Over the past year or so, technologies such as hyper-converged infrastructure, disaggregated hyper-converged infrastructure and composable infrastructure have received considerable attention for their innovative capabilities. But at the same time, many enterprise hardware vendors continue to offer compelling converged infrastructure products, even though the technology has been largely forgotten.
Here, we examine the latest happenings from six established converged infrastructure vendors -- Dell EMC, Hewlett Packard Enterprise, Hitachi Vantara, IBM, NetApp and Pure Storage -- and address where it still makes the most sense to use CI technology.
What's the difference between converged and hyper-converged infrastructure?
Before discussing the various converged infrastructure systems, let's briefly examine the difference between converged infrastructure and hyper-converged infrastructure. Both technologies bring together compute, storage and network resources that the vendor of the infrastructure certifies for use with one another. But a hyper-converged system is designed to function only as a cohesive unit. In contrast, a converged infrastructure is made up of off-the-shelf components that can be used individually, if needed.
How is a converged infrastructure deployed?
There are two ways to deploy a converged infrastructure. One option is to purchase a prebuilt system where the vendor provides compute, storage and networking hardware, along with a hypervisor and management software. Everything comes preconfigured, which simplifies setup.
The other option is to build a converged infrastructure based on a reference architecture, which is essentially a list of hardware that the vendor that provides the architecture certifies for use in a CI configuration. Building a converged infrastructure system from a reference architecture is usually more complex than purchasing a prebuilt product. But the benefit of this approach is that it enables an organization to select the hardware that's best suited to its unique needs. In some cases, a reference architecture may even give an organization the opportunity to repurpose hardware it already owns.
Converged infrastructure benefits
There are numerous advantages to deploying a converged infrastructure. First, it reduces the risks commonly associated with hardware deployment. The vendor that provides the architecture certifies the components are compatible and configures the software stack.
Deploying a converged infrastructure can also simplify the process of getting technical support. Because the vendor that provides the infrastructure tests and validates it for compatibility, it lessens the odds that vendor finger-pointing will become an issue. If an organization purchases a prebuilt converged infrastructure product, it may even receive a single point of contact for technical support, although this is more common with hyper-converged systems.
Another benefit is that initial deployment and long-term scalability may be greatly simplified. If an organization orders a CI deployment as a package -- as opposed to using a reference architecture -- the package will include everything it needs to get the system up and running. The organization shouldn't encounter any surprises such as needing extra software licenses or an obscure cable that isn't included in the package. The converged infrastructure package should include all the required hardware and software that an organization can comfortably deploy in an afternoon.
When it's time to scale the deployment, an organization only needs to add storage or nodes. Similarly, it may be possible to replace aging components with newer components, which is an especially compelling benefit, since most hyper-converged systems don't allow individual components to be upgraded.
Key features of converged infrastructure products
Converged infrastructure systems are designed to be turnkey and made up of standardized components. But there's more to converged infrastructure than just the hardware. One of the most important features is the management and monitoring software, which has been specifically built for use with the converged system. This software typically automates various deployment and maintenance tasks and monitors system components for problems. In some cases, the management software can automatically remediate certain types of problems. In most cases, converged systems also include a hypervisor from VMware, Nutanix or Microsoft.
Converged infrastructure use cases
The most common use case for converged infrastructure is server virtualization. However, an organization can use this technology for much more than just a virtualization platform. In fact, many organizations use converged infrastructure for hosting mixed workloads. Typically, this involves running containers and VMs side by side, but it's also possible to bring into the mix various databases and business applications such as enterprise resource management, financial software and HR software.
Organizations also commonly use converged infrastructure as a platform for hosting virtual desktops. Converged infrastructure platforms are also becoming popular for use with data analytics and machine learning workloads.
Converged infrastructure vendors
Each of the following six converged infrastructure vendors approaches converged infrastructure in its own unique way.
Dell EMC VxBlock 1000 is built for use in private cloud or hybrid cloud environments based on VMware. The system uses Dell EMC storage and data protection technologies, as well as Cisco LAN and SAN networking. VxBlock also uses Cisco USB blade and rack servers.
VxBlock supports mismatched hardware. For example, VxBlock can use Dell EMC Unity, Unity XT, XtremIO, PowerMax and Isilon storage. Similarly, it can use Cisco Unified Computing System (UCS) B-Series or C-Series servers.
Hewlett Packard Enterprise
The HPE ConvergedSystem 750 is built around the company's BladeSystem, ProLiant rack servers and networking platforms, along with Nimble Storage all-flash and hybrid arrays.
HPE also offers predictive analytics through its InfoSight engine. The vendor claims it can automatically resolve 86% of all issues without IT intervention, virtually eliminating level 1 and 2 support issues. This system, which is designed for use as a virtualization or container platform, has full stack knowledge up to the virtualization layer, and uses that insight to troubleshoot issues that might occur.
Hitachi Unified Compute Platform CI Series architecture enables organizations to configure compute, network and storage resources independently. The platform is designed so organizations can start small and scale up or out as needed.
The centerpiece of Hitachi's converged infrastructure offering is an AI-based system for managing complex tasks. This system, which Hitachi calls the Unified Compute Platform Advisor, assists with hardware management and is designed to reduce the complexity of managing multiple linked VMware vCenter environments.
VersaStack is built for hybrid cloud use as well as private cloud and multi-cloud. It was designed to help organizations manage their digital transformation and data center modernization efforts.
The most unique thing about VersaStack is its NVMe-based storage. IBM uses real-time compression to reduce the data storage footprint. IBM also uses consumption-based pricing, so organizations pay only for the storage they use, with no large, upfront storage investment.
NetApp FlexPod Converged Infrastructure is a highly versatile platform, but the company pitches it as a platform for powering AI and machine learning workloads.
Organizations can build FlexPod Datacenter for AI on Cisco UCS C-Series rack servers or Cisco UCS C480 ML M5 rack servers and Cisco blade servers. FlexPods consist of Cisco UCS 6000 Series Fabric Interconnects, Cisco Nexus 9000 series switches and NetApp All Flash FAS A-Series flash storage arrays. The storage can scale to more than 20 PB in a single namespace, and features built-in deduplication and data compression. NetApp OnTap data management software manages the FlexPod platform.
Pure Storage FlashStack is built to support mission-critical applications, data analytics and DevOps environments. Although Pure Storage positions FlashStack as a system for private and hybrid cloud, the company offers converged infrastructure architectures that are purpose-built for business applications such as Microsoft Exchange, SAP and Oracle.
FlashStack is based around the use of Cisco UCS rack servers and Cisco UCS Director. The platform also includes integration with OpenStack and VMware vRealize. Pure Storage also offers a version that's specifically designed for machine learning and AI applications and uses Cisco UCS C480 ML rack servers.