March 15, 2023
Hybrid and multicloud deployment are the new standard for enterprise infrastructure. In these deployments, data mobility is key. For NetApp users, SnapMirror® can be used to replicate data from on-prem to the cloud, across availability zones within a public cloud region, between different public clouds, or globally between multicloud storage environments.
Multicloud deployment can also be a DR solution, allowing you to immediately bring up your software stack in another public cloud, as a copy of your data already resides there. Or this can be done using specialized services available only from one public cloud without migrating applications to that cloud.
This blog will describe NetApp's SnapMirror data replication technology, explain the concept of multicloud deployment, provide an architectural overview, and discuss example use cases.
Use the links below to jump down to the sections on:
- Hybrid and Multicloud Deployments Use Cases
- Hybrid and Multicloud Deployments Challenges
- How NetApp SnapMirror Can Help
- Moving Data Between Clouds: A Customer Case Study
Hybrid and Multicloud Deployments Use Cases
There are a number of use cases where companies might want to deploy a secondary copy in a different cloud environment, including:
- Testing cost-effectiveness: Data can be moved to a different cloud to see how costs compare for competing services. For example, if a cloud user reads about cost-effective ARM processors in another cloud provider, that user can test applications using a secondary data volume in the other cloud to compare performance per investment in both environments.
- Test a new cloud service: Data can be moved to test out a new service in a different cloud. Consider the use case of a cloud user investigating AI image recognition, a technology where another cloud provider is at the forefront.
- Burst to the cloud: Cloud bursting helps on-prem environments achieve cloud-like scale for short periods of time. In this scenario, a customer has on-premises data but is becoming limited by on-prem CPU processing capacity and wants to burst into the cloud during peak times. Data replication is key to that.
- Services in different clouds: One of the major benefits to multicloud deployment is choosing the best services each cloud provides. In this scenario, the output of service A on provider X is written to the primary volume of asynchronous SnapMirror relationship, with secondary volume in provider Y, where the data is consumed by a preferred service there.
- Promotion opportunity: Cloud providers may approach companies to consume more of their services. For example, a customer is using a multicloud environment with a secondary Cloud Volumes ONTAP hosted in AWS. The customer is approached by AWS with an offer of free credits to try its compute. Since the customer already has a copy of the data in that cloud, it will be easy to take advantage of the offer.
Hybrid and Multicloud Deployments Challenges
Cloud providers generally offer similar services that serve the same use cases, but the different management systems use separate terms and provisioning processes with varying cost models.
As an example, data protection policies and offerings vary across cloud storage vendors and may not match what you use in your data center which can lead to backup policies in different environments.
No cloud storage providers offer native tools or methods for synchronizing data across different clouds, so administrators often rely on third-party tools. Unfortunately, these present additional complexities and cost, with another management console to learn.
Security services and encryption are vendor-specific, and storage administrators need to understand each vendor and how they differ for this critical part of storage management.
Automation tools are bespoke to the cloud provider. Both vendor and 3rd party solutions rely on the provider's API. The result is a more expansive DevOps code base and more complexity.
You can find further explanation of these difficulties in this post on the challenges of multicloud deployment.
How NetApp SnapMirror Can Help
SnapMirror is a mature and proven feature for ONTAP that creates a replica of a source volume on a destination ONTAP system. This technology has been used for years on-prem to copy data between ONTAP systems. Any two ONTAP systems which can communicate across a network or internetwork can form a SnapMirror relationship. This functionality has greatly advanced with cloud-based deployment.
Today, SnapMirror functionality has extended to the cloud, allowing users to back up and move data to and from Cloud Volumes ONTAP, between hybrid architectures using ONTAP and Cloud Volumes ONTAP, and between Cloud Volumes ONTAP systems either within a single vendor's cloud or across multiple clouds. SnapMirror functionality with Cloud Volumes ONTAP is orchestrated by NetApp Cloud Manager.
During the initiation of a SnapMirror relationship, an entire block-level copy of the source volume is created on the destination ONTAP system, including any volume snapshots existing on the source volume. The destination volume is read-only but kept synchronized with the source volume, either synchronously or asynchronously.
For an asynchronous SnapMirror, at a defined interval, NetApp Snapshot™ technology is used to copy all the data blocks in the source volume to the destination volume which have changed in that interval. At this point, the destination volume and the source volume are identical.
With synchronous SnapMirror, the source ONTAP system immediately sends all changed blocks to the destination. Synchronicity is limited only by the network's throughput and performance of the destination storage.
Blocks from the source volume containing space-efficient compressed or deduplicated data are transferred to the destination just as they are. This preservation of storage efficiencies keeps the data footprint optimized, lowers transfer costs, and saves computing resources on the destination system, which otherwise would have to recompress and deduplicate the data again.
Once a SnapMirror initialization completes, a user could at any point choose to break the SnapMirror relationship. Then, the destination volume would become a normal writable volume, valuable in moving data between locations in a controlled manner.
Moving Data Between Clouds: A Customer Case Study
A NetApp customer, a global leader in materials engineering solutions for the semiconductors industry, was initially using ONTAP on-premises and started using Cloud Volumes ONTAP on Azure for DR.
Once the DR copy was established in Azure, the customer started to create more data copies and clones, and leverage them for other purposes:
- Backup: Using NetApp Cloud Backup, the company is backing up data from the DR copy directly to Azure Blob.
- Cloud migration testing: With the DR data in Azure, the company can now evaluate how existing on-premises applications will perform as they plan to migrate applications to Azure as well.
- SW testing: Test new applications and SW versions before launch.
- Performance testing: Test applications performance in different configurations.
To show an example of how this could work, we consider an on-prem NetApp appliance and a Cloud Volumes ONTAP instance in Azure Cloud called Cloud Volumes ONTAP2. The following diagram shows a minimal infrastructure to explain the concepts without the intricate network and security design.
NetApp Cloud Manager is the portal that manages Cloud Volumes ONTAP instances, Cloud Backup, on-premises ONTAP, SnapMirror relationships, data clones, and many other features.
- DR copy: An on-premises NetApp appliance called ONTAP1 contains several volumes used by three application stacks. The volume "Destination" holds critical data for business continuity, and as part of a DR plan, the volume is replicated with SnapMirror to an Azure Cloud Volumes ONTAP instance called CVO2. Any changes to blocks on the source volume are replicated to the destination volume at a set interval.
- Cloned copies for testing: The company uses FlexClone® data cloning technology to create clones of the destination volume in Azure. NetApp FlexClone creates new volumes, which are effectively zero-space copies of their source volume. In addition, these clone volumes are writable, and only changes made to them are stored, making them very space efficient.
The tests performed using the data clones evaluate existing applications in several circumstances: when running on Azure cloud (Clone 1), to test new applications or software updates using actual data before rolling out (Clone 2), and to test application performance with different compositions (Clone 3).
- Backup: Cloud Volumes ONTAP (CVO2) is protected using Cloud Backup. The data on CVO2 is backed up directly to an Azure Blob, which is far more cost-effective than block storage.
In this article, we have shown how hybrid and multicloud environments using NetApp Cloud Volumes ONTAP storage offer flexibility over and above what is achievable in any single domain.
The case study example scenarios show how a company can leverage a DR environment for a variety of use cases, including testing or processing, whether using read-only copies or writable FlexClone volumes—reducing overall required storage capacity and therefore costs, time to market, and storage management complexity.
Use Cloud Manager to spin up a Cloud Volumes ONTAP instance and take it for a test drive yourself—the choice of cloud is up to you.