BlueXP Blog

Cloud SAN Storage Deployment in Cloud and Multicloud Environments

Written by Aviv Degani, Cloud Solutions Architecture Manager, NetApp | Feb 12, 2019 12:54:02 PM

Enterprise applications, such as database systems, have a requirement for fast, low-level data access, and the ability to manipulate storage at the block level. Though all major cloud vendors provide block storage solutions, issues usually develop when the storage is used at scale. For example, directly allocating storage across a large organization, and for myriad different purposes, inevitably leads to over-allocation and wasted storage resources.

Ensuring that data is highly available across Availability Zones, setting up cross-region disaster recovery, and combining disks together for performance using RAID are all important considerations when using block-level storage. Cloud Volumes ONTAP provides a robust solution for all of these issues, and many others, by giving you the power to deploy cloud SAN storage using NetApp’s proven technology stack.

In this article, we will look in detail at how Cloud Volumes ONTAP helps users to manage iSCSI storage and the added value this provides for cloud and multicloud deployments.



SAN (Storage Area Network) in the Cloud

What is SAN storage?


SAN (Storage Area Network) storage is used to provide block-level storage to client hosts using a remote connection, which is typically iSCSI over ethernet, but in on-premises environments may also be over a fiber channel network. This decouples the compute layer from the storage layer, allowing each to be managed independently.

Accessing block-level storage from compute instances in the cloud involves connecting to remote disks over a network connection, which is the hallmark of all SAN storage solutions. Using this technique ensures that applications can always access as much storage as they need, while at the same time providing transparent data redundancy. Amazon EBS, for example, creates a redundant copy of disk data within an Availability Zone in order to remove single points of failure at the storage level.



SAN Challenges


Managing SAN storage devices effectively in every regard requires users to address a number of different concerns. For example, it is imperative for most organizations to protect business-critical data, without data loss, across more than one site. This can be a difficult problem to solve while data is being continuously written to block-level storage devices. Similarly, having the ability to perform data recovery after regional service outages, such as those caused by natural disasters, is critical to any disaster recovery strategy.

Performing regular SAN backups is another important aspect of data protection in enterprise environments. Most cloud SAN storage solutions provide a snapshot facility that will create a point-in-time copy of a SAN storage device. An important point to note, however, is that these snapshots must be performed with the collaboration of the host application using the storage, which could otherwise lead to data corruption within the snapshot.

As well as ensuring data protection, it is also important to be able to easily scale up I/O performance as and when that is required. One way to achieve this is by combining a number of cloud storage disks together into a RAID group, however, this requires configuration on the host compute instance itself. If it becomes necessary to move the RAID group to another host in the future, this configuration would need to be reapplied, and in a way that retains the data that has already been stored.

When your cloud SAN storage is up and running, and actively used by production applications and services, software development teams will regularly require access to test copies of live data for testing. For example, a DevOps engineer creating a CI/CD pipeline may wish to incorporate an up-to-date test copy of production databases, as this would more accurately represent the live environment. This can be very challenging when the production databases are large in size, which is very often the case.



SAN Storage Management with Cloud Volumes ONTAP


Cloud Volumes ONTAP is NetApp’s data storage platform for the cloud, available in both AWS and Azure, and enables users to manage cloud storage resources, such as Amazon EBS and Azure Disk, using the mature and comprehensive set of storage features developed by NetApp. Cloud Volumes ONTAP provides a robust, flexible, and high-performance solution for building a SAN in cloud computing environments.



Data Protection


Cloud Volumes ONTAP provides data protection in a variety of complementary ways. Firstly, NetApp snapshots give you the power to create instant and space efficient backups of a SAN storage device of any size. Using NetApp SnapCenter®, application-aware snapshots can be created by quiescing the I/O performed by the host application for the very short time it takes to perform the snapshot, after which normal operation is resumed. This ensures that the data captured by the snapshot is in a consistent state on disk, and guarantees that data corruption does not prevent a successful restore of the snapshot in the future. Cloud Volumes ONTAP can restore any snapshot back to the source volume instantly, or use the snapshot for cloud cloning a new, writable copy of the volume using NetApp FlexClone®.

For high availability, Cloud Volumes ONTAP HA can be used to deploy a pair of nodes across Availability Zones, providing non-disruptive access to client hosts and applications if one of the nodes goes down. Synchronous replication keeps the two nodes fully synchronized, which means that there is no data loss if a failover is required. This provides a much greater level of protection than having all of your data available within a single site only.

Data protection in the Disaster Recovery use case can be extended across regions by using NetApp SnapMirror®, which provides highly efficient block-level data replication between any two deployments of Cloud Volumes ONTAP. After performing an initial full copy of the source data to the destination, all subsequent data transfers are incremental, in that they only copy over the changes made to the source data since the most recent transfer. NetApp SnapMirror provides built-in features for storage failover and failback, and can also be used to transfer data from on-premises ONTAP systems to the cloud. This can very useful for on-boarding new cloud deployments with onsite data.



Performance and Scalability


Cloud Volumes ONTAP allows for multiple disks to be used together in a RAID configuration that is completely transparent to the clients using the storage. This helps to drive up I/O performance, and is achieved by configuring one or more aggregates made up of the number and type of disks required for your cloud workloads. Disks can be added to an aggregate at any time after it has been created, giving you the ability to scale up as storage requirements change.

Aggregates act as a pool of storage that is divided up into volumes, where each volume represents a specific client storage allocation. Cloud Volumes ONTAP snapshots are typically created at the volume level, and will automatically take into account the various underlying disks being used by their containing aggregate. Moving a volume between aggregates is completely non-disruptive.



Thin Provisioning and Storage Efficiencies


For multiple volumes to make the best use of the available storage space within an aggregate, Cloud Volumes ONTAP provides a feature called thin provisioning, which ensures that data storage is only allocated to a volume when it is actually needed. This allows an iSCSI disk to be presented to a compute host that has a different logical size, i.e. the size reported to the OS, than its physical size. Thin provisioned volumes help to increase utilization by eliminating the need to allocate storage up front.

Cloud Volumes ONTAP also provides a number of core features that help to reduce the overall amount of storage space consumed in a SAN environment. One of these features, data deduplication, reduces the storage needed for multiple identical copies of a data block, and data compression, which compresses each block. Cloud Volumes ONTAP can combine these features together and apply them transparently to all clients accessing the data.



Conclusion


Cloud vendors provide the SAN storage basics, however, there is no alternative to Cloud Volumes ONTAP for providing enterprise-grade data protection, storage efficiency, and scalability in cloud-based SAN environments. This is why NetApp customers like AdvancedMD migrated to Cloud Volumes ONTAP for managing the SAN storage used by their SQL databases. For your next cloud deployment, there is no need to worry too much about SAN vs NAS, as Cloud Volumes ONTAP provides a unified solution for both.