is a necessity for all organizations that wish to stay ahead of the game. A cloud-first mandate to migrate databases to Azure as part of this transformation is no longer an elusive goal. Azure NetApp Files addresses both immediate and long-term storage requirements for your databases in Azure.

Simplify Your Databases in Azure

Simplify your database migration to Azure using Azure NetApp Files. Subscribe today to experience the benefits first-hand.