Critical Capabilities for Hyperconverged Infrastructure - Gartner

Critical Capabilities for Hyperconverged Infrastructure

Published 7 February 2018 - ID G00333081 - 53 min read

---

Infrastructure and operations leaders should use the 11 critical capabilities and five use cases described here to assess HCI product features and functions. I&O leaders need to align project portfolios with the use cases, rationalize technology and vendor choices, and standardize across use cases.

Overview

Key Findings

• Individual requirements continue to drive enterprises to hyperconverged infrastructure for rehosting virtualization projects and building new HCI virtualized infrastructures.
• Infrastructure consolidation and business-critical use cases are catalysts for infrastructure refresh and cost optimization projects and deployments.
• Hyperconvergence delivers value in the defined use cases. Of particular interest to infrastructure and operations leaders are remote office/branch office edge deployments and virtual desktop infrastructure hosts.
• HCI enables organizations to deploy applications into private or public cloud environments, leveraging microservices, containerization and hypervisors.

Recommendations

I&O leaders building and sustaining infrastructure modernization programs should:

• Align their project portfolios to the outlined use cases and rationalize their technology choices and the number of vendors accordingly. Where feasible, they should standardize across use cases.
• Ensure that the features and functionality IT leaders need during a project's life cycle are on the roadmaps of their shortlisted HCI vendors. They also need to track related product and technology introductions on those roadmaps and the related partnerships.
• Use scaling, cost, broad application/stack support, security and manageability metrics as critical capabilities for infrastructure consolidation and business-critical use cases.
• Prioritize their top three HCI use cases to pilot for cloud, ROBO and VDI in their organizations. Based on experiences on this first HCI project, learn HCI's limitations and determine whether other use cases can be viable.

Strategic Planning Assumptions

By 2021, more than 60% of hyperconverged infrastructure (HCI) adoption will skew into virtual desktop infrastructure (VDI) and remote office/branch office (ROBO) workloads.

By 2021, software-defined networking (SDN) will be available in most HCI offerings; however, the end-user adoption rate will remain less than 10%.

What You Need to Know

This document was revised on 9 February 2018. For more information, see the  Corrections page. 

HCI is a category of scale-out, software integrated infrastructure that applies a modular approach to compute, network and storage on standard hardware, leveraging distributed, horizontal building blocks under unified management. HCI vendors can build their own appliances using common, off-the-shelf infrastructure (e.g., hardware, virtualization and OS), or they can engage with one or more system vendors that package the HCI vendor's software stack as an appliance. Alternatively, HCI vendors sell their software directly to end users or through resellers and system integrators (SIs) for use as part of a reference architecture or on an HCI-as-a-service basis on-premises or in a public cloud.

In prior years, Gartner evaluated four classes of integrated systems in single Magic Quadrant and Critical Capabilities research:

• Integrated infrastructure system (IIS)
• Integrated stack system (ISS)
• Hyperconverged integrated system (HCIS)
• Integrated reference architecture (IRA)

There are four important changes in this year's Magic Quadrant and Critical Capabilities research, compared with the Magic Quadrant and Critical Capabilities for Integrated Systems research, which were published in 2016. The 2018 Magic Quadrant and Critical Capabilities documents:

• Focus exclusively on vendors and offerings in the hyperconverged segment.
• Drop the "system" hardware requirement, which is part of the HCIS appliance model.
• Define the market segment as HCI, allowing for nonappliance, software-only/bring-your-own-hardware, and as-a-service cloud delivery models.
• Limit the evaluation of integrated storage and data management capabilities to those technologies for which vendors have primary development responsibility and ownership.

We've made these changes because Gartner clients routinely compare HCIS with HCI software supported on a broad set of certified reference or OEM hardware partners' systems. In addition, Gartner clients are increasing the emphasis on public and private cloud capabilities and as-a-service procurement options.

HCIS forecasts and vendor market share data (see  "Market Definitions and Methodology: Integrated Systems") will not correlate with this HCI Critical Capabilities and its companion Magic Quadrant document. This is because the HCIS forecast is based on a branded hardware appliance view of the market, regardless of the developer of the underlying software-enabled capabilities. Thus, if vendors with revenue and market share in the HCIS market partner for the core technology, we have excluded them from this Critical Capabilities report. Likewise, vendors and products appearing in this report may not appear on the HCIS market share report, because they don't offer a hardware appliance under their own brand. In addition, given the change in definition and focus for this year's Critical Capabilities research, readers should make no direct comparisons with the 2016 Critical Capabilities for Integrated Systems (see  "Charting the Shifting Trends in Hyperconvergence: From HCIS to HCI [and Back Again?]"). 

Use-case requirements are diverging into VDIs, as well as ROBO point projects, infrastructure consolidation, cloud and business-critical infrastructure addressing the bimodal requirements for IT and operational technology (OT). Already, most large enterprises rely on HCI models in their environments. At the same time, HCI has emerged as an alternative platform of choice for developers engaged in agile, Mode 2 infrastructure deployments, as well as rehosting virtualized workloads.

Client inquiries show that many small or midsize businesses (SMBs) question whether they chose the right technology a few years ago, as well as whether they should migrate to a different technology now. New HCI technologies are adding to that consideration, and vendors promote long lists of features and functions; however, critical capabilities identify the key differentiators for products that are weighted based on their importance in a variety of significant use cases. For the 12 vendors' products we are evaluating in the integrated system market, we've identified 11 critical capabilities and five unique use cases.

Analysis

Critical Capabilities Use-Case Graphics

Figure 1. Vendors' Product Scores for the Consolidated Use Case

Source: Gartner (February 2018)

Figure 2. Vendors' Product Scores for the Business-Critical Use Case

Source: Gartner (February 2018)

Figure 3. Vendors' Product Scores for the Cloud Use Case

Source: Gartner (February 2018)

Figure 4. Vendors' Product Scores for the ROBO Use Case

Source: Gartner (February 2018)

Figure 5. Vendors' Product Scores for the VDI Use Case

Source: Gartner (February 2018)

Vendors

Cisco

Cisco HyperFlex: First shipping in April 2016, Cisco's HyperFlex HCI solution is a recent market entrant; however, it has quickly established a strong competitive position against longer-standing HCI suppliers. HyperFlex leverages a log-structured file system and delivers high virtual machine (VM) density, a fault-tolerant architecture and consistent high performance, with an integrated networking fabric and independent scaling of compute and converged resources. HyperFlex is primarily used in data centers and larger ROBOs for VM and VDI consolidation and business-critical use cases. HyperFlex leverages the popular Unified Computing System (UCS) server platform integrated with an HCI software stack from Springpath, which has been recently acquired by Cisco. 

During the past year, HyperFlex has delivered Cisco Validated Solutions (CVS) software support for Oracle, Microsoft SQL, SAP Apps and HANA, Microsoft Exchange, and Splunk. HyperFlex lacks support for Hyper-V, Kernel-based Virtual Machine (KVM) and Xen hypervisors, and OpenStack, and does not offer stretch clusters and native VM image backup capabilities. HyperFlex supports mirroring and replication, but not multisite replication. HyperFlex has support for containers, but not the Docker Volume Plugin. In theory, the software stack can operate on any standard server, but the solution is delivered only on Cisco UCS as integrated hardware and software. HyperFlex is a good choice for Cisco customers looking for a hyperconverged infrastructure solution for VMware, business-critical use cases, and VDI modernization initiatives in enterprises and larger ROBO installations. Customers have praised the quality of Cisco support.

Cisco Notable Service Traits

• Hardware: Cisco delivers HyperFlex exclusively on Cisco UCS C-Series servers, but compute-only nodes, including UCS B-series blades, can be added to HyperFlex clusters. 
• Software-Defined Infrastructure (SDI): Cisco combines software-defined computing (SDC), using VMware ESXi; software-defined storage (SDS), using the Cisco HyperFlex HX Data Platform software; and SDN, using Cisco unified fabric. 
• Virtualization (Hypervisors): HyperFlex supports VMware ESXi, VMware vCenter, VAAI, VMware HA, VMotion, DRS and EVC; however, it lacks support for Hyper-V, KVM or Xen. 
• Virtualization (Containers): HyperFlex supports container management frameworks, such as Kubernetes and Docker running in a VMware environment; however, it lacks support for runtime engines on bare metal or Windows native containers. 
• Data Services: The HX Data Platform provides always-on in-line deduplication and compression for hybrid and all-flash system and uses a log-structured file system that enables instantaneous snapshots and clones and space-efficient compression. 
• Data Protection: The HX Data Platform supports VM-centric, space-efficient snapshots to enable local backup and single-cluster asynchronous remote copy for disaster recovery. HyperFlex does not support stretch clusters or one-to-many or many-to-one asynchronous replication. 
• Management: HyperFlex provides RESTful APIs for integration with UCS Director, VMware vCenter and other management tools. Cisco also offers HyperFlex (HX) Connect, a web interface built on HTML5, for role-based management of HyperFlex clusters. 
• Stack: HyperFlex offers a growing number of Cisco Validated Designs for SAP, Microsoft, and Oracle applications and offers in-memory database support for Oracle Database, Microsoft SQL, Exchange and SharePoint. 
• Scaling: HyperFlex supports scaling up to 32 nodes and offers independent scaling of compute-only and converged nodes to better balance compute resources with storage capacity. 
• Service: Cisco offers 24/7, worldwide Levels 1, 2 and 3 on-site and remote support for all components of the hyperconverged solution. Cisco also offers SmartCall Home (SCH), which enables remote monitoring to identify event-driven issues and provide regular health checks. 
• Security: Cisco's HyperFlex policy-based security ensures uniform, consistent, compliant and secure encryption management and deployment across HCI clusters. 

DataCore

DataCore Software: DataCore's Hyper-converged Virtual SAN offering was launched in 2014 and is based on the mature 10th generation DataCore SANsymphony SDS product to address hyperconverged infrastructure. DataCore Hyper-converged Virtual SAN runs directly on hosts to aggregate their individual direct-attached storage (DAS), both spinning disks and flash storage, into a virtual storage pool. DataCore Hyper-converged Virtual SAN can also leverage existing SANs and cloud storage for additional capacity through its storage virtualization services. DataCore Hyper-converged Virtual SAN accomplishes high availability in the cluster by synchronously mirroring every data block between two fully active cluster nodes, which can be in a local or a metro-cluster configuration. 

Hyper-converged Virtual SAN leverages RAM as high-speed cache and features automated storage tiering with up to 15 tiers, based on access patterns. This autotiering capability is not limited to locally attached storage resources, but extends to all storage resources in the pool (local DAS, remote SANs or the public cloud). DataCore Hyper-converged Virtual SAN uses parallel input/output (I/O) technology, which accelerates I/O processing in servers to support more-efficient scheduling by using the power of multicore CPU technology.

DataCore's two-node Parallel Server holds the SPC-1 storage performance benchmarking record on 1.2 million IOPS with 0.22 millisecond latency. DataCore is a good choice for end users that need an HCI solution for distributed data centers and ROBO, addressing the needs of latency-sensitive applications, such as VDI and databases.

For years, DataCore's SANsymphony has been deployed by OEMs, system builders and SIs on all major x86 platforms; however, it does not offer a turnkey HCI appliance product. Deploying DataCore Hyper-converged Virtual SAN will require the assistance of an integration partner and separate support channels, which combine the DataCore software with standard x86 servers to deliver an HCI appliance.

DataCore Notable Service Traits

• Hardware: DataCore Hyper-converged Virtual SAN is designed as a software platform that runs on industry-standard x86 servers and provides reference architectures. 
• SDI: DataCore provides universal standardized SDI services across heterogeneous legacy storage infrastructure and HCI. 
• Virtualization (Hypervisors): DataCore Hyper-converged Virtual SAN mostly focuses on vSphere deployments, while supporting KVM and Microsoft Hyper-V. 
• Virtualization (Containers): No native integration outside PowerShell, and REST APIs support scripting into the container management interface. 
• Data Services: DataCore supports compression, deduplication, single-instance store, space-efficient snapshots and autotiering. 
• Data Protection: DataCore enables synchronous mirroring, asynchronous replication, snapshots, and continuous data protection. DataCore Hyper-converged Virtual SAN can also take advantage of SANs and cloud storage for data protection. 
• Management: DataCore Management Console provides common management platform for Hyper-converged Virtual SAN and integration with REST APIs. 
• Stack: DataCore Hyper-converged Virtual SAN supports the OpenStack API, is certified with SAP HANA and integrates with Citrix, VMware and Microsoft-based VDI systems. 
• Scaling: DataCore clusters can scale from two nodes to 64 nodes in one-node increments. DataCore is able to scale storage capacity independently by leveraging external SAN storage. 
• Service and Support: DataCore provides support only for its own Hyper-converged Virtual SAN software. 
• Security: DataCore Hyper-converged Virtual SAN is encapsulated in the security mechanisms of Windows 2012 R2 and 2016 OS. DataCore leverages self-encrypting drive technology. 

Dell EMC

Dell EMC offers a diverse set of HCI offerings, based on ScaleIO and VMware, as well as through an OEM partnership with Nutanix. The VxRail product is the flagship offering and is based on the vSAN SDS product. VxRail is available as five different series of appliances — each targeted at a different set of use cases. Customers have the option of choosing nodes based on Dell or Quanta servers. Dell EMC and VMware jointly develop VxRail appliances with co-engineering effort on the VxRail management software, as well as integrating the data protection and cloud gateway products from the Dell portfolio with VxRail.

Organizations looking for application consistency groups and synchronous replication can use RecoverPoint appliances for 1:N (with as many as four sites) site replication. Level 1 and Level 2 support are provided by Dell EMC, with Level 3 assistance provided by VMware, although all interactions are still delivered to the customer through Dell EMC. The additional data protection and cloud capabilities create lock-in, but they also provide value-add, which is a typical trade-off.

VxRail deployments benefit from the familiarity customers have with Dell EMC and VMware products, but require upfront know-how and technical depth for optimal configuration and tuning. If your desired failures to tolerate (FTT) is two in a single-chassis configuration, then the recommended number of nodes is seven, which is a high starting point for some use cases. The additional data protection and cloud-tiering capabilities that Dell provides in VxRail exacerbates lock-in for customers.

VxRail addresses a broad set of use cases, including VDI, mission-critical databases, ROBO and back-office applications.

Dell VxRail Flex Notable Service Traits

• Hardware: Dell EMC PowerEdge and Quanta Servers underpin the VxRail offerings. 
• SDI: VxRail integrates software-defined compute and storage from VMware in its offering. 
• Virtualization (Hypervisors): VMware ESX is the only supported hypervisor. 
• Virtualization (Containers): VxRail has integration with the two most popular container orchestration engines — it has a Docker-certified volume plug-in and support for Kubernetes persistent volumes .
• Data Services: VxRail supports compression, deduplication, cloning, snapshots, replication and erasure coding. 
• Data Protection: Beyond native snapshots and asynchronous replication (available from VMware), synchronous replication capabilities are provided through the RecoverPoint appliances. 
• Management: VxRail manager is a purpose-built tool to manage the hardware elements of the cluster, as well as for appliance life cycle management. Most software capabilities are managed through vCenter Server. 
• Stack: VxRail supports most independent software vendors (ISVs) that are certified with VMware vSAN. 
• Scaling: VxRail can scale from three to 64 nodes in a cluster. 
• Service and Support: Dell EMC provides Level 1 and Level 2 support, and acts as the interface for Level 3 support issues jointly with VMware. 
• Security: VxRail offers native, software-based, data-at-rest encryption. Boot volumes, vMotion and other forms of data in motion can also be encrypted. 

HPE

Hewlett Packard Enterprise (HPE) completed the acquisition of SimpliVity in early 2017, and then immediately integrated and positioned the HPE SimpliVity 380 as the company's flagship hyperconverged product in place of the HPE StoreVirtual-based offering. It is now available on HPE ProLiant Gen10 servers as a scale-out, 2U, rack-mounted HCI building block that delivers advanced data services and data protection based on OmniStack software and all flash hardware. SimpliVity was designed around in-line, hardware-assisted global data reduction that offloads write-intensive workloads and provides guaranteed data reduction ratios, while ensuring that primary CPU resources are available to serve business applications.

HPE SimpliVity software provides guaranteed enterprise-class resilience features, such as hardware high availability, instant VM recovery, remote backup and restore, and site-to-site disaster recovery. SimpliVity RapidDR is disaster recovery orchestration software that helps customers fully automate their disaster recovery processes. SimpliVity's built-in data protection features go beyond traditional snapshot technology to provide a full-featured backup capability that enables enterprises to sunset their legacy enterprise backup solutions.

The vendor offers its infrastructure and data services (such as provisioning, cloning, restore and move a VM) through VMware vCenter. Under its Hyperguarantee initiative, SimpliVity guarantees that customers will achieve 10:1 data efficiency ratios in production systems and the ability to perform a local backup or restore of a 1TB VM in 60 seconds or less. SimpliVity lacks GA support for Hyper-V, KVM and Xen hypervisors.

HPE SimpliVity is often being shortlisted by enterprises of all sizes and verticals as a general-purpose consolidation platform for VMware workloads.

HPE SimpliVity Notable Service Traits

• Hardware: SimpliVity is deployed on HPE ProLiant DL380 servers with all solid-state drive (SSD) storage. SimpliVity leverages a PCIe based FPGA OmniStack Accelerator Card for in-line deduplication and compression to offload the main CPUs. 
• Software-Defined Infrastructure: HPE SimpliVity has full interoperability with VMware ESXi for compute and has built-in SDS services provided by the OmniStack Virtual Controller VM. 
• Virtualization (Hypervisors): HPE SimpliVity has full support for VMware ESXi 6.0 and 6.5, and all operations are fully integrated in the native vCenter Management. 
• Virtualization (Containers): HPE SimpliVity has no native support for containers; however, Docker containers can run inside VMware VMs. 
• Data Services: HPE SimpliVity provides global in-line compression and deduplication using the SimpliVity Accelerator card. SimpliVity guarantees 10:1 data efficiency. 
• Data Protection: HPE SimpliVity protects data using RAID inside the node and coping data across the cluster nodes. HPE SimpliVity provides native data protection, where every backup is a full logical backup of the VM. Backups can be stored locally or to remote SimpliVity clusters through a global policy engine. HPE SimpliVity supports two-site synchronous replication, stretched clusters and multisite asynchronous clusters. 
• Management: HPE SimpliVity is integrated directly into VMware vCenter and also provides REST APIs for management. 
• Stack: HPE SimpliVity is certified with SAP HANA and supports Horizon View and Citrix VDI platforms. HPE is integrating the HCI systems into HPE OneSphere Hybrid Platform management platform. 
• Scaling: HPE SimpliVity can start with single node and supports up to 24-node VMware clusters. 
• Service and Support: HPE provides bundled services, support and pricing, including third-party software. Customers have the option to buy VMware licenses and support from HPE. 
• Security: HPE SimpliVity leverages VMware hypervisor authentication, and supports data-at-rest encryption with the HPE Proliant Gen10 release. 

HTBase

During the past year, HTBase has added fundamental new capabilities. These offer a composable infrastructure OS (COS) that abstracts resources from disparate, on-premises hardware and public clouds: Amazon Web Services (AWS), Azure and Google. They are in a pool that's centrally managed, composed and delivered to business in multiple formats (VM, bare-metal or container). COS abstracts compute and storage from multiple cloud server providers to include cloud resources as part of the pool. Maestro, which is part of COS, administers cloud management that enables users to control hybrid cloud deployments and scale out resources across private and public clouds. As a default, Maestro offers an application marketplace that comes with a few prepackaged applications, such as databases and web servers for automatic deployment and provisioning.

The application market can be extended by customers with their own applications. Customers have added applications, such as middleware and big data modules. The composable capability enables leveraging of internal SSD or hard-disk drive (HDD) disks, as well as storage appliances as part of the pool of storage. Thus, customers can grow compute and storage completely independently from each other. In addition, HTBase is building its own in-memory module as part of COS, to achieve SAP certification and certification of in-memory database management systems (DBMSs) from Oracle and Microsoft.

Abstraction across all infrastructures (physical, virtual and multicloud) is by a thin, common software layer called OneCloud Hypervisor (OCH). HTBase enables high availability and resilience across multiple clouds without transformation. Machine learning enables quality of service (QoS) to be automatically adjusted, based on an application's utilization patterns. Machine learning enables workloads to be moved automatically, based on performance and/or cost. Using Slack for Chatbot enables monitoring and chat-based alerts of infrastructure failures or proactive monitoring. HTBase has developed most of its architecture by using a microservices approach that allows small parts of the code/application base to be updated without affecting the performance of the platform. The microservices model enables HTBase to offer template plug-ins in COS that administrators can leverage as a framework to develop further functionalities that include integration with existing applications.

HTBase recently closed a worldwide partnership agreement with Dell EMC to sell and support the HTBase composable solution. Through Dell EMC, HTBase offers a bundled service and support package and pricing, including the HTBase software and third-party software.

HTBase Notable Service Traits

• Hardware: HTBase recently announced a partnership with Dell EMC to sell, support and go to market with a composable infrastructure solution on Dell servers. 
• SDN: HTBase supports levels of abstraction that span HDDs, SSDs, flash, storage appliances and cloud storage across different cloud vendors. 
• Virtualization (Hypervisors): HTBase does not leverage third-party tools (such as vCenter) to control different hypervisors. Rather, all integration and functionality is built into COS, using KVM as the base hypervisor. 
• Virtualization (Containers): HTBase enables Docker deployments to be integrated with the storage layer and data volumes to persist beyond the lifetime of a single container machine host. 
• Data Services: HTBase supports compression, deduplication, snapshot and cloning. 
• Data Protection: HTBase synchronously replicates data to one or more nodes in the cluster, ensuring that all data exists in at least two independent locations with failover. 
• Management: Maestro provides cloud management that enables users to control hybrid cloud deployments, as well as scale out resources across local and public clouds. 
• Stack: HTBase has documented and open APIs, along with a plug-in template that supports the integration of third-party software and stacks, which are currently in the validation process. 
• Scaling: COS leverages internal SSD or HDD disks, as well as storage appliances, as part of the storage pool to grow compute and storage independently. 
• Service and Support: HTBase offers strong infrastructure and HCI support; however, clients should check geographic consistency. 
• Security: Compliance and support is mostly through third parties. 

Huawei

FusionCube: Huawei's HCI systems are based on the FusionCube and support all use cases satisfactorily. Huawei has strong R&D capabilities for hardware and support; however, it lacks support in some storage functions, such as in data reduction, containers and integrated SDN. Huawei continues to invest in its storage business and SDI. Gartner expects the company to add more functionality, including more storage capabilities, data protection and open-source support, (e.g., OpenStack) in 2018. Huawei's HCI products target a variety of workloads, ranging from virtualization platforms to DBMS and mission-critical applications (e.g., SAP HANA). FusionCube is expanding its capabilities, including OpenStack, VMware and KVM support, as well as 100Gb InfiniBand, while adding more storage features, such as NVMe SSD, active-active disaster recovery, in-line deduplication and in-line compression. 

Huawei has developed a comprehensive HCI software suite of homegrown software, including ManageOne (unified management), FusionInsight (big data), FusionStorage (software-defined anything [SDx]), FusionAccess (VDI) and FusionSphere (virtualization, cloud platform). FusionCube has good cloud management tools, and early success came from outside China.

End users interested in Huawei's SAP HANA platform can feel comfortable that the product will be supported in markets outside China. FusionCube can be considered as a hybrid cloud platform. To enhance its efforts, Huawei plans to focus more on developing hybrid cloud ecosystems, involving ISVs, service providers and carriers, while adding container and other hypervisor support in extended geographies outside China.

Huawei Notable Service Traits

• Hardware: Huawei's FusionCube HCIS has been the heart of its HCI offering and software releases since its 2009 launch. 
• SDI: FusionCube supports universal standardized SDI services across heterogeneous legacy infrastructure, HCIS and HCI levels of abstraction that span HDDs, SSDs, flash, and storage appliances. 
• Virtualization (Hypervisors): Huawei supports its own FusionSphere release, which was initially based on Xen. More recently, it has added KVM and VMware support. 
• Virtualization (Containers): FusionSphere supports Linux hosting with Containers through KVM and through a virtual switch, as well as third-party support with VMware. 
• Data Services: Huawei Fusion Storage Services supports compression, deduplication, resource pools, application templates and VM resource management. 
• Data Protection: FusionCube supports snapshot and cloning of local disks and across the network to alternative nodes and a cloud backup solution. 
• Management: FusionCube Center manages Huawei's HCI software and HCIS Systems, through the integration for FusionSphere management and the hardware-based UMate Server management tools. 
• Stack: Huawei FusionCube is certified with SAP HANA and integrates with Citrix, VMware, Microsoft and now Oracle. However, the depth and breadth of third-party support of solutions built on Huawei needs qualifying to include ISV offerings in individual regions. 
• Scaling: Huawei's FusionCube supports up to 256 nodes in 128 clusters to a maximum of 4,096 nodes. 
• Service and Support: Huawei offers strong infrastructure and HCI support; however, clients should check geographic consistency with local references. 
• Security: Huawei offers compliance and support separated into tiers around virtualization, management, network, infrastructure and physical infrastructure security. 

Microsoft

Microsoft Windows Server 2016: Microsoft entered the HCI market with the introduction of Microsoft Storage Spaces Direct, part of the Windows Server 2016, released in October 2016. Microsoft's HCI solution is not independently monetized; rather, it's included in the Windows Server Data center license. Microsoft has significant potential against longer-standing HCI suppliers, based on its enormous installed base of Windows Server licenses and global ecosystem of hardware vendor partners. Even a small percentage of Windows Server 2016 customers deploying as HCI will result in Microsoft becoming a significant presence. Microsoft's HCI deployment supports a minimum of two and maximum of 16 servers, all of which are recommended to be of the same make, model and configuration. Servers can be from a broad range of certified vendors. Key to Microsoft's long-term differentiation is integration with Microsoft's Azure for backup and disaster recovery. By scaling down to two nodes, Microsoft HCI can be used in ROBOs of large organizations; however, it can also scale to meet data center performance and VM-density requirements. 

Microsoft leverages a common software stack, whether it is being deployed for HCI, network-attached storage (NAS) file serving, on bare metal or in a cloud. Microsoft's recent enhancements have been in the area of security and performance, including the showcasing of Windows HCI performance, as part of the Intel Purley launch. Microsoft management is optimized for large organizations with knowledge of scripting and management through System Center; however, it lacks a simple, affordable interface appropriate for small customers. Storage spaces direct will offer a free-to-implement extension only in data center versions of Windows Server and not smaller footprints.

Windows Server 2016 is a good choice as HCI for Microsoft customers looking for an HCI solution for VM and VDI consolidation, modernization initiatives, and ROBO installations with large numbers of sites. Customers praised the performance of the Microsoft solution, as well as the ability to avoid hardware vendor lock-in.

Microsoft Notable Service Traits

• Hardware: Servers from Data On, Dell EMC, Fujitsu, HPE, Intel, Lenovo, Quanta and SuperMicro are certified, with many more in process. All servers in a cluster are preferred to be of identical make, model, configuration and firmware release. 
• SDI: Windows Server 2016 Datacenter edition provides integrated SDC (Hyper-V), storage (Storage Spaces Direct) and networking. Networking functions include routing, firewalls, load balancing, gateways and network QoS. 
• Virtualization (Hypervisors): Microsoft supports Hyper-V and a cluster Cloud Witness in Azure to enable improved resiliency in two-node, remote-site implementations, but lacks support for VMware ESXi, KVM or Xen. 
• Virtualization (Containers): Microsoft provides full integration for the Moby/Docker engine, with support for Docker Enterprise Edition Basic at no additional cost. Other container support is in development. 
• Data Services: Microsoft supports NTFS and ReFS file systems. NTFS provides deduplication, compression and snapshotting through volume shadow copy service (VSS). ReFS supports VSS and three options for parity to meet performance and efficiency requirements. 
• Data Protection: Microsoft supports synchronous and asynchronous volume replicas with Storage Replica and support for VM-level replicas using Hyper-V Replica. However, it has no support for n-way replicas. Microsoft has a broad ecosystem of certified backup software solutions. 
• Management: Microsoft provides management capabilities through published APIs, PowerShell and System Center (for an additional license fee). Error and performance monitoring is available through Systems Center Operations Manager. 
• Stack: Microsoft maintains extensive support of mainstream Microsoft and third-party line of business (LOB) and back-office applications certified for use with Hyper-V. 
• Scaling: Microsoft supports as many as 16 nodes in an HCI configuration. Most customers are two to four nodes per site. 
• Service: Customers may contract for full-stack HCI support through a Microsoft hardware partner. They can also contract separately for hardware support from the hardware supplier and software support from Microsoft. 
• Security: Microsoft supports role-based access control (RBAC), volume-level encryption and VM-based encryption, as well as encryption of intracluster communication. Windows Server is FIPS 140-2 and Common Criteria compliant. 

Nutanix

Known for premium-based value solutions, Nutanix has made a significant shift from its original position as an HCIS appliance vendor to a multiple consumption model, with a range of pricing. Included are:

• Appliances with preinstalled HCIS software, with perpetual licenses
• Software for third-party server hardware (e.g., Cisco UCS and HPE ProLiant)
• Enterprise license agreements (ELAs) for larger, global enterprises, which were scheduled to be available at the end of 2017
• A cloudlike consumption program that gives users the option of blending traditional capital purchases (buying) with operating expenditure (opex)-driven acquisition (renting)
• Xi Cloud Services, which are planned to be available early in 2018, enabling users to consume SaaS, by leveraging stack and tooling of on-premises deployments to run all applications in an integrated cloud environment

Nutanix's principle strategy is:

• Invisible infrastructure to seamlessly merge compute, storage, networking, hypervisor, applications and clouds
• One-click operations across the full infrastructure and application life cycle, reducing the need for IT specialists
• Maximum platform choice (OEMs) and runtimes (VMs and containers)
• Maximum hypervisor choice — AHV, ESXi, Hyper-V and XenServer
• Clouds (Xi, AWS, Azure, GCP); application styles (Mode 1 and Mode 2) and consumption models (appliances, software references, buy and subscribe)

Nutanix Notable Service Traits

• Hardware: Hardware configurations cover high-performance databases (NX8150 All Flash) and high-capacity big data workloads (NX6155). They enable dynamic and nondisruptive increases or decreases in capacity, using nodes with differing compute and storage configurations. 
• SDI: In addition to making its solution available via software for installation on third-party hardware, Nutanix delivers SDC via virtualization across all hosts with per-workload choice of hypervisors. Storage controllers run as VMs on each cluster node, which aggregates and abstracts DAS using Acropolis. 
• Virtualization (Hypervisors): In addition to its own AHV, based on the open-source KVM project, Nutanix supports all major hypervisors. Nutanix has extended KVM with security hardening, platform-specific performance tuning and added enterprise features not in the standard version of KVM. 
• Virtualization (Containers): Nutanix Calm delivers DevOps application-centric automation via blueprints for services deployed, managed, upgraded and retired as containers; monitoring and analysis of container events and alerts; and predictive resource contention and remediation. 
• Data Services: Nutanix provides compression and deduplication using in-line and postprocess techniques. However, it must be enabled or disabled for specific datasets. Deduplication and erasure coding are data-type adaptive to avoid expending resources on data with low benefits. 
• Data Protection: Nutanix, in addition to providing native BU and snapshots, leverages a growing list of backup partners for local snapshots, backup and disaster recovery to secondary clusters, the public cloud and its Xi Cloud services. Partners include Veritas (2HC17), Veeam (1HC18), Rubrik, Druva and Comtrade. 
• Management: Prism provides a single control plane to offer the same user experience across hypervisor deployments. A REST API is available as a central means of managing the Nutanix cluster with selective command functionality exposed. 
• Stack: Nutanix supports a range of third-party workloads, as well as cloud stacks, including OpenStack, vRealize Automation, Microsoft Azure and Pivotal Cloud Foundry. 
• Scaling: Nutanix does not use internal or proprietary backplanes between nodes, which communicate using standard 10 Gbps or 40 Gbps Ethernet; cluster services can also use a mesh topology, so any node communicates with any other without relying on a central node. 
• Service and Support: Nutanix OEM partners manage support calls in the first instance concerning hardware failures; if the issue is within the software stack, partners pass support calls to Nutanix. 
• Security: Nutanix publishes a Security Development Lifecycle (SecDL) as a broad, cross-functional process run by its Security Engineering and Research Team. This provides attack defense and hardening as the default in GA releases, in addition to hardware/software-based data encryption and hypervisor hardening. 

Pivot3

Pivot3 vSTAC is a mature HCI offering that has been shipping for more than five years, and the core functionality, which was first incorporated into video surveillance offerings, has been shipping for more than 10 years. Pivot3 announced the Acuity product on 27 April 2017, which leverages vSTAC and technology from Pivot3's acquisition of NexGen Storage. With Acuity, Pivot3 introduced NVMe flash technology to better support Tier 1 workloads, as well as policy-based management capabilities to enable dynamic reprovisioning to meet SLAs for critical workloads.

Acuity uses patented erasure coding that enables large clusters and consistent performance with minimal impact during recovery from drive and node-failure conditions. Although vSTAC is primarily used in video surveillance and VDI implementations, where consistent throughput at scale is critical, the Acuity platform is Pivot3's platform for general-purpose and business-critical workloads, including databases. Pivot3 creates a single storage resource that distributes data evenly across all drives and nodes, increasing performance and removing the need to colocate data with its VM.

Acuity lacks native support for Hyper-V, KVM and Xen, and it requires third-party software to enable asynchronous replication and synchronous replication for two-way and n-way clusters. The minimum configuration required for high availability is three nodes. Acuity is a good choice for customers seeking an HCI solution for business-critical applications in a VMware-virtualized environment. Customers praised the quality of Pivot3's pre- and postsales support.

Pivot3 Notable Service Traits

• Hardware: Acuity can be deployed as a Pivot3 hardware appliance or through its OEM relationship with Lenovo. 
• SDI: Acuity enables SDS services and, when combined with VMware ESXi, vSphere, and vRealize, enables SDC, networking and management. 
• Virtualization (Hypervisors): Acuity supports VMware ESXi and can provide storage for Hyper-V, KVM and Xen by presenting iSCSI volumes from a scale-out virtual SAN. 
• Virtualization (Containers): Acuity supports containers and container management through the VMware vSphere Integrated Containers feature and vCenter. 
• Data Services: Acuity uses in-line erasure coding to create an aggregated pool of storage, with as high as 94% storage efficiency. vSTAC supports file- and application-level deduplication and compression. 
• Data Protection: Acuity provides snapshot-based data protection with administrator-assigned protection policies. Acuity supports as many as 10,000 recovery points and provides application-integrated recovery points at no additional charge. 
• Management: Acuity is managed through a vSphere web client Plugin. RESTful APIs are not public, but are available to development partners and customers on an as-needed basis. 
• Stack: Acuity supports all business applications certified to run under VMware's ESXi. 
• Scaling: Acuity has no theoretical limit to cluster size. 
• Service: Support is available directly from Pivot3 or through its server and application vendor partners. 
• Security: Acuity integrates with AD/LDAP through VMware's Single Sign On. Pivot3 partners with HyTrust for encryption of data at rest and in motion and to enable FIPS 140-2 compliance. 

Scale Computing

Scale Computing has been shipping its SMB-oriented hyperconverged products since 2012. Scale's Hyperconverged Compute Cluster (HC3) System, the HyperCore hypervisor, and the Scale Computing Reliable Independent Block Engine (SCRIBE) SDS are bundled in a turnkey appliance. This provides midmarket enterprises with an all-inclusive, cost-effective HCI product. Scale has built its product on the KVM hypervisor. Its proprietary, kernel-integrated SDS pools all virtual disks to be presented as local storage on the HC3 nodes. Scale claims that this method yields better system utilization, because it avoids the management overhead of virtual storage appliances (VSAs), which are used by most HCI vendors.

Scale now has HC3 hardware models ranging from small to high capacity, which can be mixed and matched in the cluster. Scale also supports all flash- and NVMe-enabled systems. During the past 12 months, Scale has added support for single-node deployments, automated data tiering and deduplication. The vendor now features data recovery as a service (DRaaS), with failover and failback capabilities through Scale direct, MPs, and public cloud partners. HC3 is based on the KVM hypervisor and cannot accommodate end users that have standardized on VMware or Hyper-V hypervisors.

Scale is a popular choice in the North American midmarket and with distributed enterprises looking for VMware alternatives and a single-vendor support solution.

Scale Computing Notable Service Traits

• Hardware: Scale provides a wide range of hyperconverged hardware models, but also can qualify and support other x86 hardware. HC3 systems can be mixed and matched across any hardware types within the same cluster. 
• Software-Defined Infrastructure: Scale embedded its proprietary SDC HyperCore OS and SDS SCRIBE as the basis of its HCI platform. For SDN overlay, Scale uses a combination of proprietary technology and Open vSwitch/VxLAN. 
• Virtualization (Hypervisors): Scale HC3 can only run the KVM hypervisor natively. HC3 has migration tools to import VMs from other hypervisors. 
• Virtualization (Containers): Container support is limited to deployments as an in-guest VM. 
• Data Services: Scale provides deduplication and thin VM cloning. Scale's SCRIBE SDS includes automated data tiering, with the most frequently used blocks stored on a flash tier. 
• Data Protection: Scale supports asynchronous replication in multisite deployments and provides backup based on VM-level snapshots. Scale also provides per VM DRaaS to replicate and start VMs in the Scale Computing cloud. 
• Management: Scale offers an HC3 web-based user interface for multicluster management. 
• Stack: Scale is validated as Citrix Ready for XenDesktop and certified for Microsoft RDSH. Scale supports fully integrated VDI with Workspot and offers the Scale Analytics complete business intelligence (BI) solution with Information Builders. 
• Scaling: The HC3 cluster can scale from three to eight nodes and can mix and match different hardware models with the ability to centrally manage up to 25 clusters through the HC3 Web Interface. They also offer a single-node appliance for distributed data centers or ROBO deployments. 
• Service and Support: Scale HC3 appliances offer end-to-end product support, software updates and a hardware warranty. Scale has Premium installation and disaster recovery planning services. 
• Security: HC3 is built on a locked-down, purpose-built OS that does not offer encryption capabilities. 

Stratoscale

Stratoscale Symphony is a software-based HCI, offering hardware-neutral and service-driven hybrid cloud solutions. It provides a full-stack private cloud infrastructure as a service (IaaS) solution, as well as an application catalog, monitoring tools, fully managed database service, and turnkey OpenStack and Kubernetes. Symphony's management tools monitor resource consumption and automatically migrate resources based on priorities. Symphony's self-service capabilities enable users to access all functions directly. Symphony has on-premises, multitenant capabilities with quota limitation and chargeback.

Symphony's AWS-compatible, on-premises infrastructures move workloads on- and off-premises dynamically, based on a predefined quota and policies. Symphony also can provision services on AWS, in addition to on-premises. Its AWS-compatible services include EC2, EBS, S3, VPC, ELB and RDS. Symphony does not manage outside of the Symphony cluster; however, it offers extensive APIs to manage nodes. In addition, Stratoscale has reference architectures with major system providers, such as Cisco, Dell, SuperMicro, Lenovo and HPE. Symphony's main use cases include data center modernization (with flexible hardware choices), hybrid cloud deployment, involving application portability between on- and off-premises, "lift and shift" data center, and container deployments for DevOps (Kubernetes).

Symphony's primary targets are end users that want to deploy an AWS-compatible infrastructure on-premises, while leveraging existing hardware to implement an HCI solution. Symphony lacks Azure and Google Cloud compatibility; however, it plans to add these services in 2018.

Stratoscale Symphony Notable Service Traits

• Hardware: Symphony offers hardware-agnostic solutions (any x86 servers). 
• SDI: Symphony is integrated into SDS and SDN layers and provides flexible choices of deployments using VMs or containers. 
• Virtualization (Hypervisors): Symphony uses KVM, but also supports migration from other VMs, such as VMware's vSphere or Microsoft Hyper-V, or other KVM-based hypervisor. 
• Virtualization (Containers): Symphony supports all container platforms and offers a fully integrated and managed Kubernetes service. 
• Data Services: Symphony unifies storage pools, including DAS and external storage arrays, as well as public cloud without limitations. Its thin provisioning ensures that only written data consumes storage space. It also offers compatible APIs with AWS S3 and EC2. 
• Data Protection: The data is replicated across multiple nodes in the cluster. Automation of protection and recovery management is achieved via a comprehensive set of REST APIs. Data protection occurs at the block level. The target can be either block or object storage. 
• Management: Single-pane management covers the entire software stack without third-party management software. 
• Stack: Symphony's focus has been building AWS-compatible services and releasing new code frequently. 
• Scaling: Symphony has no limitation for the maximum number of nodes and fully supports heterogeneous clusters. It can add nodes to scale without downtime (autoscaling) and also offers elastic capacity by extending workloads into public cloud. 
• Service and Support:Its annual per-node subscription includes 24/7 dedicated support, maintenance and product upgrades. 
• Security: Multitenant isolation and security provides separation between authentication, credentials and security policies managed by Symphony. Symphony recently announced a partnership with security vendor Safe-T to protect on-premises cloud services. 

VMware

VMware's vSAN product was launched in 2014 and is natively integrated with the ESXi hypervisor kernel. vSAN is available in standard, advanced and enterprise editions, with key data reduction features only available in the advanced and enterprise editions. vSAN's native integration with vSphere, vCenter and vRealize operations simplifies day-to-day administration, upgrades and monitoring of the environment. Features such as read/write caching on an SSD tier, snapshots, replication and QoS (through which there is a maximum input/output operations per second [IOPS] limit on a Virtual Machine Disk [VMDK] basis) are available in all software editions. VMware enhanced stretched clusters in the recent 6.6 software release, which enables customers to build active-active, geographically distributed sites that are load-balanced, providing high availability against unforeseen node or site failures.

The vSAN product has several quirks that can be limiting factors from a total cost of ownership (TCO) standpoint. Customers cannot mix disk group types (all-flash and hybrid) in a cluster. Key data services (e.g., compression, deduplication, erasure coding and encryption) aren't available in the standard edition. In addition, all of the above features, except encryption, require customers to have an all-flash configuration. VMware has built joint support and close integration with Dell EMC, which sells the turnkey VxRail appliance. VSAN is deployed across a broad set of use cases with server consolidation, business-critical applications, VDI and ROBO being the prominent ones.

VMware Notable Service Traits

• Hardware: vSAN is a software platform that can be deployed on certified hardware components or on vSAN ReadyNodes supplied by hardware partners or available as an appliance called VxRail from Dell. 
• SDI: VMware provides competitive SDC and SDS in its core offering. With VxRack SDDC, which is a rack scale appliance, customers also get NSX, which provides SDN capabilities. 
• Virtualization (Hypervisors): vSAN is deeply integrated and supports key functionalities of the ESXi hypervisor. VSAN can't be deployed on third-party hypervisors. 
• Virtualization (Containers): vSAN is integrated with the two most popular container orchestration engines — it has a Docker-certified volume plug-in and support for Kubernetes persistent volumes. 
• Data Services: vSAN supports compression, deduplication, cloning, snapshots, replication and erasure coding. 
• Data Protection: vSAN enables asynchronous replication, snapshots and stretched clusters. 
• Management: Organizations can manage vSAN through vCenter and can use vRealize Operations tool for more advanced monitoring and analytics. 
• Stack: vSAN enjoys broad ecosystem support for hardware OEMs and ISVs. In addition, the HCI product is consumable from IBM and AWS cloud through VMware's partnerships. 
• Scaling: vSAN can scale from two nodes to 64 nodes in one-node increments. The quorum node can be deployed remotely either in the main data center or in the cloud in two-node deployments. 
• Service and Support: VMware provides support for the entire software stack. Third-party hardware OEMs may provide first- or second-level support for software. 
• Security: vSAN now offers native-software-based, data-at-rest encryption. Boot volumes, vMotion and other forms of data in motion can be encrypted as well. 

Context

Hyperconvergence continues to disrupt the market for integrated systems, with major system vendors now joining the ever-growing number of startups (some of which are now mature themselves). HCIS includes HCI software; however, HCI software is not limited to a system (hardware appliance) deployment model. Hardware-appliance deployments (HCIS) continue to outpace software-only/bring-your-own server, reference architectures, cloud and as-a-service deployments. The advantages of software-only deployments, which include avoiding hardware vendor lock-in, are offset by the added complexity of the support model and the inability of software-only vendors to test and certify the myriad of configuration options customers may choose.

In some cases, hardware-vendor optionality can be achieved as HCI software solutions (e.g., Nutanix, Pivot3 and VMware) become available as HCIS appliances through multiple vendors. However, the support model becomes more complicated when users choose multiple HCIS vendors for the same HCI software. Thus, we see many HCI customers obtaining HCI products as appliances, based on a prior relationship with a preferred server vendor.

One limitation of the typical HCIS appliance model is that compute, storage and networking do not scale in tandem for all workloads. Vendors looking to compete across the broadest range of workloads are beginning to or will soon offer compute-only and storage-only nodes. Users should conduct a proof of concept (POC) to carefully evaluate the compute, storage and networking requirements of their workloads running on HCI. They should also estimate the component growth requirements, to determine the need for unbalanced scaling.

Although there are many deployments of more than 100 nodes, most HCI implementations can be measured in tens of nodes or fewer. As HCI is more broadly adopted across a wider range of nonhomogeneous workloads, requirements will increase for HCI to operate more autonomously. This will include the ability to automatically provision, rebalance, adapt to meet QoS requirements, detect anomalies, and prevent failures and data loss. When HCI is deployed at large scale, these capabilities become increasingly necessary and key points of vendor differentiation.

One of the attractions of HCI is the potential to create a cloudlike provisioning model, while maintaining physical control of IT assets and data on-premises in the data center, remote site or branch office. During the next few years, cloud deployment models will be developed to meet short-term, scale-up/scale-down needs, as well as backup and disaster recovery requirements. An important question for users is whether HCI is a steppingstone to the cloud or a "foreseeable future" resting place for applications.

HCI-based solutions will increasingly be adopted; however, outside smaller organizations, HCI is unlikely to become a full-service platform for IT services across all workloads and delivery models of on-premises, hosted off-premises and native cloud. Customers should evaluate HCI solutions and select vendors and products, not because HCI or that vendor is rapidly growing, but because it fits with their particular use case, growth expectations and application architecture direction. HCI is likely to become yet another silo to manage, so integration with higher-level management frameworks, including cloud management platforms (CMPs), becomes key to supporting an already-overtaxed operations staff.

Product/Service Class Definition

Critical capabilities are attributes that differentiate products/services in a class, in terms of their quality and performance. This methodology required analysts to identify the 11 critical capabilities for this class of products. Following is a summary of the selected capabilities for each category.

Critical Capabilities Definition

Hardware

The hardware attributes of these offerings include key differentiators, such as server/storage OEM/ODM certified with the solution; predictive failure analytics; large-scale memory in support of DDR4 DRAM; benchmark data on SPEC, IOPS, latency and throughput; and I/O optimizations for flash.

In addition, the hardware critical capability includes lights-out support and definitive templates for supported hardware.

SDI

This critical capability assesses the overall viability of SDI in SDx. Key differentiators include SDC support, SDS support, integrated SDN, support for VMware NSX, support for Cisco ACI, full life cycle management for SDI, deployment to and management of AWS/Azure, and CMP (own and third party).

Other differentiators involve software-defined facilities/data center infrastructure management (DCIM) integration/support, as well as the level of CMP support and integration.

Hypervisors

This critical capability assesses the breadth and depth of hypervisor support, predominantly in a Mode 1 use case. Key differentiators include hypervisor support (VMware, Hyper-V, Xen, KVM and own offering); live migration assistance; and host-based recovery/failover.

Other differentiators involve storage migration and mixed hypervisors in cluster support.

Containers

This critical capability assesses the breadth and depth of container support, mainly in Mode 2 use cases. Key differentiators include container technology/management support (e.g., Docker, Mesos, Kubernetes, LXC and rkt), as well as support for a container-optimized, MicroOS, lightweight guest OS.

Additional differentiators include support for persistent container storage, container/hypervisor integration and management, and storage virtualization.

Data Services

This assesses the storage functions and services as part of the storage solution. Key differentiators involve publicly available benchmarks of performance for reads/writes of large and small objects, and storage efficiency/data reduction features (compression, deduplication, single-instance store).

Other differentiators include the percentage of the customers that implement data reduction features; data efficiency schemes used in the storage service (e.g., erasure coding); tiering capabilities that allow data to be automatically placed on optimal and appropriate classes of storage by policy; and the average compression and deduplication ratio for most use cases. In addition, there is support for cache tiering; VM-level data service management; maximum/minimum throughput per VM; and in-line, federated, or serial compression and deduplication in performance and capacity tier.

Data Protection

This critical capability involves overall viability of just the storage offering, with the following key differentiators: third-party ISVs for backup/archiving; metadata distribution among nodes/modules; the stated controller, node and brick failover times; and software/hardware updates.

Additional differentiators include fault isolation techniques that ensure high availability; guaranteed availability (4x 9s, 5x 9s, etc.); replication to the public cloud; snapshots and backup; ability to disable high availability/data protection on a per-VM or per-disk basis; and metro-cluster support.

Management

This critical capability rates the viability of key management differentiators — e.g., logical, physical partitioning or multitenancy features; self-service and metering features, with showback/chargeback analysis; available end-user training; and web or mobile applications available for monitoring.

In addition, there is support for hybrid (private/public) cloud integration to manage multisite deployments; I/O optimization technique (hardware and software) available for performance tuning; management console (own and/or third party), with published APIs/software development kits (SDKs); and monitoring for error conditions/performance/trending capacity analytics. Finally, there is automated policy enforcement, remediation, alert prioritization and grooming, as well as Plug-in for VMware vCenter and/or Microsoft System Center VM Manager (SCVMM).

Stack

This critical capability assesses the software stack of the vendor. Key differentiators include integrated, vendor-owned stack/third-party stack; supported SAP HANA appliance/SAP HANA Tailored Datacenter Integration (TDI); support for mainstream VDI platforms; and prepackaged VM templates.

Other differentiators involve official support by mainstream LOB application vendors (including SAP and Oracle); public, searchable repositories of officially supported applications/ISVs; and support for mainstream BI, analytics platforms and database software. In addition, there's support for mainstream platform as a service (PaaS) environments, Java and .NET application platforms, as well as cloud APIs (e.g., Amazon S3 and OpenStack Swift).

Scaling

This scaling critical capability assesses such key differentiators as support for n-tier, monolithic, mission-critical DBMS and application-serving workloads; support for vendor-owned/third-party/high-availability failover; and workload recoverability (LAN-based synchronous/WAN-based asynchronous).

Additional differentiators include support for cross-cluster live migration; snapshots into DBMS support; DBMS cluster support (e.g., load balancing and high availability); scale-out application-serving support; and LAN/WAN-based recoverability. Finally, this also includes support for Hadoop/Splunk/Spark, as well as snapshots into application server support.

Service and Support

This critical capability rates the viability of service and support, including single-vendor support contact, dedicated service account manager, minimum three- to five-year return to vendor (RTV) warranty, third-party product support, 24/7 live support and secure handling/disposal of customer data.

Additional differentiators involve roadmaps provided to customers and hardware compatibility lists.

Security

The security critical capability assesses several key differentiators. These include software patch management, denial of service (DoS) and distributed denial of service (DDoS) defense support, and support for published hypervisor hardening best practices and encryption of guest VMs.

Other differentiators include Secure Sockets Layer (SSL) + Transport Layer Security (TLS) encryption of control plane traffic, management support of RBAC, management graphical user interface (GUI) support of multitenancy, and support of security information and event management (SIEM) tools. In addition, there is FIPS 140-2 Level 2 compliance; support for full disk encryption (FDE)/self-encrypting disks (SEDs); support of network forensic tools (NFTs), such as Blue Coat and FireEye; and support for network intrusion prevention systems (NIPSs).

After identifying the critical capabilities, Gartner analysts then weight each capability to describe its relative importance for the individual use cases. Because the total cannot exceed 100%, this forces us to allocate the relative value or importance of each capability according to the five use cases. The capabilities each feed into one of two common categories:

• Offering/product strategy
• Product and service

These, in turn, feed into the Magic Quadrant, one for vision and one for execution.

Use Cases

Consolidated

Consolidated data center infrastructure HCI projects are driven by cost and TCO savings. They consolidate n-tier architectures supporting Mode 1 applications.

Business-Critical

Business-critical HCI projects improve resiliency and scalability of business-critical applications — e.g., ERP and packaged applications that require vendor certification.

Cloud

These HCI projects are deployments of Mode 2 applications and rearchitected Mode 1 applications in private or public cloud environments.

ROBO

Non-data-center locations may be managed remotely by data centers. HCI in ROBO is used to bridge Mode 2 IoT/edge infrastructure with Mode 1 data center infrastructure.

VDI

VDI benefits from HCI by simplifying the provisioning and hosting of user VMs in dedicated deployments. VDI is accessed over a LAN or WAN, using a remote display protocol.

Vendors Added and Dropped

We review and adjust our inclusion criteria for Critical Capabilities as markets change. As a result of these adjustments, the mix of vendors' products in Critical Capabilities research may change over time. A vendor's products may appear in a Critical Capabilities report one year and not the next. This does not necessarily indicate that we've changed our opinion of that vendor. It may reflect a change in the market and, therefore, changed evaluation criteria, or of a change of focus by that vendor.

Added

Not applicable.

Dropped

Not applicable.

Honorable Mentions

Several vendors did not meet one or more of the inclusion criteria for this revised Critical Capabilities research; however, they deserve honorable mention and consideration from organizations looking to benefit from HCI software.

Fujitsu

Fujitsu is a global server and storage provider offering HCIS, based on HCI software from Microsoft, Nutanix and VMware.

Hitachi Vantara

Hitachi is a global server and storage provider offering HCIS, based on HCI software from VMware.

Lenovo

Lenovo is a global server and storage provider offering an extremely broad portfolio of HCIS and reference architectures based on HCI software from Microsoft, Nutanix, Pivot3, Scale Computing, Stratoscale, Red Hat and VMware. Lenovo has become a partner of growing importance for many HCI software vendors. In part, this is because it has a global reach and no competing, internally developed HCI software.

New H3C

New H3C is a China-headquartered supplier of IT solutions, including HCI, and the exclusive provider of HPE server and storage products in China.

Red Hat

Red Hat is a U.S.-headquartered, global, open-source software developer and supplier of open-source software solutions and support. Red Hat offers HCI integrating Red Hat Virtualization and Red Hat Gluster Storage targeting ROBO requirements, as well as mobile edge computing, and microdata centers supporting IoT edge applications.

Sangfor

Sangfor is a China-headquartered provider of network security, network optimization and virtualization solutions. Although the company is expanding geographically, Sangfor's HCI products have been sold almost exclusively in the Asian region.

Starwind

Starwind Software is a developer of storage virtualization software that delivers HCI solutions primarily targeted at ROBO of midsize and large enterprises.

Inclusion Criteria

To qualify for inclusion in the HCI Critical Capabilities research, vendors need to meet the following criteria.

Functional Criteria

Included HCI vendors must:

• Provide a software stack that includes SDC, storage and, optionally, networking with integrated software services and management.
• Deliver the solution as a hardware appliance, as software-only for use on certified, reference architecture, or use as a service in on-premises or public cloud infrastructure.
• Provide a solution that uses local DAS storage, rather than shared storage, such as a SAN and/or NAS. HCI products must provide a mechanism to pool DAS into logical, abstracted virtual storage.
• Develop the storage and data management services integrated in the offering, not resell a white-label solution produced by another company in an OEM relationship.

Business Criteria

Eligible HCI vendors must:

• Provide evidence of a minimum of 50 production customers brought to revenue in each of at least two of the major geographies — Americas, EMEA, and Asia/Pacific (APAC) region and Japan.
• Deliver complete Level 1 (call center/service desk) and Level 2 (escalation) support to facilitate quick and easy problem resolution. However, Level 3 (engineering) support can be delivered separately, based on vendors' engineering partnerships.
• Have achieved $10 million in bookings from HCI solutions for the preceding 12 months ending 30 April 2017.
• Be delivering solutions in at least three of the use cases identified in Critical Capabilities.

Integrated systems that fall into the IIS or ISS categories are no longer eligible to participate in this Critical Capabilities research.

• IISs are server, storage and network hardware integrated to provide shared compute infrastructure. Examples include Dell EMC VxBlock Systems and HPE ConvergedSystem 700.
• ISSs are server, storage and network hardware integrated with application software to provide appliance or appliancelike functionality. Examples include Oracle Exadata Database Machine and Teradata.

It is not mandatory that inclusion criteria for Magic Quadrants and Critical Capabilities be common; however, we believe it is desirable for the list of assessed vendors to be identical in both research deliverables.

Table 1: Weighting for Critical Capabilities in Use Cases

Critical Capabilities	Consolidated	Business-Critical	Cloud	ROBO	VDI
Hardware	12%	10%	4%	7%	8%
SDI	7%	8%	10%	7%	9%
Hypervisors	12%	10%	7%	9%	10%
Containers	6%	3%	10%	4%	8%
Data Services	10%	10%	11%	10%	9%
Data Protection	8%	13%	8%	12%	9%
Management	12%	10%	12%	15%	13%
Stack	8%	11%	14%	8%	7%
Scaling	11%	9%	12%	7%	11%
Service and Support	6%	7%	6%	13%	8%
Security	8%	9%	6%	8%	8%
Total	100%	100%	100%	100%	100%
As of January 2018

Source: Gartner (February 2018)

This methodology requires analysts to identify the critical capabilities for a class of products/services. Each capability is then weighed in terms of its relative importance for specific product/service use cases.

Critical Capabilities Rating

Each of the products/services has been evaluated on the critical capabilities on a scale of 1 to 5; a score of 1 = Poor (most or all defined requirements are not achieved), while 5 = Outstanding (significantly exceeds requirements).

Table 2: Product/Service Rating on Critical Capabilities

Critical Capabilities	Cisco	DataCore	Dell EMC	HPE	HTBase	Huawei	Microsoft	Nutanix	Pivot3	Scale Computing	Stratoscale	VMware
Hardware	4.2	3.7	4.5	4.4	4.0	4.4	3.0	4.5	4.2	4.5	4.0	3.8
SDI	4.1	3.8	4.2	4.1	3.5	3.9	2.7	4.1	3.1	3.5	3.4	4.5
Hypervisors	3.2	2.9	4.3	4.2	3.9	4.1	4.2	4.5	3.5	4.1	3.5	4.5
Containers	2.2	2.1	3.4	3.2	3.0	2.9	3.8	3.3	2.5	2.1	3.0	3.9
Data Services	3.0	4.0	4.0	4.0	3.0	2.0	2.5	4.0	4.0	2.5	2.5	3.0
Data Protection	2.7	3.5	4.0	3.4	2.9	3.0	3.2	3.6	3.5	3.2	3.2	3.4
Management	3.7	3.5	4.2	4.6	3.5	3.9	4.0	4.5	3.9	3.5	3.5	4.9
Stack	4.1	2.9	4.3	4.1	3.0	3.1	4.7	4.2	3.5	3.0	3.5	4.4
Scaling	4.0	3.4	4.5	4.1	3.0	4.0	4.0	4.5	3.5	4.0	4.0	4.1
Service and Support	4.7	3.7	4.5	4.7	4.3	3.2	4.0	4.6	3.7	4.0	3.3	4.8
Security	4.5	3.0	4.5	4.2	3.5	3.0	4.0	4.5	3.5	3.0	3.5	4.0
As of January 2018

Source: Gartner (February 2018)

Table 3 shows the product/service scores for each use case. The scores, which are generated by multiplying the use-case weightings by the product/service ratings, summarize how well the critical capabilities are met for each use case.

Table 3: Product Score in Use Cases

Use Cases	Cisco	DataCore	Dell EMC	HPE	HTBase	Huawei	Microsoft	Nutanix	Pivot3	Scale Computing	Stratoscale	VMware
Consolidated	3.68	3.35	4.25	4.13	3.43	3.50	3.64	4.26	3.61	3.49	3.44	4.12
Business-Critical	3.70	3.39	4.26	4.11	3.40	3.42	3.63	4.24	3.61	3.46	3.42	4.09
Cloud	3.62	3.30	4.19	4.06	3.31	3.36	3.69	4.18	3.50	3.30	3.37	4.13
ROBO	3.72	3.41	4.25	4.16	3.46	3.40	3.66	4.26	3.62	3.45	3.38	4.17
VDI	3.66	3.34	4.22	4.11	3.42	3.46	3.65	4.23	3.55	3.43	3.41	4.15
As of January 2018

Source: Gartner (February 2018)

To determine an overall score for each product/service in the use cases, multiply the ratings in Table 2 by the weightings shown in Table 1.

Critical Capabilities Methodology

This methodology requires analysts to identify the critical capabilities for a class of products or services. Each capability is then weighted in terms of its relative importance for specific product or service use cases. Next, products/services are rated in terms of how well they achieve each of the critical capabilities. A score that summarizes how well they meet the critical capabilities for each use case is then calculated for each product/service.

"Critical capabilities" are attributes that differentiate products/services in a class in terms of their quality and performance. Gartner recommends that users consider the set of critical capabilities as some of the most important criteria for acquisition decisions.

In defining the product/service category for evaluation, the analyst first identifies the leading uses for the products/services in this market. What needs are end-users looking to fulfill, when considering products/services in this market? Use cases should match common client deployment scenarios. These distinct client scenarios define the Use Cases.

The analyst then identifies the critical capabilities. These capabilities are generalized groups of features commonly required by this class of products/services. Each capability is assigned a level of importance in fulfilling that particular need; some sets of features are more important than others, depending on the use case being evaluated.

Each vendor’s product or service is evaluated in terms of how well it delivers each capability, on a five-point scale. These ratings are displayed side-by-side for all vendors, allowing easy comparisons between the different sets of features.

Ratings and summary scores range from 1.0 to 5.0:

1 = Poor or Absent: most or all defined requirements for a capability are not achieved

2 = Fair: some requirements are not achieved

3 = Good: meets requirements

4 = Excellent: meets or exceeds some requirements

5 = Outstanding: significantly exceeds requirements

To determine an overall score for each product in the use cases, the product ratings are multiplied by the weightings to come up with the product score in use cases.

The critical capabilities Gartner has selected do not represent all capabilities for any product; therefore, may not represent those most important for a specific use situation or business objective. Clients should use a critical capabilities analysis as one of several sources of input about a product before making a product/service decision.

© 2018 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered trademark of Gartner, Inc. or its affiliates. This publication may not be reproduced or distributed in any form without Gartner's prior written permission. If you are authorized to access this publication, your use of it is subject to the Gartner Usage Policyposted on gartner.com. The information contained in this publication has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information and shall have no liability for errors, omissions or inadequacies in such information. This publication consists of the opinions of Gartner's research organization and should not be construed as statements of fact. The opinions expressed herein are subject to change without notice. Although Gartner research may include a discussion of related legal issues, Gartner does not provide legal advice or services and its research should not be construed or used as such. Gartner is a public company, and its shareholders may include firms and funds that have financial interests in entities covered in Gartner research. Gartner's Board of Directors may include senior managers of these firms or funds. Gartner research is produced independently by its research organization without input or influence from these firms, funds or their managers. For further information on the independence and integrity of Gartner research, see "Guiding Principles on Independence and Objectivity."