vSphere Integration for Array Management
A check mark indicates the array has a vSphere plug-in to perform array based management functions.
vSphere Integration for Snapshots
Support for full integration with vSphere’s snapshot capability either natively, or through a plug-in provided by the storage array vendor.
Full Copy
Used to copy data and/or create clones, avoiding the need to send data back and forth to a host. Full File Clone, or Full Copy, is similar to the Extended Copy (XCOPY) hardware acceleration primitive provided for block arrays. This primitive enables virtual disks to be cloned by the NAS device rather than by using the Data Mover, which consumes ESXi host, CPU, and memory resources as well as network bandwidth.
Hardware Assisted Locking
Allows vCenter to offload SCSI commands from the ESX server to the storage array so it can control the locking mechanism while the storage array does data updates.
Licensed to Nimbus Data Systems, Inc. with unlimited and unrestricted distribution rights.
© 2014 DCIG, LLC. All rights reserved.
2014-15 FLASH MEMORY STORAGE ARRAY BUYER’S GUIDE
The Insider’s Guide to Evaluating Flash Memory Storage Arrays
A-4 DCIG: Empowering the IT industry with actionable analysis.™
Appendix A—Definitions, Explanations and Terminology
(continued)Block Zeroing
The VMware vSphere ESXi host issues a command to the array that it recognizes. This command instructs the storage array to write zeros to the virtual VMDK or VMFS LUNs. This generally occurs when initializing the virtual disk for a new VM. Support for this feature indicates that the storage array recognizes this vSphere primitive and performs this zeroing activity. Enables the storage array to zero out a large number of data blocks to speed the provisioning of virtual machines (VMs) and reduce I/O.
Thin Provisioning Dead Space Reclamation (SCSI UNMAP)
Using the SCSI UNMAP command, an ESXi host can inform a storage array that space may be reclaimed that previously had been occupied by a VM that has been migrated to another datastore or deleted.
Thin Provisioning Stun
The first enhancement was introduced to address concern regarding the impact on virtual machines when thin-provisioned datastore usage reaches 100 percent of capacity. Historically, this affected all virtual machines running on the datastore. With the release of the vSphere 5.0 VAAI Thin Provisioning primitives, if a thin-provisioned datastore reaches 100 percent usage, only those virtual machines requiring extra blocks of storage space are paused; those not needing additional space continue to run. After additional space is allocated to the thin-provi-sioned datastore, the paused virtual machines can be resumed. (Source: http://www.vmware.com/files/pdf/
techpaper/VMware-vSphere-Storage-API-Array-Integration.pdf)
Full File Clone
This command is part of VAAI 5.0 and is a NAS hard-ware acceleration primitive. The VMhard-ware vSphere ESXi host issues a command to the array that it recognizes.
This command instructs the storage array to clone a virtual disk. Support for this feature indicates that the storage array recognizes this vSphere primitive and performs this cloning activity.
Fast File Clone
This command is part of VAAI 5.0 and is a NAS hard-ware acceleration primitive. The VMhard-ware vSphere ESXi host issues a command to an array that supports this feature. This command offloads the creation of VM snapshots to the storage array. Support for this feature indicates that the storage array recognizes this vSphere primitive and performs these offload activities.
Out-of-Space Conditions
This command is part of VAAI 5.0 and was introduced to mitigate the impact on VMs when thin-provisioned datastores reach 100 percent of capacity. The array alerts the VMware vSphere ESXi host and/or vCenter when specified thresholds are reached. Should a datastore reach 100 percent of capacity, only those VMs requiring additional capacity are paused while VMs needing no additional capacity continue to run. Support for this feature indicates that the storage array recognizes this vSphere primitive and performs this out-of-space alerting.
Reserve Space
This command is part of VAAI 5.0 and is a NAS hardware acceleration primitive. The VMware vSphere ESXi host issues a command to the storage array that it recognizes.
This command instructs the storage array to create thick VMDK files on NAS datastores. Support for this feature indicates that the storage array recognizes this vSphere primitive and performs this file allocation activity.
Extended Statistics
Extended Statistics enables vSphere functionality to display actual space usage statistics on NAS datastores without the use of third-party tools. Before the introduc-tion of Extended Statistics, it would have been neces-sary to use array-based tools to monitor the space being used on a thinly provisioned Virtual Disk Machine (VMDK) on a back-end data store.
VASA
vSphere Storage APIs for Storage Awareness is a set of APIs that permit storage arrays to integrate with vCenter for management functionality. Storage Awareness collects configuration, capability and storage health information from storage arrays. This allows the adminis-trator to build Storage Profiles based on capabilities.
AQDT
VMware ESX 3.5 Update 4 introduced an adaptive queue depth throttling (AQDT) algorithm that adjusts the LUN queue depth in the VMkernel I/O stack. This algorithm is activated when the storage array indicates LUN I/O congestion by returning a BUSY or QUEUE FULL status.
When LUN congestion is detected, VMkernel throttles the LUN queue depth. The VMkernel attempts to gradually restore the LUN queue depth when LUN congestion conditions subside. AQDT prevents the array from being flooded with LUN I/O requests and enables the array to recover to a normal operational state. NOTE: This is not applicable for NFS mounted storage.
Licensed to Nimbus Data Systems, Inc. with unlimited and unrestricted distribution rights.
2014-15 FLASH MEMORY STORAGE ARRAY BUYER’S GUIDE
The Insider’s Guide to Evaluating Flash Memory Storage Arrays
Appendix A—Definitions, Explanations and Terminology
(continued)SIOC
Storage I/O Control (SIOC) for VMware is a dynamic control mechanism for proportional allocation of shared storage resources to VMs running on multiple hosts.
VADP
vStorage APIs for Data Protection (VADP) is a data protection framework introduced in vSphere 4.0. VADP which enables centralized, off-host LAN free backup of vSphere virtual machines, reduces ESX host resources to do backup processing, and enables flexible backup windows.
vMSC
VMware vSphere Metro Storage Cluster (VMware vMSC) is a new configuration within the VMware Hardware Compatibility List. This type of configuration is commonly referred to as a stretched storage cluster or metro storage cluster. It is implemented in environments where disaster/
downtime avoidance is a key requirement. (Source: http://
blogs.vmware.com/vsphere/2012/05/vsphere-metro-storage-cluster-white-paper-released.html)
VASRM
vStorage APIs for Site Recover Manager (VASRM) offers remote replication features enabling a device to failover to a recovery site.
PSA
To manage storage multipathing, ESX/ESXi uses a special VMkernel layer—Pluggable Storage Architecture (PSA).
The PSA is an open modular framework that coordinates the simultaneous operation of multiple multipathing plugins (MPPs). PSA is a collection of VMkernel APIs that allow third party hardware vendors to insert code directly into the ESX storage I/O path. This allows 3rd party software developers to design their own load balancing techniques and failover mechanisms for particular storage array.
The PSA coordinates the operation of the NMP and any additional 3rd party MPP. (Source: http://kb.vmware.com/
selfservice/microsites/search.do?language=en_US&cmd=
displayKC&externalId=1011375)
Storage DRS
Storage Distributed Resource Scheduler (DRS) provides virtual machine placement and load balancing mecha-nisms based on I/O and space capacity.
Hardware
Max Raw Flash Capacity (Appliance)
The maximum amount of raw flash memory storage capacity in terabytes (TBs) that the array can have under
its management. This number only includes the array’s footprint and not the storage capacity of other systems it may have virtualized.
Max Raw Flash Capacity (Cluster)
The maximum amount of raw flash memory storage capacity in terabytes (TBs) that a fully configured cluster of arrays can have under its management. This number only includes the cluster’s footprint and not the storage capacity of other systems it may have virtualized.
Capacity per Rack Unit
The maximum amount of flash memory that this product can provide per standard EIA rack unit. This measure of storage density is presented in terms of Terabytes (TB) per rack unit.
Min Rack Units
The minimum number of standard EIA rack units required to deploy this solution.
RAID Options (Total #)
Lists the standard RAID (Redundant Array of Independent Disks) levels supported by the device.
Examples include: RAID 0 (striping without mirroring or parity), RAID 1 (mirroring without striping or parity), RAID 5 (block-level striping with distributed parity), etc.
Information Dispersal Algorithms
Information dispersal algorithms/erasure coding allows an array to store data across multiple locations as a way to enhance redundancy and protect against data loss.
(Generally used as an alternative to RAID.) Cache (Max DRAM/Max)
The maximum number of gigabytes (GBs) of cache that each array controller can support.
Max RAM: Specifies the maximum number of gigabytes (GBs) of DRAM cache.
Max: Cache may consist of any combination of DRAM, NVRAM and SSD.
Controller Config
Active-Active: Two controllers are configured with multipathing software and have concurrent access to the disk drives that are then presented as LUNs to attached hosts. In the event a controller fails or is taken offline, attached host(s) may detect little or no interruption in service as the other controller takes over for the failed unit though a drop in performance may occur. This is considered an enterprise class feature.
Licensed to Nimbus Data Systems, Inc. with unlimited and unrestricted distribution rights.
© 2014 DCIG, LLC. All rights reserved.
2014-15 FLASH MEMORY STORAGE ARRAY BUYER’S GUIDE
The Insider’s Guide to Evaluating Flash Memory Storage Arrays
A-6 DCIG: Empowering the IT industry with actionable analysis.™
Appendix A—Definitions, Explanations and Terminology
(continued)Active-Standby: All LUNs are controlled and assigned to a single active controller through which all storage I/O traffic is routed. A secondary standby controller takes over the storage I/O traffic for the active controller should it fail or be taken offline. The secondary controller also assumes control of the array’s LUNs at any time.
Dual Active: Both controllers are active and have concurrent access to the drives that are presented as LUNs to the attached hosts. The difference from Active-Active is that half of the storage array’s LUNs are assigned to and controlled by one controller and the other half of the LUNs to the other so a LUN can only be accessed down one path through one controller at a time. If a controller should be taken offline or fail, the other controller assumes control for its LUNs.
CPU Cores (Max)
The maximum number of CPU processor cores available in one fully-loaded rack-mountable controller chassis.
Scale-out Controllers (Max)
The maximum number of controllers that the model can support in a single logical configuration. An “X” indicates that this array does not support a scale-out configuration.
Concurrent SAN & NAS
The ability for the storage array to communicate with applications using both block (SAN) and network file system (NAS) clients at the same time.
FC
Fibre Channel (FC) is used as a networking protocol to transmit data between computer devices, such as a server and a storage device.
iSCSI
Internet Small Computer System Interface (iSCSI) is a block-based protocol for running SCSI commands over Internet Protocol (IP) on Ethernet to access storage resources.
Concurrent FC & iSCSI
The ability for the storage array to connect to hosts over Fibre Channel and iSCSI at the same time.
Storage Networking Ports (Max)
The maximum number of ports available in any configu-ration to connect the array to hosts or to a dedicated storage network. Controllers that support both Ethernet and Fibre Channel may also support concurrent use of both technologies. See “Concurrent FC & iSCSI” for
additional clarification. Ports for managing the array or interconnecting multiple storage controllers are not included in this count.
1/10/40 Gb Ethernet Ports (Max)
Indicates the maximum number of 1 Gb, 10 Gb and 40 Gb Ethernet storage networking ports supported by the array in any configuration.
8/16 Gb FC Ports (Max)
Indicates the maximum number of 8 Gb and 16 Gb Fibre Channel storage networking ports supported by the array in any configuration.
InfiniBand
Indicates the maximum number of InfiniBand storage networking ports supported by the array in any configu-ration. InfiniBand ports are generally offered as 40 or 56 Gb connections.
Non-disruptive Ops (Total #)
The native array capability to: update controller code, replace an entire controller, add a controller, add storage shelves/nodes, and migrate data without disrupting the availability of the storage system. Redundant and hot swap power supplies, fans, and drives are also included in this measure. The number specifies the number of different supported elements available. The specific elements supported for each product are available by accessing the DCIG Interactive Buyer’s Guide (IBG).
Operating Temp Range (Celsius)
The supported range of operating temperatures for this unit.
Max Watts
The maximum power draw of one fully-loaded rack-mountable controller chassis.
Max BTU
The maximum heat dissipation of one fully-loaded rack-mountable controller chassis.
Support
Contract Support Availability
Specifies the hours when support is available under service contract. Examples include: “business hours”
(9am – 5pm, weekdays), “24x7” (24 hours a day, seven days a week, but not holidays) and “24x7x365” (24 hours a day, seven days a week, including holidays).
Licensed to Nimbus Data Systems, Inc. with unlimited and unrestricted distribution rights.
2014-15 FLASH MEMORY STORAGE ARRAY BUYER’S GUIDE
The Insider’s Guide to Evaluating Flash Memory Storage Arrays
Appendix A—Definitions, Explanations and Terminology
(continued)Contract Support (Total #)
A count of the methods of support offered by the vendor. Options include: phone, email, web chat, remote monitoring, remote login/problem resolution, dispatch technician onsite, dispatch technician onsite with 4-hour GTD response, same-day delivery of replacement parts, pre-failure shipping of replacement parts, and proactive remediation based on phone home data. The number specifies the number of different supported elements available. The specific elements supported for each product are available by accessing the DCIG Interactive Buyer’s Guide (IBG).
Non-Contract Support (Total #)
A count of the methods of support offered by the vendor. Hours of availability and support methods may differ from those offered for on-contract arrays, and may require a fee. The specific elements supported for each product are available by accessing the DCIG Interactive Buyer’s Guide (IBG).
Hardware Warranty
Indicates the length of standard warranty that is included with the array at no extra cost.
Licensed to Nimbus Data Systems, Inc. with unlimited and unrestricted distribution rights.
© 2014 DCIG, LLC. All rights reserved.
2014-15 FLASH MEMORY STORAGE ARRAY BUYER’S GUIDE
The Insider’s Guide to Evaluating Flash Memory Storage Arrays
B-1 DCIG: Empowering the IT industry with actionable analysis.™