Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
Converged and hyperconverged storage systems are changing the way enterprises outfit their data centers and deliver IT services. And both these technologies are part of data storage trends that are gaining momentum.
In its recent analysis of the worldwide converged systems market, including both the converged and hyperconverged product categories, IDC found that sales had climbed 9.1 percent in the last quarter of 2017, on a year-over-year basis, for a total of $3.6 billion. For the year, revenue reached $12.5 billion, a 9.4 percent improvement over 2016.
Top converged infrastructure vendors include Dell and HPE. Cisco and NetApp make the cut thanks to their joint FlexPod converged infrastructure product line.
The leading hyperconverged infrastructure vendors are Dell, Nutanix, HPE and Cisco, according to IDC.
What are converged and hyperconverged storage?
Before exploring how converged and hyperconverged systems fit into an enterprise's IT strategy, here's what CIOs and IT managers should know.
What is converged storage?
Converged storage combines storage, compute, networking and some management software essentials into a single, preconfigured system. Diverging from IT strategies that rely on traditional servers and networked storage, this hardware-based approach is increasingly taking hold in mainstream data centers.
One reason for the growing popularity of converged storage infrastructures is how it centralizes IT management. Converged systems are also relatively easy to provision and scale, compared to their siloed counterparts. Deployment times and costs are reduced and management is simplified.
What is hyperconverged storage?
Similar to converged storage, hyperconverged storage infrastructure also combine storage, compute and networking into a single system.
Based on system built on commodity x86 hardware, hyperconverged storage bundle storage, networking and compute into a highly virtualized system. The biggest difference is that a hyperconverged infrastructure, often shortened to HCI, takes a software-defined, scale-out approach to systems and workload management for more agile and flexible IT operations.
How converged and hyperconverged storage differ
On a fundamental level, both converged and hyperconverged storage systems bundle a number of key elements into a single system, reducing the management load.
Converged infrastructures take a hardware-based approach to combining storage, compute and networking for improved IT management. Those aspects are all included in a pre-built system available from a single vendor, typically configured for a specific application or workload.
Rather than sourcing a server, storage array and networking equipment from a variety of hardware providers, businesses only have to deal with one company for sales and support, simplifying the buying process.
Another benefit comes in the form of simplified cabling and an overall more efficient use of data center floor space. Since all the components are stuffed into a compact form factor, organizations can essentially pack more processing power and data storage capacity into less space versus operating separate server clusters and networked storage systems.
On the storage front, converged systems integrate software-defined storage controllers and disks, whether traditional hard disk drives (HDDs), solid-state drives (SSD) or a combination of the two, into a single system. Typically, these are separate components within storage area networks (SANs), each occupying no small amount of rack space.
Hyperconverged, on the other hand, uses a software-based or software-defined approach to decouple IT service delivery from its underlying hardware.
One of the key converged technologies, virtualization is used to abstract storage, compute and networking and present them as building blocks that provide data center operators with an economical and comparatively easy-to-manage way to run and protect IT workloads. Considered a key enabling technology for the software-defined data center (SDDC) and IT as a service delivery models, hyperconverged infrastructures allow businesses to manage their resources as elastic pools of capacity.
Hyperconvergence lends itself to mixed workload environments. In terms of storage, it enables scale-out architecture that can help businesses deal large-scale data growth. Additionally, hyperconverged allows for reliable, web-scale distributed file systems.
When does it make make sense to embark on the journey of hyperconvergence? Here are a few tips on when it makes sense to implement HCI.
|Similarities||Both combine storage, compute, networking and some management software essentials into a single preconfigured system, reducing management load|
Takes hardware-based approach to combining storage
Prebuilt system from a single vendor
|Takes a software-defined, |
|Workloads||Typically configured for specific application or workload||Good for mixed workload environments|
|Advantages||No hardware compatibility issues|
Plug and play
|Less operational expense|
|Cost||Both have steep upfront costs or capex, |
but HCI has a cost advantage from less IT staff support and maintenance
Converged and hyperconverged use cases
There are numerous use cases for converged and hyperconverged storage systems. Realize that, typically, hyperconverged can accomplish anything that a converged storage system can; think of hyperconverged as the advanced version of converged. So we'll start off by listing converged, but these first use cases also apply to hyperconverged.
* Critical business applications: Converged infrastructure is typically used to support critical business applications, databases and virtual desktop infrastructure (VDI) environments.
* Data center consolidation: Given its knack for condensing various IT systems into one, it is particularly suited for data center consolidation projects.
There is some overlap between converged and hyperconverged infrastructures. For example, HCI also excels at VDI and can also support critical, or "tier 1," applications.
* Hybrid Cloud/Big Data analytics: HCI excels in some other areas, like big data analytics and hybrid cloud computing. Moving workloads between on-premises systems and the public cloud is a relatively simple endeavor. Running and managing virtual servers and application containers is similarly easier in hyperconverged environments.
* Extending server functionality: HCI can also breathe new life into not-so-old servers and help enterprises get more bang for their already-spent buck.
There's no reason to ditch perfectly good servers and blades of a recent vintage (servers that are three-years old or so), even if their designs don't quite live up to the combined storage, compute, networking and software-defined vision promoted by HCI. By combining a software defined storage (SDS) platform with existing server hardware and adding a few HDDs or SSDs, enterprises can have a hyperconverged system running application workloads quickly and cost effectively.
* Getting more from data: When applied to secondary storage, hyperconvergence can also be a boon for organizations seeking to derive more value out of their non-production data. Benefits include more efficient and economical business continuity, courtesy of built-in data replication in most cases.
* Lower cost analytics: Hyperconverged secondary storage also makes it possible to run analytics on more data at a lower cost and easily provide developers with production replicas, allowing them to test their code with less management overhead.
* But watch out for legacy apps: There is one major downside to HCI. Older applications and legacy workloads that seem to linger at large, established enterprises may not run well in HCI environments.
Cost factor for converged and hyperconverged storage
Both converged and hyperconverged solutions from IT vendors can have high sticker prices, which translates into steep upfront costs or capital expenditures (CAPEX). Such is the penalty of outfitting a data center with OEM hardware, in which the provider is responsible for converging IT systems into pre-built and pre-configured solutions of their own design, although they are often a bargain compared to buying standalone components.
For the do-it-yourself types, HCI can be implemented using commodity, off-the-shelf hardware. It's an alluring proposition, if keeping initial hardware and software costs low is a concern. Some experts recommend going the plug-and-play appliance route if scaling quickly and avoiding setup hurdles are goals.
Of course, each business is unique and so is the calculus used to acquire and maintain its IT systems. Although both converged and hyperconverged infrastructures have the potential to slash IT costs, HCI has the cost advantage.
A well-implemented HCI setup provides centralized management over all of its capabilities, often under a single interface or pane of glass. This means that less IT staff is generally required to provide support and maintenance – which represents a major cost savings in the long run.
Converged and HCI: The market suggests the future
What does the future hold for converged and hyperconverged technology landscape? A look at the latest moves by leading IT vendors offers some clues. Although these market event are just a few of the constant updates, they do suggest larger trends. Namely, deep investment in converged and hyperconverged infrastructure.
HPE doubles down on HCI with Plexxi buy
HPE announced on May 2018 that it is snapping up software-defined network (SDN) specialist Plexxi for an undisclosed amount. Building on its 2017 acquisition of software-defined hyperconverged infrastructure provider SimpliVity, HPE said Plexxi's technology will enable the company to offer its customers a "hyperconverged offering that incorporates compute, storage and data fabric networking into a single solution, with a single management interface and support."
Dell bets on sheer performance
During its Dell Technologies World conference, Dell took the wraps off two new HCI solutions, Dell EMC VxRail Appliances and VxRack SDDC Systems. Workload acceleration options in the new VxRail Appliances include NVMe cache options, Nvidia GPU accelerators and 25GbE networking. The cloud-friendly VxRack SDDC product line gets configurations based on 14th generation Dell EMC PowerEdge servers and Nvidia-backed VDI capabilities.
Cisco and NetApp bank on speedy storage
Also looking to improve application performance for enterprises, the companies took the wraps off FlexPod SF last summer, adding scale-out storage services based on NetApp's SolidFire all-flash platform to the converged infrastructure solution.