The rapidly growing hyperconverged infrastructure industry is starting to consolidate, with tech giants HPE, Juniper Networks, Cisco and Red Hat all buying promising HCI startups. But there remains a strong group of young, independent companies focused on HCI and surrounding technologies that has attracted big financial investment.
The 10 hot startups selected here have pulled down nearly half a billion dollars in venture capital. They're developing everything from full-stack HCI to scale-out NVMe, and from HCI for containerized environments to memory converged infrastructure. It’s a safe bet that at least a few of them will be coming to a data center near you very soon.
Year founded: 2013
Funding: $35 million
Headquarters: Folsom, Calif.
CEO: Chuck Smith, who spend 20 years at HPE, including a stint as VP and GM of Datacenter and Hybrid Cloud
What they do: Provide scale-out NVMe over Ethernet (NoE).
According to Apeiron Data Systems, the majority of enterprise NVMe storage technologies build scale-out storage to work over fabrics (NVMeoF). These still run up against legacy bottlenecks inherent in SAN environments, however.
Apeiron argues that NVMe over Ethernet (NoE) offers several advantages, including superior performance, linear scale-out, and better TCO because of higher density utilization per rack unit. However, even NoE faces such legacy bottlenecks as storage controllers and external switching hardware needed for routing data among clusters.
Apeiron’s ADS NoE platform is intended to deliver the performance of server-based scale-out storage with the manageability of enterprise-class external pooled storage. Apeiron ADS’ architecture scales processing and storage resources independently. ADS provides elastic management over large pools of NVMe SSDs running under software tools like OpenStack, Docker or Hadoop.
Apeiron claims ADS’ lossless Ethernet architecture can scale to thousands of external NVMe drives that look to the server as direct attached storage. ADS’ data path uses high-speed FPGAs over Ethernet and passes NVMe commands natively across the fabric.
Apeiron argues that native transport is critical to making sure no performance is lost in the pooled environment. By removing all of the IO blocking components and software and passing the data transport in its entirety over an Ethernet tunnel, Apeiron says that it breaks through bottlenecks and maintains performance at scale. The main application Apeiron targets with its storage system is Big Data analytics.
Competitors include: Cohesity, Dell EMC, E8 Storage, Excelero Software, NetApp, Rubrik, Vexata, Zadara and Zstor
Customers include: L3, the U.S. Food and Drug Administration and NEC
Why they’re a hot startup to watch: Smith, the company’s recently landed CEO, knows this industry well, having spent nearly 20 years at HP/HPE.
Founder Lee Harrison moved from CEO to CTO after Smith came aboard. He previously served as Director of Planar Magnetics at Tyco Electronics, and had been CEO of PlanarMag, Before that, he was VP of Operations for KeyEye Communications and GM of the Optical Ethernet Group at Intel.
Year founded: 2014
Funding: $18.3 million
Headquarters: Santa Clara, Calif.
CEO: Zivan Ori, former Chief Architect at Stratoscale
What they do: Provide shared NVMe storage.
E8 Storage argues that latency at the storage layer, rather than the network layer, is often the bottleneck causing major problems for data-hungry applications.
E8’s storage platform consists of NVMe storage controller appliances (the top of the line includes RAID 6 data protection) and storage software that can be installed on any NVMe-capable server.
To boost performance, E8 separates the data path (E8 Agent) and the control path (E8 Controller) operations. The E8 Agent offloads 90% of the data path operations to its host, providing additional compute for each host added to the system, while the E8 Controller ensures availability with RAID 5/6.
This approach differentiates E8 Software from other NVMe and HCI vendors that have not directly addressed the legacy I/O bottleneck that torpedoes shared-storage use cases.
Competitors include: Apeiron Data Systems, Cohesity, Dell EMC, Excelero Software, NetApp, Rubrik, Vexata, Zadara and Zstor
Customers include: Queen Mary University of London
Why they’re a hot startup to watch: E8 Storage has used its $18.3 million in funding to get product to market, and recently announced reseller/channel partnerships with Cognosystems in Canada, ClusterTech in Hong Kong and China, Clustar in China and Datera in the Czech Republic.
E8 also has an impressive founding team, who all gained experience at IBM. Most notably, CEO Ori and CTO Alex Freidman held management positions there focused on storage.
Prior to IBM, Ori also earned industry and exit experience serving as Chief Architect at HCI vendor Stratoscale, heading software development at Envara (acquired by Intel) and serving as VP of R&D at Onigma (acquired by McAfee).
The company has just one named customer, but its channel efforts should start paying off soon, and hopefully it will publicize wins it earns there.
Year founded: 2015
Funding: The startup has not released funding totals, but it closed its Series B round in 2017. The round was led by Rally Ventures with Citrix Systems, Osage Venture Partners and El Dorado Ventures participating. HiveIO’s Series A closed in 2016.
Headquarters: Hoboken, N.J.
CEO: Dan Newton, previously GM and SVP of Rackspace
What they do: Provide HCI services.
HiveIO’s Hyperconverged Fabric (HCF) platform enables users to build a software-defined datacenter on x86 commodity hardware. HCF combines an Intelligent Message Bus with a virtualized computing stack that includes everything from the hypervisor to storage to compute and networking capabilities.
HiveIO’s HCF platform is based on its Hive Fabric technology. At the end of May, HiveIO released version 7.3 of Hive Fabric, an AI-ready fabric solution that enables organizations to deploy vendor-neutral virtualization technology. Hive Fabric combines a KVM hypervisor with software-defined storage (SDS), software-defined networking, a VM broker, orchestration and virtual desktop management.
Hive Fabric consolidates disparate data into a single stream, so organizations can feed it directly into AI software. IT professionals can then quickly collect and analyze data without having to wait for data to be cleaned up, formatted and ingested into third-party software.
Hive Fabric is built with an Intelligent Message Bus and an intuitive UI that provides a real-time, all-encompassing view of the datacenter and its connected components, making it easier for administrators to find and act upon problems to reduce downtime.
The main use case HiveIO focuses on is VDI. Organizations can also deploy Hive Fabric for shared storage, HCI, virtual servers, and PaaS.
Competitors include: Citrix (also an investor), Exago, Izenda and TIBCO
Customers include: JP Morgan Chase, Northrim Bank, CDW, the U.S. Army, Tata and University Hospitals of Leicester
Why they’re a hot startup to watch: HiveIO has a strong leadership team, and a good chunk of HiveIO’s messaging focuses on eliminating multi-vendor complexity, which includes complicated licenses and SLAs. Also, multi-vendor environments invariably create data silos, which is a major problem at scale.
HiveIO not only focuses on eliminating data silos, but also on transforming data into AI-ready streams, enabling data to be ingested by AI tools in real-time, so IT professionals can predict problems and intervene before potential issues impact uptime.
Year founded: 2017
Funding: $24.5 million
Headquarters: San Jose, Calif.
CEO: Charles Fan, former CTO of Cheetah Mobile
What they do: Provide storage-infrastructure services.
According to MemVerge, 2.5 quintillion bytes of data are created every day, much of it machine-generated data from AI, machine learning, IoT and analytics applications. These high-performance workloads require an infrastructure that can process the constant flood of data in real time.
MemVerge argues that this problem forces enterprises to make trade-offs between speed and capacity, and its flagship Memory-Converged Infrastructure (MCI) system attempts to solve this problem.
Designed to run on Intel’s new Optane DC persistent memory technology, MCI delivers memory and storage services from a single distributed platform. This approach eliminates boundaries between memory and storage to better support data-centric and/or real-time enterprise workloads.
MemVerge’s proprietary Distributed Memory Objects (DMO) software provides a logical memory-storage convergence layer that uses Intel’s Optane persistent memory to allow data-intensive workloads – such as AI, machine learning, big data analytics, IoT and data warehousing – to run without errors at memory speed.
MCI expands memory and stores data consistently across multiple systems so enterprises can analyze an enormous amount of data in real time.
Competitors include: Other storage, hybrid cloud and/or HCI companies building on Intel Optane DC persistent memory include Cisco, Dell EMC, Google, Microsoft and VMware.
Beta customers include: LinkedIn, JD.com and Tencen
Why they’re a hot startup to watch: The company just emerged from stealth mode in April 2019 with $24.5 million in Series A funding from Gaorong Capital, Jerusalem Venture Partners, LDV Partners, Lightspeed Venture Partners and Northern Light Venture Capital. Given that, it has an impressive list of named beta customers.
Co-founders, CEO Fan and Chairman Shuki Bruck have the right backgrounds to lead a startup in this space. Fan was an SVP/GM at VMware, where he founded the storage business unit that developed Virtual SAN. He joined EMC via the acquisition of Rainfinity, where he was a co-founder and CTO, and he founded EMC’s China R&D center.
Bruck was the founding director of the Caltech Information Science and Technology (IST) program. He also co-founded and served as Chairman of both XtremIO and Rainfinity.
Year founded: 2013
Funding: $22.7 million
Headquarters: Irvine, Calif.
CEO: Nader Salessi, former VP of Western Digital’s SSD Business Unit
What they do: Develop NVMe and SSD storage.
In March 2019, NGD Systems released its Newport Platform, which provides high-capacity NVMe SSDs that use in-storage processing to eliminate the need to move data to main memory prior to processing. NGD argues that in situ processing removes a major bottleneck that has prevented NVMe from performing in edge and hyperscale environments.
Newport is an ASIC-based computational storage platform that accommodates up to 16 flash channels and delivers NVMe 1.3 PCIe Gen 3.0x4 optimized storage performance.
Competitors include: Western Digital (Tegile), StorCentric (Nexsan), Cohesity, NetApp, Actifio, Qumulo and Veritas
Customers include: None announced.
Why they’re a hot startup to watch: NGD Systems has raised $22.7 million in funding, has a third-generation of its platform in the market as of March 2019 and a unique spin on how to use in-storage processing to extend NVMe storage to intelligent edge and hyperscale use cases.
The startup is also led by an executive team that gained decades of experience with such storage companies such as Western Digital, STEC, Memtech and Micron. Co-founders CEO Salessi, CTO Vladimir Alves and EVP Richard Mataya all came from Western Digital.
Year founded: 2015
Funding: $5 million
Headquarters: Lille, France
CEO: Laurent Denel, previously Head of Mail and Cloud Service Management at Worldline
What they do: Provide hyperconverged storage infrastructure services.
Based on the open-source OpenIO project, which OpenIO’s team developed while at Worldline Global, this startup provides software-defined, scale-out object storage.
As users demand remote access to larger and larger files and as applications increasingly generate more of their own data, latency and bandwidth limitations undermine performance. OpenIO tackles this problem by decentralizing storage alongside distributed applications.
With OpenIO software uses commodity servers as large object-storage and compute pools. By combining storage and integrated data processing on a single platform, OpenIO gives organizations a more flexible and scalable method for building backend application services. Its event-driven, serverless computing framework streamlines data access and simplifies complex workloads by offloading some tasks directly to the storage infrastructure. Use cases include Big Data, video streaming, hybrid-cloud storage, IoT and machine learning.
Competitors include: Scality, Cloudian, Nasuni, Nexenta and StorPool
Customers include: Vade Secure, Internet Initiative Japan, Dailymotion and Reflex Solutions
Why they’re a hot startup to watch: OpenIO proved out the concept for its storage platform while the founding team was still at Worldline Global. The open-source code for OpenIO is available on GitHub, and OpenIO claims that it currently serves up applications to more than 60 million end users.
The startup has more than 25 customers for its service, including several named customers.
Year founded: 2012
Funding: $222.3 million
Headquarters: Seattle, Wash.
CEO: Bill Richter, formerly president of EMC’s Isilon Storage Division
What they do: Provide hybrid-cloud file-system services.