Exploring HCI

Runninghead:EXPLORINGHYPER-CONVERGEDINFRASTRUCTURE 1
Exploring Hyper-Converged Infrastructure
David Han
ITEC 610
Professor Abdollah Abtahi
November 6, 2015

Abstract
Over the last 10 years the emergence virtualized server technology has been adopted at a
rapid pace. Consolidating many server functions into fewer hardware platforms can reduce
hardware maintenance costs, the need for floor space expansion, centralized management, rapid
new system deployments, increased uptime, and ease of system backups and implementing
disaster recovery sites. However, as legacy equipment and technology was first used to deploy
server virtualization, those hardware limitations are increasingly affecting system performance as
demand increases. To address this issue, hyper-converged infrastructure solutions for virtual
environments have been developed over the last five years by combining storage arrays with the
hypervisor and networking into a single platform for efficiency and performance increases. This
research paper will discuss how the combination of computing, storage, and networking into
hyper-converged infrastructure solutions will continue to advance server virtualization
technology and define the term software defined data centers (SDDC). Different solution
providers will be introduced and compared, and how those solutions benefit virtualized
environments. Advantages and disadvantages of converged systems is discussed and compared
to hyper-converged infrastructure solutions. Future trends of hyper-convergence solutions for
virtualizing SAN and data centers will be introduced and discussed. The goal of this paper is
introducing new hyper-converged technology and how the solutions will be the foundations of
future data centers.

Contents
Abstract 2
Table of Contents 3
Introduction 4
Converged vs. Hyper-Converged 5
Advantages and Disadvantages 8
Hyper-Converged Infrastructure Vendors 12
Future Trends 16
Discussions 17
Conclusion 18
References 20

Introduction
While hardware virtualization can trace its roots back to the 1960's, it was the release of ESX
Server, a dedicated operating systems (OSs) for hosting virtual machines called a type 1
hypervisor, by VMware in 2001 that led to mainstream adoption of virtualized server technology
(Conroy, 2011). Eventually ESX was joined by other type 1 hypervisors from other vendors,
such as Citrix, KVM, and Microsoft Hyper-V. Traditionally hypervisors are installed on existing
stand-alone server hardware or into server blades with traditional server OSs deployed as virtual
servers within each host. Virtual machines (VMs) are either stored on local drives, which limits
features such as migrating VM’s to another data store, or on shared enterprise storage solutions,
such as NAS or SAN. Physical connections between the computing, storage, and network
components are still required, but the logical connections to the VMs are managed and
configured through a virtual switch. As each element of the VM is managed and stored by
separate systems and resources are increased to meet growing demands, the footprints of the
individual components grow proportionally. Each segment of computing, network, and storage is
managed and configured individually, requiring subject matter experts in their respective fields.
With the recent emergence to dynamically provision virtualized desktops, the performance
bottleneck of traditional SAN and NAS storage became an issue. To address the shortfalls of
these separate systems, increase performance and efficiency, the latest trend is combining the
computing, storage, and networking into a single centrally managed hardware platform through
commodity computing and distributed storage. Hyper-converged infrastructures (HCI) are the
latest efforts to simplify the entire computing, network, and storage stacks into the simplest form
for ease of deployment, upgrade, and management and eliminates legacy infrastructure silos. The
future trend for private and public cloud implementations is HCI.

Converged vs. Hyper-Converged
A converged solution is a pre-integrated platform combining the computing, storage, and
networking into a single rack, or block. While it has elements similar to traditional VM
infrastructure as there are physical connected to each other, components of converged solutions
are optimized and preconfigured to operate dependently. Purchasing additional blocks to increase
resources and capacity require minimal configuration updates as the entire stack is managed
centrally. Unlike hyper-converged infrastructure solutions, components from the block can be
configured to operate independently from each other (Lawton, 2014). A subsidy of EMC called
Virtual Computing Environment, or VCE, is the industry leader of converged solutions with
Vblock, which integrates Cisco servers and networking with EMC storage and VMware
virtualization software (VCE, 2015). Vblocks utilize Cisco Unified Computer Systems (UCS)
integrating computing nodes, switching fabric, and Cisco UCS Manager management software
into a single platform to host hypervisors. Cisco USC Manager centrally manages all
configurations, such as MAC address, firmware, and BIOS for the computing nodes. Enterprise-
level Vblocks include EMC applications Avamar, Data Domain, and RecoverPoint for data
backup, deduplication, and recovery. Cisco Nexus modular switching equipment to EMC storage
are pre-configured and tested for optimal performance. The traditional market for Vblocks has
been enterprise level datacenters with the starting price tag of $650,000. VCE has recently,
however, started offering lower cost options for smaller data centers at $100,000 (Babcock,
2013). Below Figure 1 is an example of a VDI implementation of VCE Vblock capable of
supporting 7000 users.

EMC’s VSPEX is another option for organizations to migrate to a converged solution with the
goal of implementing a private cloud while retaining existing hardware. EMC provides a
comprehensive hardware compatibility list that has been pre-tested for meeting performance
benchmarks called reference architecture. Organizations have the flexibility to choose either
VMware View or Citrix XenDesktop for virtual desktop solutions and VMware or Microsoft
Hyper-V private cloud solutions. While this solution retains the central management capability
of a converged solution and advantage of independent hardware selection over VCE Vblocks,
organizations would still need support contracts from separate vendors.
The term hyper-converged is derived from the combining of computing, networking, and storage
into a single form-factor. Unlike rack-sized converged solutions, hyper-converged infrastructure
solutions are physically smaller-sized appliances and measured in rack units. Each hyper-
converged node resembles a single rack mounted server. Management software is deployed as a
Figure 1 (Image courtesyof jasongaudreau.com)

virtual machine which can centrally manage one to multiple appliances supporting industry
standard hypervisors such as ESXi, XenServer, Hyper-V, or KVM. Hyper-converged solution
implements the concept of commodity computing by utilizing a central management console to
integrate inexpensive hardware components in parallel to deliver high-performance and high-
availability virtual environments. The environment is seen by the management software as a
single pool of computing and storage resources. As additional nodes are deployed to increase
computing cores, memory, or storage, the resources are dynamically added into the resource
pool. Because hyper-converged infrastructure solutions operate from a single node, a data center
is not required and therefore is ideal for small or medium-sized businesses. Below Figure 2 is a
visual representation of how servers, switches, and storage are virtualized into a single hyper-
converged hardware platform by Nutanix.
Figure 2 (Image courtesyof nutanix.com)

Advantages and Disadvantages
Both converged and hyper-converged infrastructure solutions offer advantage of reduced
hardware footprint in the datacenter which reduces power consumption, cooling costs, and other
environmental requirements. The modular design of both solutions allow for scalability that is
easily integrated for additional resources. Both converged and hyper-converged infrastructure
solutions are designed with substantial virtual infrastructure performance increases over
traditional non-converged solutions through optimized transmission architecture from computing
to storage and network components. Known as single pane of glass, configuration, management,
and upgrades are greatly simplified through single a management software application
eliminating configuration requirements on the individual server, network, and storage
components. No longer are configurations maintained on individual hardware components, but
can now be dynamically assigned by engineers through a single console.
One of the greatest advantages of a converged hardware solution is lower support and
maintenance costs. In a traditional data center, hardware issues are complicated to troubleshoot
since there are different vendors for computing, storage, and networking. Many different subject
matter experts are required to properly support the variety of hardware types from various
manufacturers. While converged infrastructure is similar in components, since the hardware is
single-sourced, troubleshooting is simplified to only contacting one vendor. This reduces the
training and familiarization requirements for the IT staff as the components are centrally
managed and updated. Another advantage is the integration of the computing, storage, and
network hardware into a single management console. Engineers would not have to go the
hypervisor management page to deploy a VM, then go to the sever chassis management page
such as HP iLO or Dell DRAC to check on server health, and then use a telnet session to SSH

into network switches to make configuration changes. Since all the hardware is already
preinstalled, prewired, and tested in the racks, deployment time is considerable shorter than
having to install individual components in existing data center racks, cable and then test the
systems.
Hyper-converged infrastructure is designed to be the easiest to deploy and implement. The
greatest advantage of hyper-converged infrastructure is the rapid deployment ability of the
solution. Implementation time is usually measured in hours instead of days or weeks in
traditional virtual environment deployments and requires a smaller engineer team as the solution
is implemented in individual nodes. Another advantage is lower cost compared to purchasing
additional enterprise storage or converged infrastructure blocks since a specific requirement for
resources of computing and storage can be scaled in nodes. Performance advantages to the
solution are built-in data deduplication, compression, and snapshot backup capabilities. Data
deduplication reduces wasted data storage space by compressing and removing duplicate data
blocks. Distributed storage is another key advantage of HCI as the storage across all the nodes
can be configured as one logical pool. This provides a robust fail safe system against loss as
multiple disk or entire node failures would not affect data integrity. Another advantage of
distributed storage is performance in terms of input/output as the bandwidth is spread across all
the nodes in parallel read/write operations similar to IBM’s General Parallel File System (GPFS)
or open source Lustre file systems for high performance computing. Unlike traditional SAN
configurations where spindle disks and solid state drives (SSDs) are specifically allocated to
separate storage pools, hyper-converged storage can utilize the performance advantages in a
hybrid solution across the entire infrastructure. By allowing options of mixing solid states disks
with hard drives, organizations can save money by utilizing more spindle drives and upgrade to

more expensive SSDs with future upgrades. Since the core of a hyper-converged solution is the
management software, as the nodes become obsolete, new hardware nodes are easily
incorporated without having to reconfigure the entire system. Most hyper-converged
infrastructure solutions eliminate the need for traditional RAID storage configurations and the
overhead associated with it as well as dynamically configure the storage pool size as needed.
Unlike a traditional storage area network (SAN) block level data store which requires
virtualization software such as VMware vStorage File System (VMFS) to allow sharing of a
logical unit numbers (LUNs) by multiple virtual servers, clustered file systems, also known as
distributed filesystems in a hyper-converged infrastructure solution, are able to utilized network
file system (NFS) file level data stores without decreased performance from network chatter.
This is accomplished by virtualizing the NFS controller, eliminating the physical connections
from NFS storage to server (Nutanix, n.d.). By also utilizing flash array hardware, the highest
performance in input/output per second (IOPS) is also achieved. Virtual desktop infrastructures
(VDIs) and databases are two technology areas that benefit the most from the performance
advantages of hyper-converged infrastructure file systems. In figure 3 below, Nutanix
Distributed File System illustrates how a hybrid of solid state and hard disk drives are utilized in
a clustered storage pool. Nutanix requires a minimum of three physical nodes are required to
form a cluster.

The major disadvantage of converged solutions is being dependent on a single vendor for the
entire computing, storage, and network hardware and inability to replace any individual
component from another vendor. Because systems are preconfigured and tested, there are limited
configuration changes that can be made compared to a customized solution from various
vendors. Finally, there is a higher cost to converged solution compared to HCI since it is a total
system that is purchased together. As converged infrastructures are rack sized solutions, they
require a larger physical footprint in a data center than HCIs, requiring greater power and cooling
requirement than HCI. Upgrades require purchasing another full block or blocks instead of
having the flexibility to upgrade individual components as nodes.
Figure 3 (Image courtesyof www.netwatch.me)

While there are fewer disadvantages with hyper-converged infrastructure solutions, since it
offers greater flexibility and scalability compared to converged infrastructure, it also has the
same configuration limitations as granular upgrades or modifications are not possible. The ability
to upgrade individual resources is also limited since computing and storage are integrated into
one hardware chassis. However, many vendors are addressing this issue by offering storage
nodes to increase data capacity only. While data deduplication and inline compression are major
features of hyper-converged infrastructure solutions, performance is also reduced when the
process is executing as additional computing and memory resources are consumed.
Hyper-Converged Infrastructure Solution Vendors
According to market analyst firm IDC, hyper-converged infrastructure is one of the fastest
growing and changing emerging markets (Feng, Permenter, Scaramella, 2014). The core strength
of a vendor's hyper-converged product is the solution software's ability to integrate computing,
storage, and networking hardware into a single seamless architecture that is easily scalable.
Nutanix is currently the industry leader in hyper-converged infrastructure that supports VMware
ESXi, Hyper-V, and open source KVM. Their main product is Virtual Computing Platform
(VCP) and implements a propriety file system called Nutanix Distributed Filesystem (NDFS).
NDFS allows the appliance to see all the disks as a single storage resource pool discarding the
traditional storage RAID configurations and associated limitations. Depending on the total
number of nodes in server, multiple drives and entire nodes can fail without affecting data loss
greatly increasing reliability and redundancy of the system. However, data deduplication is only
available with solid state disks. Virtual hosts outside the Nutanix storage clusters can also be
integrated and managed through NFS and SMB3 file systems allowing for seamless VM

migration into the HCI environment and replicated to Amazon or Azure public clouds. Upgrades
to vSphere and Hyper-V hypervisors and firmware upgrades to all the nodes can be
accomplished with one click from the management console. Google is currently the highest
profile customer utilizing Nutanix in their data centers (Maheshwari, 2014). The major
disadvantage of Nutanix is not offering a storage-only node. Industry average shows storage out
growing compute at the rate of 5 to 6 times. Since each Nutanix node requires a licensed
hypervisor, unless organizations migrate to open source KVM hypervisor, cost of additional
Nutanix nodes grows exponentially more expensive. A representation of HCI single pane of
glass management console is shown in figure 4 and is a screenshot of a Nutanix dashboard
display of storage and server health.
In addition to OmniCube HCI appliances, SimpliVity also offers Omnistack Accelerator cards
installed on legacy onboard PCI Express slots to manage data deduplication and compression
Figure 4 (Image courtesyof nutanix.com)

while maintaining RAID drive configurations. This allows data centers migrate to hyper-
converged infrastructure at the same time integrating existing legacy infrastructure. VM hosts on
non-HCI are still manageable through NFS shares. OmniCubes can be geographically dispersed
since only differential data is replicated between sites enabling a great level of data reliability as
well as replicating to Amazon public cloud. However, as legacy hardware does not change, the
configuration and upgrade of the vSphere hypervisor is still a manual process.
Gridstore is a Hyper-V based appliance is an all-flash solution engineered for Windows and only
supports VMware through iSCSI. All-flash drives provide consistent performance over hybrid
solutions that utilize flash drive for caching SATA drives for data storage. The advantage of
Gridstore over Nutanix is being able to run without a hypervisor on the appliance. Management
software is installed on any hardware platform and then integrated into the storage pool.
Gridstore also provides greater flexibility than Nutanix by offering dedicated storage nodes that
can be added to increase data storage capacity. Unlike Nutanix which relies on performance by
combining compute and storage in the same hardware chassis, Gridstore implements grid storage
configuration allowing for higher performance through parallel I/O and all flash storage arrays
that can be incorporated as dedicated storage nodes. However, only organizations utilizing
Hyper-V hypervisor and Azure cloud are the advantages of Gridstore realized as system backups
and disaster recovery are replicated to Azure cloud providers only (Murphy, 2015).
Nimboxx utilizes the open source KVM hypervisor freeing organizations from paying license
fees to VMware, Citrix, or Microsoft and can be deployed as a single node and scale up to
hundreds of nodes. For organizations starting to implement virtual machines, Nimboxx offers the
greatest level of flexibility as it will utilize any ratio of solid state and hard disk drives, dynamic

storage configuration, load balancing, data deduplication and thin provisioning of storage space,
backups through snapshots and replication, and built-in data encryption through single pane of
glass management software.
EVO:RAIL is software hyper-converged solution by VMware and partnered with many
hardware providers including Dell, EMC, Fujitsu, HP, Hitachi, NetApp, and Super Micro
Computer (Evans, 2015). It utilizes VMware Virtual SAN (VSAN) to provide distributed storage
management as well as quick deployment. The major disadvantage of EVO:RAIL is the only
supported hypervisor is VMware, and as VSAN is configured at the vSphere kernel level, it
cannot connect to virtual machines outside of the VSAN cluster. Other limitations of EVO:RAIL
include not providing data deduplication nor compression, data encryption, and public cloud
replication. Backups are either done through VMware’s Virtual Data Protection and Recovery
(VDP) add-on module to vSphere or third-party solutions such as EMC Avamar.
While not a true hyper-converged infrastructure solution, DataCore’s SANsymphony-V10 is a
direct competitor to VMware’s Virtual SAN. SANsymphony is a storage virtualization solution
that allows for the integration of flash storage arrays to an organization’s existing storage
infrastructure for scalability of up to 32 nodes, 32 petabytes of data storage, and up to 50 million
IOPS. In comparison, VSAN can only support 4.4 petabytes and up to 2 million IOPS with 32
nodes. The caching of data into working memory (DRAM) for buffering instead of solid state or
hard disk drives is key to SANsymphony’s high performance capability as well as split an
application on the block level and move higher priority blocks into flash storage, lower priority
blocks to hard disk, and archiving infrequently used blocks to public cloud storage .

SANsymphony is hypervisor agnostic as it is able to support Windows virtual machines on
VMware vSphere, Citrix XenServer, and Microsoft Hyper-V (NT4Admins.de, 2014).
Windows Storage Server 2012 is a Microsoft customized operating system for storage platforms.
It is an OEM-only operating system provided to manufacturers for integration into appliances for
centralized storage management supporting Hyper-V hypervisors with features such as data
deduplication and clustered-storage failovers through NAS, iSCSI, and server message block
(SMB) protocols. It is purposed designed for Hyper-V environments migrating to the cloud.
Future Trends
Storage virtualization is the future industry trend of hyper-converged infrastructure. Of the three
major components of computing, storage, and network in a virtualized environment, storage is
the most critical element as storage performance has the greatest effect on virtual machines. This
is evident by the offerings of Windows Storage Sever 2012 by Microsoft on storage platforms to
integrate with Windows Server 2012 Hyper-V, VMware Virtual SAN with vSphere, and
DataCore’s SANsymphony to virtualize storage through software clustering and simplify data
allocation through a central management console. The technology of clustered storage allows for
dynamic allocation of storage resources legacy SAN and NAS solutions cannot offer including
the ability integrate flash and hard drives dynamically to a storage pool through a single
management console. Clustered storage, also known as grid storage and distributed file systems,
offer critical performance increases to workstations provisioned through virtual desktop
infrastructure and databases deployed as virtual machines.
Both converged and hyper-converged infrastructures are making Software Defined Data Centers
(SDDC) possible because both solutions are able to virtualize servers, storage, and network

through a central management console, or single pane of glass management application. In
addition to the traditional offerings of IaaS, SaaS, and PaaS associated with cloud computing,
SDDC offers the next level of service as IT as a service (ITaaS). ITaaS has the implication of
being able to encompass private, hybrid, and public clouds.
On October 12, 2015, Dell announced plans to acquire EMC for $67 billion. This has major
implications for VCE as Cisco is an integral partner with Dell as a direct competitor of server
and network hardware. This future trend is a great concern to customers with continuing support
for existing VCE Vblocks integrated with Cisco UCS servers and Cisco switches as Dell plans to
integrate their own solutions into future Vblocks (Haranas, 2015).
The greatest barrier for organization from migrating into to the cloud is legacy infrastructure.
The future trend to cloud migration is eventual replacement of existing legacy siloed hardware
with software managed converged and hyper-converged computing, storage, and networking to
enable organizations to migrate seamlessly into the private and hybrid clouds.
Discussions
Organizations must consider the current state of their IT infrastructure and the lifecycle of the
hardware and incorporate the latest technology when planning for upgrades. While converged
and hyper-converged software solutions can be used to integrate existing infrastructure,
compatibility of the hardware and managing various vendors must be considered versus
complete hardware replacement. Analysis of total cost of ownership (TCO) must be analyzed
when considering upgrading data storage as support for SAN and NAS exponentially increases
after five years. Organizations must choose between complete storage infrastructure replacement,
known as forklift upgrades, or implement a new solution. The labor resources in terms of

expertise and training required to manage the various components in the data center must also be
considered. Traditionally, IT engineers must access different management consoles for servers,
storage, and network to make configuration changes and perform updates. Both converged and
hyper-converged infrastructure solutions are able offer single glass pane management of the data
center infrastructure integrating all the tasks of server deployment, network and storage
configurations, and hardware patching. The greatest advantage of converged and hyper-
converged infrastructure solutions is also the greatest concern to an organization as they are
entrusting their entire data center infrastructure to a single vendor's longevity.
Conclusion
Hyper-converged infrastructures are the foundation of future data centers. Converged solutions,
while centrally managed, are still using legacy hardware and require similar power and cooling
requirements. Converged infrastructures are cumbersome as they must be deployed by entire
blocks and are inflexible since each component in the block is static and cannot be replaced.
Hyper-converged infrastructure offers the greatest level of scalability because computing,
storage, and network are deployed as individual nodes, and thus fewer data center resources in
terms of power consumption or cooling are required. Nodes can also be deployed to other
geographical sites while remaining in the same resource pool for increase data resiliency over
legacy infrastructure. Utilizing commodity clusters and distributed storage with hyper-converged
infrastructure offers the most flexible and highest performing solution over legacy systems.
Hyper-converged infrastructure solutions simplify the managing of a complex data center
environment by virtualizing servers, storage, and networking through a single management
console allowing organizations to effectively streamline their engineering staff and reduce
training requirements. Troubleshooting is greatly simplified as the entire infrastructure stack is

supplied by a single vendor. When organizations analyze the total cost of ownership of
maintaining legacy silo components of computing, storage, and network, hyper-converged
infrastructure solutions offer the best return on investment. Hyper-converged infrastructure is an
organization’s answer to the highest performance, most reliable, and simplest solution possible
for their virtualized environment.

References
Conroy, S. (2011, August 01). History of Virtualization. Retrieve from
http://www.everythingvm.com/content/history-virtualization
Evans, C. (2015, March). Hyper-converged infrastructure systems deliver storage in a bundle.
Retrieved from http://searchvirtualstorage.techtarget.com/feature/Hyper-converged-
infrastructure-systems-deliver-storage-in-a-bundle
VCE. (2015). History. Retrieved from http://www.vce.com/about/company/history
Lawton, S. (2014, October 16). Converged vs. hyper-converged infrastructure solutions.
Retrieved from http://www.tomsitpro.com/articles/converged-hyper-converged-i nfrastructure-
solutions,2-817.html
Babcock, C. (2013, February 02). VCE adds smaller vblocks to attract SMBs. Retreived from
http://www.informationweek.com/cloud/infrastructure-as-a-service/vce-adds-smaller- vblocks-
to-attract-smbs/d/d-id/1108786?
Sheppard, E., Feng, I., Permenter, K. & Scaramella, J. (2014, December). IDC Marketscape:
Worldwide Hyperconverged ssytems 2014 vendor assessment. Retrieve from
http://www.idc.com/getdoc.jsp?containerId=253267
Nutanix. (2015). Nutanix liberates NFS from network – reincarnating it as NDFS. Retrieved
from http://www.nutanix.com/press-releases/2012/06/12/nutanix-liberates-nfs-from-the-network-
reincarnating-it-as-ndfs/
Kovar, J. (2015, April 29). 13 powerful hyper-converged infrastructure solutions. Retrieved from
http://www.crn.com/slide-shows/virtualization/300076666/13-powerful-hyper- converged-
infrastructure-solutions.htm/pgno/0/1
Murphy, K. (2015, April 01). FAQ from all-flash hyperconverged infrastructure webinar.
Retrieved from http://www.gridstore.com/faq-from-all-flash-hyperconverged-infrastructure-
webinar/
Ward, Chris. (2015, April 10). To converge infrastructure or not, that is the question. Retrieved
from http://www.tomsitpro.com/articles/converged_infrastructure-coverged_storage-roi-
data_center-hp,2-273.html
Newe, G. (2015, September 22). What does hyper-converged infrastructure mean for the future
of enterprise application delivery? Retrieved from http://www.cbronline.com/blogs/cbr- rolling-
blog/what-does-hyper-converged-infrastructure-mean-for-the-future-of- enterprise-application-
delivery
Pluribus networks named to CRN's 2015 virtualization 50. (2015, May 28). Retrieved from
http://www.businesswire.com/news/home/20150528006339/en/

Reference Architecture addresses hyper-convergence market. (2014, August 26). Retrieved from
http://www.thomaspublishing.com/about_thomas/business_units/index.php
Maheshwari, N. (2014, February 20). The secret to Google's datacenter and why it matters to
every enterprise. Retrieved from http://www.nutanix.com/2014/02/20/the-secret-to- googles-
datacenter-and-why-it-matters-to-every-enterprise/
Posey, B. (2015, August). Which hyper-converged systems are best for VDI? Retrieved from
http://searchvirtualdesktop.techtarget.com/answer/Which-hyper-converged-systems-are- best-
for-VDI
Mutegi, L. (2015, September 24). Kenya: VMware to target Kenyan SMEs with their new
hyperconverged infrastructure solution. Retrieved from
http://searchvirtualdesktop.techtarget.com/answer/Which-hyper-converged-systems-are- best-
for-VDI
10 hyperconvergence trendsetters. (2015, March 12). Retrieved from
http://www.networkcomputing.com/storage/10-hyperconvergence-trendsetters-/d/d- id/1319413
NT4Admins.de (2014, August 21). Virtual SAN product review – Datacore versus VMware.
Retrieved from http://datacore.com/sf-docs/default-source/opinions-reviews/product-
reviews/nt4admins-virtual-san-review.pdf
Jain, S., Kumar, A., Mandal, S., Ong, J., Poutievski, L., ... Vahdat, A. (2014, January 21). B4:
experience with a globally-deployed software defined WAN. Retrieved from
https://www.eecs.berkeley.edu/~sylvia/cs268-2014/papers/b4-sigcomm13.pdf
Haranas, M. (2015, October 8). Dell EMC acquisition would nix Cisco's participation with VCE,
partners say. Retrieved from http://www.crn.com/news/networking/300078391/dell-emc-
acquisition-would-nix-ciscos-participation-with-vce-partners-say.htm/pgno/0/1

Exploring HCI

More Related Content

What's hot (20)

Similar to Exploring HCI (20)

Exploring HCI