Stay informed Subscribe to our Blog
Enterprises are by definition complex, multi-operational businesses that serve customers on a large scale. The cloud has increasingly become the most effective way for enterprises to operate their most critical business applications. For enterprises that are finally ready to make the move to the cloud, creating an enterprise cloud migration plan is a process that can be daunting. In the initial stages, the personnel that will make up the enterprise migration team should be selected, and a full view of the current environment should be assessed.
When it comes to Azure migrations, tools such as the Migration Assessment Tool and ASR Deployment Planner will give users insight into which parts of their workload will benefit from a move to the cloud, allowing them to consider how the migration will impact overall performance and affect costs.
There are six main points to consider when planning your enterprise’s Azure migration:
1. The migration model that the application will use
2. Current state assessment and future state planning
3. How to accomplish the move
4. Storage configuration options
5. Security and availability solutions
6. Migrating and maintaining your Azure deployment
In this post we’ll take a deep dive into each of these points, and show you how NetApp Cloud Volumes ONTAP can also help your enterprise’s Azure migration.
1. Application Migration Models
Analysis has shown that optimal migration to the cloud can be nested under two model types: Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS). Each of these two models has two variations:
Redeploys your current setup on an IaaS without making changes. Also known as Lift and Shift.
Fits existing coding into the new cloud environment by modifying or extending the code currently in place
The PaaS options are as follows:
Existing code is run on top of the cloud
Discards the existing code in favor of re-architecting on top of the cloud
The model a company chooses to use should reflect its business type and goals.
The PaaS model, for instance, will benefit companies that work primarily online and with software development. This is because Azure PaaS offers services such as Azure App Services, which provide serverless architecture capabilities that allow R&D teams to focus on innovating for the company, rather than maintaining infrastructure.
The fastest method to accomplish migration is the IaaS rehosting, aka “Lift and Shift.” Rehosting uses the infrastructure components of Azure. Other options include a mixed IaaS-PaaS configuration, or completely replacing the service a company offers with a new one based in the cloud. Deciding between an Azure PaaS vs. IaaS services will largely be dependent on your existing application’s requirements.
Another Azure migration path is for an enterprise that is currently deploying in another cloud, such as AWS, to redeploy in Azure. There are a number of reasons why an enterprise would want to move from one cloud provider to another, including more attractive pricing, unique Azure features, better compatibility with existing applications, or to unify data silos. However, carrying out such a move can be difficult, as the platforms are not built to interact. Their native migration tools simply aren’t able to move data between one cloud, such as AWS or Google Cloud, to Azure. That’s where NetApp Cloud Volumes ONTAP can come in, making it possible for data to transition seamlessly between clouds.
2. Assessment and Planning the Azure Migration
The initial phase of any cloud migration is a current-state assessment of the on-premises environment to identify the connectivity of the different tiers of the application, their dependencies on on-premises resources if any, and legacy configurations that need to be updated before migration. This process helps in deciding the scope of the migration of the application, i.e. whether it will be possible to rehost and revise to IaaS or if it will be necessary to refactor and rebuild to use PaaS. The information gathered in the assessment phase goes directly into the planning phase, where deeper evaluations of the application will be done to fine-tune the migration approach.
In large-scale environments, manually assessing environments for cloud migration is not feasible. For this purpose, Microsoft offers multiple tools to help customers in this process. The basic tool that can be used is a MAP assessment toolkit that can generate an Azure VM readiness report by scanning the on-premises environment. It can provide precise information regarding the compatibility of the existing servers with the Azure platform. However, note that it does not take the dependency aspects on other services into account.
The Azure Migrate service helps to assess VMware or Hyper-V (preview) environments and report on their feasibility for migration. This tool also helps you to group machines related to the same application to visualize and assess the dependencies, something which helps in the planning process. The assessment report can be downloaded in an Excel format. After running an Azure Migrate assessment, you can leverage services like Azure Site Recovery or Azure Database Migration service for moving your application and data to Azure. The Azure Website Migration Assistant tool can help in automated assessment and migration of websites running on Microsoft Internet Information Services (Microsoft IIS).
3. Accomplishing the Migration to Azure
The objective of the move should result in a lower TCO and improved performance at lower costs. By setting your goals early on, you will be able to take advantage of Azure VM Sizing to make sure these parameters are met. Map toolkit and Azure Migrate also provide recommendations on the target Azure VM size based on their assessment results.
Azure runs on the proprietary Hyper-V platform from Microsoft, which makes lift-and-shift rehosting procedures onto the Azure blob storage from an on-premises Hyper-V environment as simple as transferring a VHD from the old environment to the new one.
Azure Site Recovery (ASR) enables customers to continuously replicate virtual or physical servers to Azure, and then migrate using a simple failover process. It also offers a test failover approach, which can be used to test functionality of applications before the actual Azure deployment. The popularity of running Linux on Azure virtual machines has also led Suse and Red Hat to offer their own services in Azure through BYOL or monthly fees.
4. Using the Appropriate Azure Storage Types
Regardless of whether a migration is based on an IaaS or PaaS model, storage costs will increase over time. To keep rising costs under control, it is crucial to get to know the different Azure storage types and to select the storage account type that appropriately meets the application’s performance, read/write operations, and replication requirements.
The right storage service will be determined by the data types in use: both IOPS and throughput will be factors to consider. Make certain you understand the IOPS nuances for different types of storage and test this with tools like IOMeter and SQLIO to simulate workloads.
Pro tips: For some virtual machine tiers, you can use storage spaces to combine the IOPS from several disks. Also, you get temporary SSD storage when creating a VM, which can be utilized for things such as Pagefile or SQL TempDB.
One way to tackle storage in Azure is to use the concept of managed disks, which overcomes a lot of limitations inherent in storage accounts
Moving cold storage data to the Azure Cool Blob storage format is another way to keep costs down. It is both cost effective and simpler to accomplish than alternatives. Since Cool Blob is meant for infrequent access, it is ideal to use for data that cannot be deleted but is accessed sporadically, such as backup data for example. For long-term retention of data archives for compliance or regulatory purposes, the Azure Archive storage tier can be used, which offers a low-cost, high-value cloud cold storage option.
With Azure’s StorSimple solution, for instance, you get to build on your prior expertise of on-premises storage solutions, but with a unified management console for both on-premises and cloud storage workloads. Another option to maintain compatibility with enterprise-grade products is to use NetApp® Cloud Volumes ONTAP, which not only provides the familiar Cloud Volumes ONTAP features of NetApp FAS products, but also automates data movement in the cloud and provides a central console for operational insights and resource cost monitoring.
Certain virtual machine tiers allow users to combine IOPS from more than one disk by taking advantage of storage space. When creating a virtual machine, users are granted temporary SSD storage that can be used for Pagefile or SQL TempDB.
Becoming familiar with the use of managed disks will also help deal with storage issues with Azure.
5. Azure Security and Availability Solutions
Public cloud usage comes with a major concern: security. With some loss of data sovereignty and infrastructure control, it is important to make sure to have a strategy for encrypting all at-rest data in place.
For Windows, Bitlocker Drive Encryption and Azure Key Vault can be an excellent starting point while devising this strategy.
Linux users can supplement Azure Key Vault with DM-Crypt. Another important technology that helps encrypt data at rest is Azure Storage Service Encryption (SSE), which encrypts the data as it is stored in Resource Manager Storage Accounts.
The Azure marketplace also hosts many third-party vendors who offer their own encryption services. A service such as BYOL can offer a low-cost solution to protecting expensive appliances.
Networking resources will not get lost in an infrastructural move to the cloud. Built-in technologies like network security groups and Azure WAF can control the flow of traffic with the aid of technologies like route tables and User Defined Routes.
These technologies can also be used to setup advanced use cases like forced tunneling.
Azure Virtual network gateway helps setup traditional VPN connections for use cases such as Site-to-Site VPN and Vnet-to-Vnet VPNs . Apart from this, always ensure that you have a holistic view of threat vectors through the Azure Security Center for Azure or even hybrid deployment scenarios.
It should go without saying that any security measures that are enacted should stay within RTO and RPO objectives. Depending on your environment and availability requirements, it will be wise to take advantage of Azure availability tools such as availability sets, and availability zones, ASR and Azure backup, and guest clustering.
All of these features have capabilities that will meet specific backup and data recovery goals. Availability sets use fault domains and update domains to guarantee uptime while performing planned or unplanned maintenance. Availability zones can be used to protect your applications from data center failures. VMs in availability zones are deployed in multiple data centers with independent power, cooling, and networking to protect against data center level availability issues.
Guest clustering uses Storage Spaces Direct for high-availability and load-balancing features of VM workloads. For more on security solutions and Azure resilience, refer to this guide.
6. Migration to Azure and Beyond
After spending considerable time on assessment, planning, and testing out your new environment, you are finally ready to move your workload to Azure. Now you need to think about migrating huge amounts of data to the Azure data centers.
What if your data is too big to migrate through these conventional tools?
Consider using import/export, with which you can ship hard drives with the on-premises data to Azure data centers.
In the case of hybrid deployments, use of exclusive networks such as ExpressRoute can significantly reduce transfer time. Azure Site Recovery, already mentioned, is also a good feature to help you move VM-based and physical server workloads to the cloud.
Microsoft provides out-of-the-box support of hybrid cloud for enterprises. With products such as Windows Azure Stack, for example, Microsoft has enabled enterprises to deploy a true hybrid cloud. Windows Azure Stack brings Azure Resource Manager, blob storage, and fault domains (among other features) to your on-premises data center, transforming it to a true, scalable private cloud solution.
And don’t forget to make use of the SQL Database Migration Tool to move SQL databases from on the premises to the cloud. A tool that can help you in this process is Data Migration Assistant (DMA), which detects compatibility issues and allows you to move data to an Azure SQL database. Similarly, the Azure Website Migration Assistant tool can help in automated assessment and migration of websites running on IIS.
For enterprises that are currently using NetApp storage systems, moving to Azure with SnapMirror® data replication technology is a cost-effective and fast way to carry out the migration. SnapMirror and Cloud Volumes ONTAP for Azure can also continue to help you manage a deployment, especially if your enterprise is deploying a hybrid deployment, keeping data synced in all the repositories automatically and at low cost.
Want to get started? Try out Cloud Volumes ONTAP today with a 30-day free trial.