The impact of hyperconvergence is far more than the mere aggregation of processing, storage and data protection resources into standardised modular nodes.
The true power of hyperconverged platforms is that they allow you to get digital products to market faster while streamlining your expansion, provisioning and management tasks. In a hyperconverged environment, the deployed systems handle multiple tasks – compute, storage, network, data protection – which means your IT team needs fewer specialists devoted to each of those tasks. This frees up IT professionals to deliver other components of the business technology agenda.
Many of those adopting hyperconverged platforms seek the benefits of a traditional storage system without the need to add proprietary storage arrays and the expense of managing and optimising them. But IT leaders who have selected hyperconverged platforms must beware of defensive internal manoeuvring – storage and server engineers are used to making purchasing decisions independently, and with hyperconverged platforms, they’ll have to rely much more on the cloud and virtualisation teams.
If you’re taking the hyperconverged path, there are a few things you need to do differently. The first is to make purchasing a joint responsibility. Multiple teams, including server, storage, networking cloud operations and application teams, must evaluate and purchase solutions jointly.
Read more about hyperconverged IT infrastructure
- As hyper-converged infrastructure continues to gain momentum, suppliers have raced to build their own platforms, sometimes at the cost of quality.
- Scottish law firm MacRoberts replaces more than 80U of ageing, heterogeneous HPE infrastructure with 4U of hyper-converged servers and storage from Gridstore.
Mid-sized companies that are heavily virtualised and have modest storage requirements can satisfy nearly all their storage and processing needs with a hyperconverged platform. Large enterprises can cover their primary storage needs for virtual machines (VMs) with hyperconverged platforms, but will likely use scale-out network-attached storage (NAS), object storage and cloud storage services for unstructured data. In the long run, IT professionals should use external storage systems for high-performance and high-capacity archival and file storage use cases.
The second change needed is to accept supplier lock-in. While hyperconvergence is convenient, it also locks your firm into a proprietary platform, particularly at the software-defined storage level, which is at the core of the hyperconverged offerings.
Over time, IT professionals must migrate from integrating best-of-breed products at all levels of the stack to selecting the best underlying hardware and layering an integrated hyperconverged offering on top. This continues the trend towards successively higher levels of integration that began with early integrated and engineered systems.
Many of the early hyperconverged suppliers also offer a software-only version that can be mated with multiple hardware partners other than the one they use in their turnkey offering.
Next is to change storage provisioning roles, responsibilities and operations radically. IT teams with large infrastructures have already transitioned daily storage provisioning chores from storage to virtualisation administrators to cope with the stream of urgent requests from stakeholders.
Storage administrators have carved out large allocations of storage capacity to virtualisation administrators, who then use that capacity to meet daily provisioning demands. Venturing deeper into private and hybrid cloud architectures, with automated provisioning based around global catalogues, will require further substantial role changes.
Hyperconverged platforms will accelerate automation because they centralise and standardise processor, memory and storage resources, and also embody all storage control logic in the central system software, with no reliance on specialised storage processors or embedded software.
Forrester’s Business technographics global infrastructure survey, 2014 shows that hyperconverged systems are the clear preference for integrated management, with 91% of technology decision-makers expressing their desire for integrated storage and virtualisation management. This preference reflects IT professionals’ response to continual pressure to reduce
Finally, organisations must shift the broader IT team’s skills. For several years, Forrester has been hearing anecdotal evidence that early adopters of first-generation converged infrastructure, such as Cisco, HP and IBM, have already collapsed some of their IT silos. Forrester believes the adoption of hyperconverged systems will accelerate this trend because it further abstracts the underlying management complexity, particularly in the storage domain.
Hyperconvergence aids technology accessibility
Hyperconverged platforms include many leading-edge technical features, including scale-out storage, in-line data deduplication, cloning and cross-domain management integration – but hyperconvergence’s greatest asset isn’t any one of these technologies. For example, while IT teams today use scale-out storage in high-performance computing (HPC) farms, in webscale cloud architectures, and for challenging unstructured customer analytics implementations in verticals such as media and entertainment, hyperconverged platforms make scale-out storage accessible to server and virtualisation administrators without storage expertise.
Nutanix shipped its first hyperconverged system in 2011. Within 18 months, arch-rival SimpliVity was shipping products. And in 2013, VMware announced Virtual SAN V (VSAN), its federated file system for hyperconverged systems, followed by its Evo:Rail appliance to allow multiple hardware partners to ship integrated hyperconverged hardware/software systems.
Today, suppliers including Atlantis Computing, Maxta, Nexenta, Pivot3, Scale Computing and StorMagic provide software-only solutions, and formerly integrated systems are being disintegrated – Nutanix and SimpliVity have now enabled their software on multiple independent hardware platforms.
Nearly all the hyperconvergence innovators know how to position their hyperconverged offerings as efficient alternatives to dedicated server and storage systems. The startups have taken their expertise in building scalable and resilient storage systems and integrated their technologies with hypervisor platforms to create hyperconverged platforms that infrastructure generalists can deploy quickly and manage easily.
In early 2015, large storage players EMC and Hitachi Data Systems introduced new platforms to the market. IT leaders evaluating hyperconvergence platforms must keep in mind that teams that don’t want a completely integrated hyperconverged offering can instead select from software-only storage players such as Hedvig, Maxta and SpringPath, which have products to converge server and storage hardware resources.
These software-only storage products allow customers to convert direct-attached storage (DAS) servers into networked storage, similar to how VMware’s VSAN and HP’s StoreVirtual VSA functions. While this option does not provide the simplified packaging of current hyperconverged offerings, it does allow IT teams to avoid the supplier lock-in of physical hyperconverged platforms.
Established suppliers are scrambling to react to the influx of hyperconverged system suppliers, either by manufacturing the hyperconverged software themselves or attempting to engineer their own alternatives – or both.
Dell resells Nutanix on its servers and sells its own version of Evo:Rail, and Cisco has a meet-in-the-channel programme with SimpliVity. There are currently seven other Evo:Rail makers. Pure software players such as Atlantis and Maxta have multiple hardware partners, and the ecosystem of certified hardware platforms for the hyperconverged software supplier community continues to grow.
This is an extract of the Forrester report Vendor Landscape: Hyperconverged Platforms by analyst Richard Fichera.