Data protection is in a state of evolution, transitioning to a service that offers the business more than simply backup.
Leading the charge in this move are scale-out backup appliances – purpose-built hyper-converged (HCI) hardware solutions that perform all the tasks of data protection, and more.
But why use an HCI-style appliance-based solution over traditional data protection? What can these solutions provide in addition to backup and restore?
In recent years, mainstream IT has moved through converged and on to hyper-converged infrastructure. At each stage, the benefit to the user or customer has been to provide simplification and remove complexity when building IT solutions.
We can see parallels in data protection, as suppliers have delivered products to eliminate the headache of building, managing and scaling backup solutions by developing backup appliances.
Appliances provide the same level of simplicity that IT departments love about hyper-converged infrastructure. Typically, this means taking the effort out of designing a solution that will scale with customer needs and fully exploit the hardware on which it is deployed.
Where hyper-converged meets data protection, we see appliances that look and act a lot like hyper-converged infrastructure. This provides the capability to meet backup needs and to add more value than a traditional backup design can offer.
One aspect of hyper-converged that evolves the technology from simply packaged hardware is the integration of storage natively into the platform.
The three main hyper-converged backup appliance suppliers (discussed below) all integrate a scale-out file system within their solutions.
As HCI scales primary storage and compute capacity by adding new nodes, scale-out backup works in the same way by providing more storage capacity and throughput to a backup cluster.
Having the ability to easily scale backup is a strong operational benefit for IT organisations. Typically, backup would be reviewed and scaled only when capacity was constrained, because the overhead of extending a backup solution was time consuming and arduous.
With the scale-out appliance model, as primary capacity is increased backup capacity can be added in lockstep just by racking a new node, powering it on and connecting to the existing backup cluster. The new capacity is then instantly available for use.
As hardware ages, removing nodes from a cluster is also easy to achieve, removing the headache of either recycling offline media or moving data from deduplication appliances.
The HCI model provides greater clarity around backup license charges. Suppliers typically license per node, making it easy to quantify the costs of expanding data protection capacity.
If chargeback is in use a simplified backup cost structure makes it easier to pass these costs onto the business – the cost of backup can be aligned and built into the cost of primary storage.
All data protection software solutions deliver added value by providing access to metadata that can be used to search backup archives. With scale-out solutions, the ability to place this metadata in one place becomes much easier and provides a single search endpoint for compliance or e-discovery.
As a storage platform in their own right, HCI backup clusters can be used for instant restores, as secondary storage for test/development environments or as the platform for object and file storage (in the case of Cohesity).
HCI backup appliance limitations
Since we last wrote about HCI backup appliances, the limitations of solutions are slowly being eroded. Suppliers are increasing their support for a broader range of applications and platforms, including bare-metal servers, eventually consistent databases and public cloud (including SaaS offerings).
One area that still hasn’t been addressed is the ability to ingest historical backups from other platforms. Many IT organisations choose to simply allow older backup environments to scale down over time, but many have data that can’t be discarded.
The first supplier to offer the mass import of backups from traditional backup media and products will make their solution much more attractive for large-scale enterprise customers.
Divergent data models
As more data is created and retained at the edge and in public cloud, so backup solutions need to cater for a divergent data model.
All of the vendors covered here offer either software-based and/or public cloud instances of their platforms. In doing this, they still need to offer a common interface for search.
As a result, we’ve seen SaaS platforms emerge, such as Rubrik’s Polaris and Cohesity’s Helios that place a management umbrella over distributed or multiple appliance clusters.
HCI backup vendor roundup
Cohesity has extended its C2000 series appliances with the introduction of the C3000 platform. The C3500 (currently the only model available) is a high-density solution, with 176TB of hard disk capacity, 7.68TB of flash and double the memory (128GB) of the C2000 series, all in a 2U server.
As Cohesity offers file and object storage in addition to backup, the C3500 meets the needs of customers looking to create archives or large-scale data repositories.
DataPlatform, the software driving the Cohesity solution, is also available as a cloud and virtual edition for public cloud and remote/branch offices respectively.
Commvault recently updated its range of HyperScale backup appliances that were originally introduced in 2017. New additions provide greater scalability at the high end and also meet the requirements of remote/branch offices.
The HS3300 offers from 174TB to 262TB of usable capacity using the same three-node architecture as the existing mid-range HS1300. The new Commvault Remote Office HS1100 appliance provides either 5TB or 15TB of usable capacity in a single 1U server, running Microsoft Windows 2016 Server. The hardware uses a combination of M.2 flash, SSDs and hard drives.
Read more about backup
- Applications that run in the cloud are protected, but only so much. For full protection of data generated by cloud-based apps, you need cloud-to-cloud backup.
- Backup methods pros and cons: Full, incremental, differential, synthetic full, incremental forever, and reverse incremental. We set out the advantages and drawbacks of each.
Commvault HyperScale is also available as a software solution that can be purchased to run on a wide range of server platforms, where customers want consistency in their hardware choices.
Rubrik has updated the previous 3000-series with the new 6000-series appliances. The changes are incremental between the two ranges of products, with slight increases in processor cores per CPU (8 to 10) and more memory on the high-end models. The 6000-series now benefits from support for 25GbE networking for greater throughput. Node storage capacities remain the same.
Customers can also deploy the Rubrik software on self-sourced HPE, Cisco and Dell hardware. The Rubrik platform is also available as a virtual edge appliance and as a virtual instance in public cloud.