WavebreakMediaMicro - Fotolia

How the big five storage array makers tier data to the cloud

We look at the big five storage array makers' efforts to connect on-premise hardware with cloud storage and find automated tiering, on-ramps, and backup and archive capability

In recent articles we have looked at the extent of cloud storage products and services available. These have included the file, block and object storage available from the main cloud providers, and virtual storage appliances available in the cloud from the big storage array makers.

In this article we take a snapshot of integration between on-premise storage arrays and the cloud.

Methods used tend to break down into three main categories.

First, there are features and functionality that offer actual tiering to the cloud with various degrees of automation, mostly aimed at migrating inactive data off to cheaper storage.

Second, there are products and features that offer some form of backup and archiving to the cloud, through software or hardware appliances.

Finally, some suppilers – notably IBM and Hitachi Vantara – focus their cloud tiering efforts around a product that provides some kind of on-ramp to the cloud, as a facilitator of hybrid- or multi-cloud storage.

Dell EMC

Dell EMC’s midrange/enterprise Unity storage arrays offer file and block tiering to the cloud, “seamlessly” according to its publicity materials, using its Cloud Tiering Appliance (CTA).

This sits between the Unity on-prem deployment and the cloud. Files are migrated to the cloud according to user-defined policies and an 8kb stub left on the on-prem hardware. For block storage, snapshots are taken and these can be migrated to the cloud while the originals are erased. The snapshots can be restored to the original system or any other.

Cloud tiering from CTA is supported for Microsoft Azure, Amazon S3, and IBM Cloud Object Storage as well as Dell EMC’s Virtustream and Dell EMC Elastic Cloud Storage.

Read more about cloud storage

Dell EMC also offers CloudArray, which is a cloud tiering tool available as hardware or a software virtual appliance. CloudArray – gobbled up from TwinStrata in 2014 – can work with any SAN or NAS on-prem hardware, and can tier data to public cloud. It also offers snapshots, data deduplication and encryption functionality.

In addition, Unity arrays can be managed from the cloud and also come with Cloud IQ, a free cloud-based software-as-a-service suite with predictive analytics, alerts and remediation suggestions. Cloud IQ is supported in Unity, SC, XtremIO, VMAX, and PowerMax storage hardware.

Dell EMC’s scale-out NAS product, Isilon, has CloudPools, which allows policy-based automated tiering of data to the three key cloud providers as well as to private clouds.

Xtremio all-flash systems can tier data off to Dell EMC’s Virtustream, as can Unity and VMAX. No such option is available for PowerMax NVMe-equipped arrays but then if data is on those it’s not likely to be cold anyway. Dell EMC doesn’t seem to provide any cloud tiering for its SC series storage arrays. 

HPE

HPE’s StoreOnce data protection appliances have a feature called HPE Cloud Bank Storage. This offers use of the cloud as a target for backup and archiving, with change block tracking and data deduplication.

Cloud Bank Storage works with AWS and Microsoft Azure as well as private clouds built with Scality (see below) and can restore to any – presumably HPE – system in case of recovery from a disaster.

HPE 3Par publicity refers to use of Cloud Bank Storage as a “cloud tier” but it’s pretty clear this is backup/DR capability rather than a storage tier as such.

With HPE’s acquisition of Nimble Storage, it gained that company’s Cloud Volumes offering. This sees customers able to set up and provision Nimble flash-driven cloud storage instances in the Azure and AWS clouds. HPE calls it a tier, but there doesn’t seem to be any automated tiering functionality between on-premises deployments and Cloud Volumes.

HPE’s Scalable Object Storage – based on Scality’s RING architecture – presumably comes with the Zenko multi-cloud controller, announced in March.

IBM

IBM’s link between on-premises storage and the public cloud is IBM Spectrum Virtualize for Public Cloud.

This is the public cloud-capable update of IBM’s venerable SAN Volume Controller, formerly a hardware storage virtualisation box, but now runnable as a software appliance in the cloud and on-premise.

IBM Spectrum Virtualize for Public Cloud allows access to the public cloud – only IBM’s own, for now – from IBM storage to move data between on-premises datacentres and the cloud, to use the cloud for disaster recovery, devops and to provide asynchronous and synchronous remote replication.

Hitachi Vantara

Hitachi Vantara’s VSP all-flash F and hybrid flash G series arrays offer automated tiering via the company’s own Hitachi Content Platform to the Amazon, Microsoft Azure and IBM clouds. The focus is on reduction of storage costs by movement of inactive data to the cloud.

Hitachi Content Platform is based on an object storage platform, which can run in hardware and software versions and operate as private cloud storage with access to Azure, Amazon and Google clouds. With access from existing storage infrastructure it acts as an on-ramp to public cloud storage.

NetApp

For NetApp’s FAS all-flash and hybrid flash hardware it offers FabricPool. This allows tiering of inactive data off to public cloud storage, with Amazon S3 and Microsoft Azure Blob Storage supported, as well as private clouds. Tiering is automated and policies for data movement can be set on per-volume basis.

NetApp’s E-Series all-flash arrays can use NetApp SANtricity Cloud Connector for block-based backup, copy, and restore of E-Series volumes to an Amazon S3 account, with RESTful API job management of backup and restore tasks.

NetApp doesn’t appear to provide any connection to the cloud for its Solidfire all-flash storage. But then that may be because Solidfire is targeted as storage for those that want to provide cloud storage. 

Read more on Cloud applications