Hyperconverged storage has become a key factor in enterprise data storage.
The trend toward hyperconverged storage is causing technology leaders to rethink their data storage strategies. It enables enterprises to support agile development cycles, streamline IT management and bring their storage, compute, virtualization and networking initiatives into tighter alignment. Its growing popularity also spells good news for vendors that have hopped on the bandwagon early.
Shifting to a hyperconverged infrastructure from a traditional server and SAN (storage area network) environment can have a profound impact on how an organization delivers and manages its IT services. But first, let's examine what makes a hyperconverged infrastructure (HCI) tick.
What is hyperconverged storage?
A step up from converged systems, hyperconverged storage combines storage, networking and compute into a highly virtualized system based on commodity x86 servers. Of course, IT organizations have a range of options that allow them to tailor these systems to their application and workload needs.
This approach simplifies management and enables organizations to create flexible pools of shared storage and compute that are comparatively easy to manage and scale. Need more capacity? Just add more nodes to the cluster.
Hyperconverged infrastructures are appealing to organizations looking for an alternative to SANs, but there is one factor they need to closely consider: the network.
Since all nodes in a hyperconverged setup store data that other nodes will likely need to access, this can lead to larger volumes of network traffic. Completing storage operations will require a robust network, catching up to the latest software-defined networking techniques and close collaboration between on organization's network and storage teams.
Who sells hyperconverged systems? Some of the usual suspects, it turns out. Dell, Nutanix, Cisco and HPE are the leading hyperconverged vendors, according to market research firm IDC.
Hyperconverged storage systems replace legacy systems
Now let's look at the impact hyperconvergence has on the data center.
In a hyperconverged environment, traditional servers and storage arrays and appliances are a thing of the past. Compute, storage and networking components are bundled into a single system or node based on commodity servers, reducing hardware costs, at least compared to specialized storage arrays, servers and other equipment.
Hyperconverged storage may seem like an all-or-nothing proposition, but IT buyers today have a plethora of configuration options that help ensure that they aren't stuck with a system that doesn't fit their requirements. For example, they can order a box with more storage capacity and a less powerful CPU subsystem if increased storage capacity is the aim.
On the software front, they include software-defined storage, hypervisors and virtual networking, allowing many of the operational elements to be abstracted from the underlying hardware. In terms of data protection, this is typically accomplished with erasure coding and replication, helping ensure that virtual machine data isn't lost as it is shuttled across a network, at the cost of some storage capacity, of course.
One benefit is simplified management and improved flexibility in how applications are deployed. Automation and orchestration tools further help IT teams optimize their HCI solutions and stamp out complexity.
Virtual desktop infrastructure (VDI) is an enduringly popular use case for hyperconverged solutions. Now, they are also used to provide file services and run a multitude of workloads, including production databases and critical business applications.
Software-defined data center (SDDC)
Hyperconvergence can also serve as a software-defined data center (SDDC) building block.
Software defined data centers feature virtualized infrastructures where storage, compute and networking are delivered as a service. The aim is to provide cloud-like flexibility, agility, scalability and seamless workload migration while lowering management overhead with the use of policy-based automation and orchestration.
See how closely hyperconvergence and software-defined data centers are aligned? By eliminating siloed approaches to storage and compute, hyperconverged systems can play a critical role in helping organizations realize their IT as a service vision.
Hyperconvergence and mainstream adoption
The market for hyperconverged systems is picking up.
In fourth quarter of 2017, sales of hyperconverged systems jumped nearly 70 percent to $1.25 billion year-over-year, according to IDC. For all of 2017, hyperconverged systems vendors raked in $3.7 billion, a 64.3-percent gain compared to 2016.
Clearly, more data center operators are turning to hyperconverged storage as a means of modernizing their IT operations and service delivery models. And this increased appetite for the technology is boosting the HCI businesses of some major provides, sometimes by astronomical amounts.
Hyperconverged system sales surged by a whopping 340.4 percent at HPE during the fourth quarter of 2017, year-over-year. Cisco, too, experienced a healthy jump of nearly 200 percent.
Another sign that hyperconvergence is gaining mainstream attention comes from another major IT analyst firm.
In February 2018, Gartner published its first hyperconverged magic quadrant report. Familiar to CIOs and IT decision makers, the magic quadrant format groups vendors into four categories, leaders, challengers, niche players and visionaries. They are then ranked by their ability to execute and the completeness of their vision.
Needless to say, landing in the prestigious leaders quadrant is a momentous occasion for a vendor, one that is often highly publicized. Today, those leaders include Dell EMC, HPE, Nutanix and VMware.
Changing landscape of hyperconverged storage and its technology
Hyperconverged storage is a big evolution from traditional approaches. And there are more changes in store. Here are some trends that will affect hyperconverged IT environments in the years to come.
- Cloudier data centers
Although private and hybrid clouds are hardly new, enterprises are quickly discovering that legacy infrastructure and as a service IT delivery strategies don't mix, or at least not as well as they would hope. By their very nature, hyperconverged environments lend themselves to cloud-native workloads.
One trend affecting the hyperconvergence scene involves completely disaggregating the storage, compute and networking resources from another, allowing organizations to create massive, flexible pools of capacity for all applications.
- NVMe over Fabrics (NVMe-oF)
Although hyperconverged systems tightly integrate compute and storage resources, larger workloads may require accessing data from multiple nodes over the network, introducing latency. Vendors are exploring the use of NVMe over Fabrics specification for snappy applications response times.
- Hyperconvergence continues mainstream push
Spiceworks' State of IT survey for 2018 revealed that 23 percent of organizations had already deployed hyperconverged systems. Another 22 percent expected to adopt the technology in the next 12 months.
- IT staffing changes
The storage administrator's job is about to change. In a survey of Dell EMC HCI solutions customers, IDC discovered that it took 81 percent less time for IT workers to deploy storage. With less unplanned downtime (98 percent) and improved application performance (42 percent), technology professionals will have to adapt to a new way of doing things.