Data-center virtualization is about much more than servers these days. Everything from storage to applications to entire infrastructure stacks are being virtualized by startups seeking to push virtualization throughout the entire IT stack.\nOr, to borrow from Marc Andreessen, virtualization is eating the data center.\n\nThe startups in this roundup are well-funded, with the exception of brand-new ones just out of stealth, but even those have seasoned founders and compelling products under developments. The ventures detailed here represent the future of the data center, offering everything from software-defined servers to software-defined block storage to managed hybrid-cloud services and infrastructure.\nAvi Networks\u00a0\nWhat they do: Provide multi-cloud application services\u00a0\nYear founded: 2012\u00a0\nFunding: $115 million in four rounds from Cisco Investments, DAG Ventures, Greylock Partners, Lightspeed Venture Partners and Menlo Ventures\nHeadquarters: Santa Clara, Calif.\u00a0\nCEO: Amit Pandey, who formerly served as CEO of Zenprise, which was acquired by Citrix in 2013\nProblem they solve: Every application needs application services, such as load balancing, application firewalls, and service meshes. In complex, fluid IT environments, however, delivering those services is a challenge.\nThis is especially true of multi-cloud environments. As enterprises mix traditional architectures with microservices in both private and public clouds, traditional application-delivery methods, such as application delivery controllers (ADC), can\u2019t keep up.\nThe traditional manual approach of requisitioning, installing and configuring ADCs is a process that can take weeks, if not longer. ADCs are hard to automate, must be managed individually and they don\u2019t scale in the cloud.\nHow they solve it: Avi Networks\u2019 flagship product, Vantage, is a multi-cloud application-delivery platform. Vantage consists of three core components: a software load balancer, a web-application firewall (WAF) that relies on machine learning and automation, and what the startup calls an elastic service mesh.\nThe mesh delivers\u00a0multi-cloud application services\u00a0for containerized applications with microservices architectures through dynamic service discovery, application maps and micro-segmentation.\u00a0\nThe software-defined platform separates the control and data planes, unlike traditional appliance-based ADCs. The ability to dispatch services to applications wherever they live without having to worry about the underlying infrastructure gives administrators more flexibility to embrace new application technologies.\u00a0\nCompetitors include: F5, Citrix and NGINX\nCustomers include: Adobe, EBSCO, Paddy Power Betfair, Palo Alto Networks, Swisscom, Swisslos, Travelport and ZOLL Data\u00a0\nWhy they\u2019re a hot startup to watch: Avi Networks is positioned in a high-growth, rapidly innovating market where it\u2019s already carved out a sizable niche. An Avi spokesperson emailed that 20% of the Fortune 50 are Avi customers. With such named customers as Adobe, Palo Alto Networks and Swisscom, the claim sounds credible.\nAvi has raised $115 million in funding from top-tier VCs, and its senior leadership team has a string of successful exits behind them, including Zenprise to Citrix for $335 million, Terracotta to Software AG (terms not disclosed), Andiamo to Cisco for $750M, and Shasta Networks to Nortel for $340 million.\nFinally, if the entire data center is being eaten by software, the underlying infrastructure will need to be as flexible and elastic as the applications it must serve. The Vantage platform looks to be a big leap in that direction.\nExcelero Storage\u00a0\nWhat they do: Provide software-defined distributed block storage\nYear founded: 2014\nFunding: $35 million in two rounds of funding. Investors include Square Peg Capital, Battery Ventures, Micron, Qualcomm Ventures and Western Digital Capital.\nHeadquarters: San Jose, Calif.\u00a0\nCEO: Lior Gal, who previously served as Global VP at Data Direct Networks.\nProblem they solve: Newer Flash media such as NVMe Flash have raised expectations on storage latency. An individual local NVMe drive can deliver performance with single-digit microsecond latency. However, sharing NVMe storage resources across a network degrades latency significantly, something that becomes an issue for web-scale data centers that must scale out their networks and serve demanding high-growth applications, such as machine learning and Big Data.\u00a0\nHow they solve it: Excelero delivers low-latency distributed block storage for web-scale applications. Excelero\u2019s NVMesh enables shared NVMe across any network and supports any local or distributed file system. The software features an intelligent management layer that abstracts underlying hardware with CPU offload, creates logical volumes with redundancy, and provides centralized, intelligent management and monitoring.\nNVMesh\u2019s distributed block layer allows unmodified applications to utilize pooled NVMe storage devices across a network at local speeds and latencies. Distributed NVMe storage resources are pooled with the ability to create arbitrary, dynamic block volumes that can be utilized by any host running the NVMesh block client.\nApplications receive the throughput and input\/output operations per cecond (IOPS) of a local NVMe device, while avoiding proprietary hardware lock-in.\u00a0\nCompetitors include: E8, Apeiron, Dell EMC, E8, HPE, NetApp, Pavilion Data and Pure Storage\nCustomers include: CMA, Technicolor, SciNet and Teuto.net\nWhy they\u2019re a hot startup to watch: Excelero is an early entrant to a fast-growing market. G2M Research predicts that by 2021 the\u00a0NVMe market will expand to $60 billion. Excelero already has several massive deployments, such as the one with SciNet, Canada\u2019s largest semiconductor facility.\nThe senior leadership team has a background that features successful exits to EMC, Dell and F5, and a ton of experience gained at such storage leaders as DataDirect Networks, EMC and Western Digital. Finally, Excelero has raised $35 million in funding, which should provide plenty of runway in this landgrab space.\nHammerspace\nWhat they do: Provide Data-as-a-Service\u00a0\nYear founded: 2018\u00a0\nFunding: Self-funded, amount not disclosed.\nHeadquarters: Los Altos, Calif.\u00a0\nCEO: David Flynn, who previously founded Fusion-io and served as its CEO\nProblem they solve: Data silos limit the ability to create value from data. Data consumers struggle to gain access to the data they need, so IT responds by copying data. Unfortunately, this approach creates sprawl, increasing cost and risk. As enterprises adopt hybrid clouds, this problem becomes more acute and more difficult to solve, since most IT strategies continue to manage data by managing storage.\nIf IT tries to manage hybrid-cloud data via storage technologies \u2013 a strategy pushed by a number of competing vendors \u2013 you end up creating more data silos, while also introducing bottlenecks that continue to restrict your ability to extract value from the data.\nHow they solve it: Hammerspace is focused on enabling the use of cloud services for unstructured data (File\/Object), which makes up the majority of enterprise data and is growing 30 percent year-over-year.\nOrganizations struggle to manage and create value from the unstructured data needed to feed into analytics and AI-type workloads.\u00a0Hammerspace virtualizes data, separating data from metadata while working through standard protocols like NFS, SMB and S3.\nOnce virtualized, machine learning optimizes the flow of data across hybrid clouds to balance cost and performance. Hammerspace provides a global namespace that is available across multi-cloud environments, giving data consumers the ability to run their workloads anywhere.\nHammerspace metadata management features allow for global data visibility and discovery, data catalogs, self-service hybrid-cloud data management and file-level data governance. All of this is achieved using the infrastructure and storage architecture already in place.\nHammerspace allows data operations teams to leverage metadata, self-servicing their hybrid cloud data management to increase their agility, while reducing risk and cost for IT by eliminating manual tasks like copying, load-balancing and downtime remediation.\nCompetitors include: Nasuni, Igneous, Actifio and Panzura\nCustomers include: This startup just launched in 2018 and has yet to name any customers.\nWhy they\u2019re a hot startup to watch: Hammerspace has yet to raise VC funding and name customers, but founder and CEO David Flynn has a relevant, if rocky, background in this sector. He founded Fusion-io and, as its CEO, took it to a successful IPO. SanDisk later acquired Fusion-io for $1.2 billion. However, Flynn\u2019s next startup, Primary Data, flopped, despite raising about $100 million in funding. \u00a0\nFlynn has convinced a good chunk of the Primary Data management team to give it another go, while also retaining some of Primary Data\u2019s virtualization IP.\nGetting into a market too early is risky, a key lesson from Primary Data. But if Hammerspace manages to build on top of Primary Data\u2019s IP to create a viable data-as-a-service business, Flynn will burnish his reputation as a serial entrepreneur. Bad bet or sure thing, this startup is definitely one to watch.\nHyperQube\u00a0\nWhat they do: Server virtualization\u00a0\nYear founded: 2017\nFunding: $550,000 in three rounds of seed\/angel funding from Mach37, CIT, Michael Wellman, Kathryn Stewart and others.\nHeadquarters: Fairfax, Va.\nCEO: Craig Stevenson, who previously served as Director of Cyber Training at Raytheon\nProblem they solve: Sharing and distributing virtual machines is a complex, labor-intensive chore for most IT teams. When they must also handle virtual infrastructure, the problem becomes even more acute.\n\u201cWhile at Raytheon, my division spent over $1 million a year on labor creating virtual infrastructure for customers. We were maxed out at building and delivering 10 products a year,\u201d said CEO Craig Stevenson.\nHow they solve it: HyperQube\u2019s software allows enterprises to automatically distribute and create additional copies of virtual infrastructure. Enterprises no longer have to devote engineering resources building the same environments over and over again.\nHyperQube\u2019s virtualized test environments can be built in minutes and just as quickly modified, re-used and shared. With the ability to look to the future and explore infrastructures in a risk-free environment, engineers save hours of labor, and security officers can test far more extensively before releasing VMs to the wild.\nHyperQube claims to be able to reduce the labor associated with managing VMWare environments by as much as 90 percent.\u00a0 According to Stevenson, if he had something like HyperQube when he was at Raytheon, they \u201ccould have delivered hundreds of products a year,\u201d rather than maxing out at 10.\nCompetitors include: VMware, Hyper-V, Open Stack, Cloud Shell, CirCadence and SimspaceCustomers include: Cisco\nWhy they\u2019re a hot startup to watch: HyperQube\u2019s idea of having software lets you test and deploy virtual infrastructure as easily as calling up an Uber. Plus, it\u2019s nailed down a top-tier named customer in Cisco, which makes it worth paying attention to in order to see if it\u2019s able to streamline and automate virtual infrastructure at scale.\nNetlify\nWhat they do: Provide website-build automation and hosting\nYear founded: 2014\nFunding: $44 million in total. The startup\u2019s most recent round, a $30 million Series B, closed in October 2018. Investors include Kleiner Perkins, Andreessen Horowitz and individual investors.\nHeadquarters: San Francisco, Calif.\nCEO: Mathias Biilmann Christensen, co-founder and CEO, previously founded BitBalloon, an API-driven hosting platform, and served as its CEO. President and co-founder Christian Bach previously founded CAPSIZE, which was acquired by CONNECTED A\/S in 2012. Upon joining CONNECTED, Bach served as Chief Digital Officer.\nProblem they solve: While website backends have experienced a wave of innovation, with everything from virtualization to microservices to SDx reshaping the data center, front ends have lagged behind.\nWhile websites are increasingly designed to be dynamic in order to offer personalized, interactive experiences, the underlying infrastructure to support them is complicated, fragile and often insecure.\nAnother problem with dynamic modern sites is performance. Due to their complicated natures, a simple change or misconfiguration can negatively impact performance.\u00a0\u00a0\nHow they solve it: Netlify believes that the way to alleviate these problems is to create less complicated, automation-friendly static websites with fewer moving parts. The dynamic features can then be layered on top as microservices.\nThe Netlify platform automates code and deployment, systematizing the creation of high-performing sites. The Netlify platform features an automated workflow that combines global deployment, continuous integration and automatic HTTPS.\nDevelopers are able to connect to repositories such as GitHub or Bitbucket, add build settings that can be customized and deploy sites directly from the platform. The platform features multi-cloud redundancy and deployment, built-in CDN, identity management features, automatic HTTPS, DNS management and support for various third-party tools and microservices.\u00a0\nCompetitors include: CloudCAnnon, Heroku, surge and Zeit Now Customers include: Cisco, Docker, Facebook, Google, NBC and Verizon\nWhy they\u2019re a hot startup to watch: When it comes to startup fundamentals, Netlify is strong across the board: experienced leadership team, hefty funding from top-tier VCs, and a long list of named customers, including the likes of Cisco, Google and Verizon.\nToday\u2019s websites break, degrade and fail to block attackers all the time. An enterprise-class, systematized, automated approach to building, maintaining, updating, deploying and securing websites is long overdue.\nPlatform9\nWhat they do: Provide a managed hybrid-cloud service that turns existing enterprise infrastructure into a cloud.\nYear founded: 2013\nFunding: $36.5 million raised in three rounds of funding. Investors include Redpoint Ventures, Menlo Ventures and Canvas Ventures.\nHeadquarters: Sunnyvale, Calif.\nCEO: Sirish Raghuram. Before co-founding Platform9, Raghuram was an early engineer at VMware, where he held technical and management leadership roles and helped ship multiple vSphere products.\nProblem they solve: According to Gartner, 95 percent of enterprise IT teams struggle or fail when implementing private clouds. In addition, the typical team must manage an average of 4.6 environments.\nPlatform9 argues that typical large organizations run \u201chyper-hybrid, multi-cloud environments,\u201d where they must support many different types of applications and architectures all at once. Some of those are on private clouds, some on public clouds and some run on bare metal.\u00a0\nEach type of application requires its own technology stack and different types of infrastructure, with different environments built and maintained in fundamentally different ways. For instance, enterprises manage their on-premises infrastructure, like VMware, much differently from the way they consume their infrastructure on AWS or manage container environments.\nThis situation creates challenges for operations teams, which struggle just to keep things running, let alone to scale them up. As the pressure to move faster while reducing costs, increases, the operational challenges will continue to mount.\nHow they solve it: Platform9 Managed OpenStack service is intended to make it easy for enterprises to create an AWS-like private cloud, using their own on-premises virtualized infrastructure. Enterprises can pool resources in a single data center or across multiple regions.\nPlatform9 hosts the OpenStack control plane and handles installation, monitoring, troubleshooting and upgrades.\nOther features of the Platform9 cloud include cloud visibility and control tools, reusable automation frameworks for various OpenStack offerings, multi-hypervisor management capabilities, SSO integration, persistent storage, isolated networking and image management.\nPlatform9 recently released a new service, Platform9 Managed Kubernetes, which, as the name indicates, is a managed Kubernetes service. Managed Kubernetes also handles backend configuration and maintenance.\nCompetitors include: Red Hat, AWS, Mirantis, StratosCAle and ZeroStackCustomers include: AutoDesk, LogMeIn and Veritas\nWhy they\u2019re a hot startup to watch: Platform9 has continued to raise VC money for several years, gained new customers and refined its products and market positioning. The startup\u2019s leadership team is filled with VMWare alums, which has helped them compete against deep-pocketed incumbents.\u00a0\nSea Street\nWhat they do: Provide an AI-based autonomous-operations platform\nYear founded: 2012\nFunding: $41 million raised in two rounds of funding, investors not disclosed.\nHeadquarters: Wakefield, Mass.\nCEO: Harley Stowell, who previously was Cisco\u2019s CTO, Worldwide Service Provider Cloud & Managed Services\nProblem they solve: Delivering enterprise application services and the cloud infrastructure that supports them is a labor-intensive, error-prone, difficult-to-scale task.\nSea Street argues that in order to scale to meet end-user demands, all applications and all IT services will eventually need to be rendered as services operating in converged virtual infrastructure.\nHow they solve it: Sea Street\u2019s AI-based autonomous operations platform, StratOS, enables enterprises and service providers to deliver services as fully autonomous, closed-loop operations.\nStratOS manages both applications and multi-cloud infrastructure based on policy, and it continuously creates, updates and programs system architectures to meet the needs of the applications and the policies.\nStratOS is built on designable micro-intelligences that the startup calls Objectives. These are goal-seeking, intelligent software objects. They interoperate and cooperate to create autonomous services, compound services and even division-level service operations. \u00a0\nSea Street describes Objectives as being like bees in a hive: a set of Objectives runs a service, while a swarm of Objectives runs an enterprise. They enable organizations to automate the enterprise element by element, a service at a time.\nCompetitors include: BMC, itential, SaltStack and ChefCustomers include: Cox Communications\nWhy they\u2019re a hot startup to watch: CEO Harley Stowell\u2019s background at Cisco is a big plus, as is the fact that he joined Cisco by selling his previous startup, LineSider Technologies, to them. Plus Sea Street has $41M in funding and a major customer in Cox Communication..\nStratoscale\nWhat they do: Provide AWS-compatible, automated hybrid cloud infrastructures delivered as a\u00a0 service\nYear founded: 2013\nFunding: $70 million-plus. Investors include Battery Ventures, Bessemer Venture Partners, Cisco, Intel, Qualcomm Ventures, SanDisk and Leslie Ventures.\nHeadquarters: Herzliya, Israel\nCEO: Ariel Maislos, who previously founded Anobit, an SSD startup that was acquired by Apple in 2012. At Apple, he served as a Senior Director in charge of Flash Storage.\nProblem they solve: As cloud adoption surges and developers increasingly turn to managed cloud services, they struggle to deal with legacy apps that don\u2019t fit with the new model. Whether it\u2019s due to regulatory issues or data ownership issues, many enterprises will never fully embrace the public cloud, meaning hybrid models are here to stay.\nHowever, hybrid models often mean certain applications (and thus end users) receive less attention and resources, resulting in an inferior experience.\nHow they solve it: Stratoscale\u2019s cloud service transforms on-premises infrastructure into self-service, on-demand resources with built-in multi-tenancy.\nBy delivering on-premises versions of popular cloud services, including EC2, EBS, VPC, RDS, ELB and EMR, developers and IT professionals are able to leverage all of the advantages of the public cloud, without compromising control, governance and security.\nStratoscale automatically manages all aspects of open-source services, including Kubernetes and various database engines, so the burden of provisioning, monitoring, scaling and upgrading open-source services is eliminated.\nStratoscale also provides AWS-compatible APIs, enabling customers to easily implement consistent and infrastructure-agnostic automation toolchains across the entire organization.\nCompetitors include: AWS, Microsoft Azure, Red Hat, Pivotal and Platform9Customers include: Arlington Orthopedics, Be Takeoff, InCArta, SparkinIT and Skyberate\nWhy they\u2019re a hot startup to watch: Stratoscale has raised a significant amount of VC funding, is positioned in a competitive, expanding market \u00a0and has a decent roster of named customers.\nCEO Maislos has founded several startups and has successful exits to PMC-Sierra and Apple. Ronen Kofman, VP of Product and Engineering, has held leadership roles at Mirantis, Oracle and Intel, while CFO Anat Zach joined Stratoscale from Waze, where she managed the due diligence process for the company\u2019s $1 billion acquisition by Google.\nTidalScale\nWhat they do: Provide software-defined servers\nYear founded: 2013\nFunding: $41.8M. In October 2018, the startup announced its $24M Series B. Investors include Bain Capital Ventures, Hummer Winblad, Sapphire Ventures, Infosys and SK Hynix. Update: On December 3, TidalScale announced that it has extended its Series B, adding Forte Ventures and undisclosed investors. This brings its Series B total to $27M.\nHeadquarters: Campbell, Calif.\nCEO: Gary Smerdon, who formerly served as EVP, Chief Strategy and Product Officer for Fusion-io\nProblem they solve: Organizations increasingly face workloads that overwhelm even their largest servers. In the modern data center, storage, networking and other assets are designed to expand and grow as workloads change over time. But servers are fixed resources.\nThat's a problem when unexpected spikes in transactions, seasonal fluctuations in business activity, rapid growth in data size and complexity and other factors overwhelm typical servers. Thus, many enterprises install excess capacity based on the projected future requirements, often vastly overpaying along the way.\nHow they solve it: TidalScale argues that what the modern data-center needs is the ability to right-size a server on the fly to fit any data set. TidalScale\u2019s software-defined server technology is designed to deliver that flexibility.\nTidalScale allows customers to flexibly allocate existing low-cost infrastructure, so they always have enough computing power to fit their data sets. TidalScale pools or aggregates the compute, memory and I\/O resources of multiple physical servers to create SD-servers using what it calls HyperKernel technology. HyperKernel software binds multiple physical computers into a single, coherent virtual system that enables running an unmodified guest operating system, without making any changes to existing applications.\nHyperKernel is built on TidalScale\u2019s inverse-hypervisor technology. Unlike traditional virtualization technology, TidalScale\u2019s inverse hypervisor works between the guest operating system and bare metal hardware, acting as a hardware-level interface to the guest OS.\nBut unlike traditional virtualization, where multiple virtual machines run on a single physical server, TidalScale provides a single virtual machine across multiple physical servers. Each SD-server appears as a single system, even though it may be comprised of dozens of individual servers. TidalSCAle accomplishes this by joining the physical machines via a standard 10Gb Ethernet network that acts as the resource bus, enabling the server pool to behave like a single large server.\nOnce an SD-Server is up and running, TidalScale\u2019s real-time machine-learning layer continuously optimizes system performance.\nCompetitors include: Datacore, VMWare and Red HatCustomers include: NCS Analytics\nWhy they\u2019re a hot startup to watch: TidalScale has raised a respectable amount of funding, has a named customer and has a strong management team.\nPresident & CEO Gary Smerdon played a leadership role in Fusion-io\u2019s $1.3 billion acquisition by SanDisk. Smerdon has also served in leadership roles for LSI, Greenfield Networks (acquired by Cisco), Tarari (acquired by LSI), Marvell and AMD.\nOther members of the senior team have served in executive roles for SAP, Apple, Cisco, Lotus and HP.\nFinally, capacity planning is a headache for most enterprises, and most of them end up paying for more capacity than they need until something happens, and they need even more capacity than they were overpaying for. TidalScale could help solve that conundrum.\nVexata\nWhat they do: Provide NVMe Flash storage systems and data management software for hybrid cloud environments\nYear founded: 2014\nFunding: $54 million. Investors include Lightspeed Ventures, Intel CApital, Mayfield and Redline Capital.\nHeadquarters: San Jose, Calif.\n\u00a0CEO: Zahid Hussain, who was most recently SVP and GM of Flash Products at EMC\nProblem they solve: Database, analytics and machine learning environments require large volumes of structured and unstructured data to be processed in real-time, but performance and I\/O bottlenecks prevent these workloads from getting the horsepower they need. In almost all cases, Vexata contends, the bottleneck is a result of the storage-controller architecture, especially in all-flash arrays. \u00a0\nAt the same time, as multi-core compute performance increases, the number of virtual machines supported by hypervisors continues to grow, as well. This can translate to hundreds of VMs per data-center rack.\nThis type of traffic load can easily overwhelm all-flash storage infrastructures, because a single multi-core server can drive more throughput than most all-flash arrays (AFAs) can support. This forces architects to distribute, load and buy more AFAs than needed to scale performance.\nStorage bottlenecks throttle system performance and slow response times, directly impacting the customer experience.\n\u00a0How they solve it: Vexata\u2019s storage systems remove latency from the controller by separating the fast control plane from the slow data plane through its Vexata Operating System (VX-OS) software. Separating the data from control planes is a best practice that has been implemented in high-performance networking products for years.\u00a0\nBy removing latency from the controller architecture, users get the full performance benefits from NVMe and storage-class memory solid-state media, which means AI, machine learning, real-time analytics, and other throttled workloads are now able to scale. \u00a0\nThe VX-Manager component provides visibility, management and analytics for the applications running on the system, as well as intelligent troubleshooting and root-cause analysis.\u00a0\nCompetitors include: Dell EMC, Pure Storage, HPE and NetAppCustomers include: Oath, Tata Consultancy, Pacific Data Center and Sanmina\nWhy they\u2019re a hot startup to watch: Vexata emerged from stealth mode in 2017, backed by $54 million in VC funding and touting named customers with complex use cases, including Oath and Tata Consultancy.\nIn the past year, the startup launched a VMware-optimized version of its product and also attracted former Brocade and EMC executives to an already strong leadership team. CEO Zahid Hussain led the EMC team that acquired XtremIO and also headed the mid-range product line. Before that he was head of engineering at VMware ESX and VP of Engineering at Brocade. CTO Surya Varanasi led the ViPR engineering team at EMC.