yoshitaka272 - Fotolia
How to buy hyper-converged infrastructure
Simplicity, high availability, scalability and reduced cost are among the benefits of hyper-converged infrastructure. We survey the benefits, pitfalls and the key suppliers
Businesses always need to simplify operations, but many companies struggle to configure and manage their virtual environments efficiently. And legacy infrastructure – with separate storage, storage networks, and servers – is not well suited to meeting the needs of enterprise and cloud native applications, or to the needs of businesses grappling with digital transformation.
As digitalisation takes increasing prominence in CIOs’ budgets – up to 28% by 2018 – the demands placed on IT infrastructure to deliver reliable, reactive services will increase, and hyper-converged infrastructure will be a key tool which facilitates this changing business environment.
Hyper-converged infrastructure offers a consumer-inspired user experience, making the management of virtual machines simple for IT staff.
In essence, hyper-converged infrastructure (HCI) is much like a datacentre in a box managed by smart software and addresses the challenges of traditional legacy infrastructure – inherent complexity, inefficient silos, forklift scaling and painful management.
Hyper-converged infrastructure integrates servers, storage and virtualisation along with end-to-end systems management and operations management capabilities to allow enterprises to deploy infrastructure in minutes and shift the focus to applications that power the business.
There are many choices that face the buyer when it comes to HCI in terms of supplier, size of deployment, hypervisor choice, and ability to scale.
Buyers must also consider issues such as supplier lock-in and software update lag.
Hyper-converged: Benefits, pitfalls, key features
What makes hyper-converged appealing is that it can reduce deployment times to near zero, increase resource efficiency, reduce OPEX, completely remove the possibility of problems related to interoperability, and most importantly, it comes supplier-certified.
As hyper-converged systems contain an entire infrastructure within one chassis, it takes up a lot less floor space than hardware that consists of traditional hardware. If space is at a premium, this can be a very attractive prospect for an organisation. There are also usually far fewer cables and a single power output.
By having all elements (compute, storage, networking, security, and availability) all virtualised via a software-defined layer and managed through a single integrated console can benefit IT organisations looking to simplify deployment and management.
This shared core approach allows suppliers to develop and sell at a lower entry point than with traditional architectures. In a traditional architecture, each node would need to be individually managed.
More on hyper-converged
- Hyper-converged infrastructure can be an easy win when it comes to datacentre infrastructure upgrades, but there are pitfalls. We discuss how best you can sidestep them.
- Can compute, storage and networks in one box be a better option than shared storage? And which organisations can most benefit from it, and for what scale of deployment?
Hyper-converged infrastructure helps to tackle increasing IT costs. Price to entry is very low because you can often start scaling at just two nodes.
Due to their configuration, hyper-converged systems allow you to build, scale and protect IT infrastructure more affordably and effectively than any other option available today. And software-defined intelligence reduces operational management, to provide automated provisioning of compute and storage capacity for dynamic workloads.
Another benefit is scalability. Hardware is usually bought as demand increases within an organisation, with price and specification primarily influencing purchasing decisions. This often results in an IT department that has to manage multiple pieces of hardware from different suppliers, with different support contracts.
That is not to say HCI is without its problems. Troubleshooting can be complicated; each system is tightly layered amalgamation of compute, storage, and networking, so it can be problematic to find the cause of bottlenecks. This could be caused by network, storage, too many services, lack of internal storage, etc. When the cluster size is increased, so does complexity.
When deploying HCI, separate teams for server, storage and networking will need to be pooled into one whole group, possibly causing organisational problems. However, with the right training, communication and reassurance, staff can transfer their existing skills to work in a hyper-converged environment.
Hyper-converged products overview
HCI suppliers have begun to embrace hybrid cloud and multiple cloud environments. Suppliers that only offered appliances now also offer software-only and reference architecture solutions.
Nutanix: One of the pioneers in hyper-convergence. Offers its software on its own appliances, OEM platforms, and third-party servers. Its Nutanix Enterprise Cloud OS, launched in 2011, combines storage, hypervisor, security, software-defined networking and management. Organisations can buy Nutanix as software or as an appliance through resellers and server suppliers.
Nutanix has also launched new hardware and software aimed at edge computing and remote office/branch office applications. It has also launched a number of software products aimed at HCI, such as Calm for application and IT automation; Flow for software-defined networking, Beam for cost and security compliance; and Epoch for application visualisation and performance management.
HPE: SimpliVity was launched in May 2017 and is available as an appliance that integrates HPE servers with third-party server virtualisation software and data services. HPE also offers software-defined networking from the company’s purchase of Plexxi in June 2018.
HPE has expanded hypervisor support as well as products using HPE’s latest-generation servers. There are also offerings without hardware-based acceleration and ones with integrated networking.
Cisco: Cisco’s HCI product is Cisco HyperFlex, which was launched in April 2016. The appliance integrates compute, Network Fabric, Intersight cloud-based management, third-party hypervisors, and HX Data Platform data and storage management software. Cisco HyperFlex is available only on Cisco servers. It has no software-only, hardware-independent products.
Its offerings originally targeted critical data centre workloads in a VM environment. Since then, it has expanded to encompass more deployment and workload requirements.
Dell EMC/VMware: This company has two products: VxRail (introduced in 2016) and a rack-scale software-defined solution called VxRack SDDC. For multi-hypervisor deployments, it has VxRack FLEX and VxFlex Ready Nodes, based on VxFlex OS. It also boasts the XC Series appliances based on Nutanix.
The company’s product lines cover a lot of different use cases, including midsize businesses, global enterprises and service providers. It also covers cloud, SaaS and critical business applications.
NetApp: NetApp takes a slightly different approach to HCI by separating compute and storage in its architectures, based on its claims datacentres and applications require both at the same rate.
NetApp HCI is designed for workload consolidation and VMware private clouds and can scale to 40 storage nodes and 64 compute nodes.
VMware: VMware’s software can be deployed in over 500 certified x86 platforms such as Dell EMC VxRail, Hitachi Unified Compute Platform (UCP) and Lenovo ThinkAgile VX Series systems.
VMware Cloud Foundation (VCF) offers a broad HCI portfolio, with software-defined services for compute, storage, networking, security and cloud management.
Pivot3: Its Acuity HCI platform features QoS software, that supports multiple and mixed-application workloads. It focuses on enterprises with high-density and virtualised workloads such as databases, VDI and mission-critical applications at remote office or edge computing.
It also has NVMe-level high-performance products that expand its abilities for hybrid clouds and virtualised workload consolidation.
Scale Computing: Scale Computing offers a branded appliance to end users through OEMs. Managed service providers can access the product as a software offering. It combines its own KVM-based hypervisor with software-defined storage and networking, and multi-cluster management.
Scale’s products were found mainly in the datacentres of medium-sized companies, but it has now expanded to remote office locations and distributed enterprises.
Microsoft: As you would expect from Microsoft, its offerings are based around Windows. Core HCI functionality is provided by Windows Server 2016/2019 Datacenter Edition – which includes Microsoft’s Hyper-V server and Storage Spaces Direct software virtualisation and Microsoft System Center. Windows Admin Center is a web-based management console for Windows Server and HCI. There are a number of hardware platforms Microsoft has validated for its software.
The HCI software is aimed at cloud, edge, remote office and hybrid use cases. Companies can deploy a full stack of HCI without the need of another supplier’s software.