The three top public cloud service providers are Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP)—with AWS clearly in the lead with 40% market share.
However, under the cloud-first leadership of Microsoft CEO Satya Nadella, Microsoft Azure is determinedly closing that gap. In October of 2017 Microsoft announced that its annualized revenues in 2017 from Microsoft Azure and other cloud-related services (such as the Office 365 workplace software) will be $20.4 billion — already achieving the cloud revenue milestone that the company set for itself in 2018.
When it comes to storage, Azure offers cost-effective options for unstructured data and object storage for both rarely-accessed (cold) and frequently-accessed (hot) data. Azure’s Blob Storage has three different tiers (Cool, Hot and Archive), which offer trade-offs between the cost of the storage vs. the data accessibility SLA.
But despite Azure’s native ability to reduce storage costs through its rate of access, there is still room for an additional service that can store that data efficiently. For more savings on Azure storage, Azure users have to turn to the powerful storage efficiencies available in NetApp’s Cloud Volumes ONTAP (formerly ONTAP Cloud). Cloud Volumes ONTAP extends NetApp’s industry-leading on-premises enterprise data management solutions to the cloud. This article discusses how Azure users can layer NetApp’s Cloud Volumes ONTAP on top of Azure-native compute and storage services to achieve hyper-efficient enterprise-grade storage efficiencies.
Unmatched Inline Data Storage Efficiencies
Cloud Volumes ONTAP can be run as an Azure compute instance using either standard or premium Azure storage, as shown below:
As public cloud storage costs have a tendency to increase rapidly, Cloud Volumes ONTAP for Azure carries out a series of inline data operations to help reduce the storage space and costs required for maintaining data in Azure.
Cloud Volumes ONTAP for Azure’s inline adaptive compression process efficiently determines whether the data block entering the storage controller can be compressed by 50% or more. If that is determined to be the case, then Cloud Volumes ONTAP compresses the data to minimize storage overhead.
Most datasets contain a certain amount of duplication at the file and/or block levels. After performing compression (where relevant), Cloud Volumes ONTAP for Azure deduplication detects if the block currently being written is identical to an existing block in the storage media. If so, it saves a pointer rather than writing the block again. Inline deduplication can lead to substantial storage savings in and of itself. It is also possible to run scheduled deduplication in the background across stored data at rest. Deduplication is extremely effective in specific types of data, where some parts of it may be repetitive, such as with images.
Both data that has been compressed and small uncompressed files are eligible for compaction: a process that fits two or more smaller chunks into a single 4KB physical block before sending the block to storage. Substantial storage space is saved when each physical block is used as close to capacity as possible.
These three inline processes—compression, deduplication, and compaction—are complementary and designed to work together. For example, the higher the compression rate, the higher the compaction rate. Where deduplication may be less beneficial, compression can step up to the plate and yield significant storage space savings.
Thin Provisioning: True On-Demand Capacity
Cloud Volumes ONTAP for Azure implements thin provisioning so that the storage capacity for an app or a DB instance is allocated dynamically from a single shared storage pool only when data is actually being written to a volume or LUN. In this way you save the cost of provisioning storage space that will be unoccupied most of the time. Furthermore, when data is deleted, free space is released back to the pool. Overall, Cloud Volumes ONTAP for Azure’s robust thin provisioning not only saves on storage costs, it also simplifies capacity planning.
Here’s a classic example of how Cloud Volumes ONTAP thin provisioning saves storage and reduces costs. One of NetApp’s Cloud Volumes ONTAP customers is a well-known university, which estimated that it would need close to 30TB of physical storage in order to meet its storage allocation obligations to its students and faculty members. In reality however, most of the students and faculty were using little or no storage. With ONTAP Cloud thin provisioning, the college was able to seamlessly meet its storage obligations with only 3.5TB of physical storage — and saved $90,000 a year in storage costs in the process.
Snapshots are read-only, point-in-time, virtual copies of file systems or volumes, used primarily for backup and recovery purposes. Most snapshot implementations use what’s called a “copy-on-write” (CoW) approach. The initial snapshot points to the original data, and data blocks are only copied when the original data is to be overwritten (or deleted) by a change. In this way, a true point-in-time version of the original is preserved. Although more space-efficient than snapshots that copy the entire volume every time a change is made, CoW snapshots still require reserved storage capacity to store the snapshots—typically 10-20% of the volume size.
NetApp’s proprietary Snapshot technology is at the heart of Cloud Volumes ONTAP for Azure’s suite of disaster-tolerant data protection solutions such as SnapMirror®, SnapRestore® and SnapVault®. NetApp was the first to introduce “redirect-on-write” (RoW) snapshots, leveraging its unique storage virtualization technology that maintains a set of pointers to individual blocks of data. When data changes, the updated data is written to a new block and changes the pointer. Thus, a NetApp snapshot is a point-in-time, read-only copy of the new pointers; it takes less than a second to create, incurs no performance overhead, and is highly space-efficient (only taking up several KBs). Each point-in-time Snapshot copy remains stable as changes are made to the original data, and up to 255 Snapshot copies can be stored per volume. However, each newly-created Snapshot copy points only to the blocks added or changed since the previous Snapshot. Cloud Volumes ONTAP Snapshots make it possible to keep an extremely high, up-to-the-minute SLA, while keeping costs low.
Near-Zero Capacity Volume Cloning
There are many cases in which it is necessary or desirable to clone a volume, such as creating forks for the application dev/test cycle, provisioning new virtual machines, or creating DR volumes. Cloud Volumes ONTAP for Azure leverages the Snapshot technology described above in its industry-leading FlexClone® solution:
Cloned volumes are created instantly and at near-zero capacity regardless of the source data size. Thus, FlexClone technology dramatically cuts the storage you need for dev/test or virtual environments by 50% or more. FlexClone also makes it easy to fully test your DR processes, which is essential to any robust DR plan. Last, but not least, you can deploy tens or even hundreds of virtual machines in minutes — with only a small incremental increase in your storage needs (and costs).
There are many other ways that Cloud Volumes ONTAP for Azure enhances Microsoft Azure cloud services. For example:
- The Snapshot and FlexClone technologies can significantly simplify the migration of workloads to the Azure cloud for current NetApp data management customers.
- NetApp’s cloud-integrated AltaVault storage lets an organization benefit from cloud economics while continuing to leverage existing backup infrastructure.
- The OnCommand Cloud Manager provides a centralized, intuitive data storage management interface across cloud-based and on-premises infrastructures.
We invite you to use our Azure Calculator to see how Azure and NetApp services work together to provide a cost-efficient data storage solution for your organization. And you are welcome to try a free 30-day trial of Cloud Volumes ONTAP for Azure.