When one examines enterprise data protection and data storage products through the lens of hyper-converged infrastructure (HCI) designs, one would think each product either supports an HCI architecture or it does not. But as one begins to see when one scrutinizes this topic, the answer is not a simple “Yes” or “No”. Nuancing how well or if a product fits into an HCI design, one first needs to think about the question or even the series of questions that he or she should ask to properly make this assessment.
As DCIG has begun to re-position its analysis to look at the emerging and existing set of data storage and data protection products in the context of HCI, one can quickly see many of them introducing or coming to market with HCI architectures. Consider:
- Cohesity and Rubrik have come to market with data protection solutions built upon hyper-converged architectures
- Deduplication backup appliance provider ExaGrid has been using the phrase “hyper-converged” more frequently when referencing the architecture of its product
- Commvault recently announced ScaleProtect that, delivered along with Cisco’s HyperScale Software, provides an HCI data management and protection software solution
- This year Comtrade Software introduced its HYCU software that specifically targets the protection of Nutanix Acropolis Hypervisor (AHV) environments running on Nutanix HCI platforms
- Pivot3 acquired NexGen Storage in 2016 and has since incorporated NexGen’s Quality of Service technology into Pivot3’s HCI Acuity platform
- NetApp, a long-time enterprise storage player, in mid-2017 announced its own HCI platform aptly named HCI
Looking at these multiple announcements from long standing players in the enterprise data protection and storage spaces who are adopting hyper-converged solutions as well as how many of the new entrants into the market are delivering their solutions on HCI architectures, one would think checking the HCI check box would be straight forward.
In one respect, it is. Many of the emerging and existing data protection and data storage solutions can now run on HCI platforms. You will get no argument from me on that point.
However, the bigger question is, does a product running on its own HCI platform really solve the bigger management problem that enterprises should ultimately seek to address? Or do all these latest and greatest product iterations just create another new rat’s nest of management complexity that enterprises need to deal with?
When I look at what HCI platforms should deliver in the context of a perfect world, they should only have one solution that spans on-premises, off-premises, and/or in the cloud that gives them the flexibility to:
- Run applications and/or VMs where they are best suited to run
- Provides them with the analytics they need to optimize data placement, performance, and quickly trouble shoot issues
- Minimizes or eliminates down time associated with upgrades and patches
- Manages data holistically across the environment
Granted, no such environment existed in any organization for which I have ever worked in either the public or private sector. However, it would be awesome if such a solution existed that could deliver on this ideal. In short, one could put all these products together to create a cohesive, single architecture upon which they could use to build their underlying IT infrastructure for tomorrow’s data centers.
Here’s the problem that emerges with HCI solutions as they stand today. If one selects an HPE SimpliVity, Nutanix, VMware VxRail, Pivot3, Scale Computing, or any other HCI platform as their primary production HCI platform, it is unlikely – and I would even say improbable – that any of these new secondary HCI solutions will integrate into one of these primary HCI platforms.
Granted, these other solutions may provide primary or secondary storage for the primary HCI environment or protect the applications and/or data residing on them. However, can they plug into and/or be managed as part the broader HCI enterprise environment that enterprises want and need? Or do any of these other solutions once again just become another island that organizations must manage outside of their primary HCI platform?
To help mitigate these management pitfalls that can result from using multiple HCI-based solutions, alternative solutions such as those from Datrium have already emerged. Like other HCI platforms intended for use in primary, production environments, Datrium’s software offers:
- Disaster recovery workflows
- Data encryption
- Flexibility to run on inexpensive x86 servers
- High availability
- Snapshot and replication capabilities for VMs
- Support for leading hypervisors such as VMware vSphere and Linux KVM
Datrium also gives organizations the option to create two types of nodes: Compute and Data Nodes. Compute nodes provide the high levels of performance that production applications require while its Data Nodes provide cost-effective options to optimize storage capacity housed in server-size footprints.
The wave of adoption for hyper-converged platforms is just getting started and much yet needs to be sorted out in terms of how well these HCI platforms interact with one another. However, any organization that adopts a primary HCI platform and then expects any of its data storage and/or data protection products running on adjacent HCI platforms to seamlessly plug into their primary one may be in for a rude awakening. Data storage and data protection products running on secondary HCI platforms do result in individually making them easier to deploy, manage, and upgrade. However,unless each of the HCI solutions connect and run as one cohesive solution, they still create their own silos of data that organizations must manage apart from their primary HCI platform.
Note: This blog entry was updated on January 2, 2018.