89 fifth avenue, 7th floor new york, ny 10003 www ... · edison: hp 3par 7200 versus netapp fas...
TRANSCRIPT
89 Fifth Avenue, 7th Floor
New York, NY 10003
www.TheEdison.com
@EdisonGroupInc
212.367.7400
White Paper
HP 3PAR StoreServ 7200 with File
Controller versus NetApp FAS3250
Storage System
Usable Capacity Analysis and
Comparison
Printed in the United States of America
Copyright 2015 Edison Group, Inc. New York.
Edison Group offers no warranty either expressed or implied on the information contained
herein and shall be held harmless for errors resulting from its use.
This report was developed by Edison Group, Inc. with HP assistance and funding. This report
may utilize information, including publicly available data, provided by various companies and
sources, including HP. The opinions are those of Edison Group, Inc. and do not necessarily
represent HP’s position. All products are trademarks of their respective owners.
First Publication: January 2015
Produced by: Chris Evans; Senior Analyst; Kalicharan Rakam, Analyst; Manny Frishberg, Editor;
Barry Cohen, Editor-in-Chief
Table of Contents
Executive Summary ..................................................................................................................... 1
Lab Configuration and Evaluation Methodology.................................................................. 3
Test Process ................................................................................................................................ 3
Test Results ................................................................................................................................... 4
Capacity at Initialization .......................................................................................................... 4
Share Provisioning Efficiency.................................................................................................. 6
Deduplication and Compression Efficiency .......................................................................... 6
Conclusion ..................................................................................................................................... 8
Appendix ..................................................................................................................................... 10
SUT Configurations ................................................................................................................ 10
System Capacity Metrics ........................................................................................................ 12
Datasets .................................................................................................................................... 13
Reference Screen Shots ........................................................................................................... 20
Storage Provisioning Reserve ............................................................................................ 23
Endnotes ...................................................................................................................................... 27
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 1
Executive Summary
The continuing massive growth in requirements for data storage capacity is a given for
almost any organization today. Growth is seen predominantly in what is known as
unstructured data, information that is typically text heavy and often stored as discrete
files rather than in a database or other controlling program. Unstructured data can
include e-books, documents, health records, audio, video, images, e-mail messages and
more. IDC expects a 50-fold increase in these data from 2010 to 20201.
The exponential rate of growth seen in unstructured data means that throwing more
storage capacity at the problem is not a tenable solution. Instead, storage vendors have
been working on ways to reduce costs and optimize the amount of data physically
stored on both disk and solid-state drives. As a result, we have seen features such as thin
provisioning, data deduplication and compression being introduced into primary
storage systems. These data reduction techniques provide varying levels of efficiency,
both at initialization, and when systems have been loaded with data, which is highly
dependent on system architecture.
Data duplication identifies chunks of repeated data across entire storage systems,
reducing the replicas to a single copy retained on disk. Compression is a technique that
identifies and optimizes repeated patterns of information, for example within a single
file. Data deduplication and compression are features offered by both HP 3PAR
StoreServ with File Controller and NetApp FAS storage systems.
Edison performed an analysis of the benefits of data reduction technologies between the
HP 3PAR StoreServ 7200 with File Controller and NetApp FAS3250 systems. The data
used was a typical representation of unstructured data found within most organizations
and was derived from Edison’s own file servers. Testing was performed to show the
savings on deployed capacity both before and after data was written to the system. This
shows the relative advantage of the HP 3PAR platform before data as even been written
to the array.
The results showed that system overhead on HP 3PAR StoreServ systems was only
3.09% compared to 10.93% on NetApp FAS. Furthermore when comparing the usable
capacity of 100TiB of storage, HP 3PAR StoreServ provided 8.8% more capacity than the
NetApp platform. Once the system was loaded with test data, the HP 3PAR StoreServ
system with File Controller reduced the allocated capacity by 24% compared to a
NetApp FAS total of 17% (6% dedupe, 11% compression).
Given 100TiB of raw storage, the HP 3PAR StoreServ solution would provide 64.61TiB of
usable capacity after system overhead and RAID-6(4+2). In comparison, the NetApp
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 2
FAS system provides only 59.38TiB after overheads and RAID-6 (3+2). When applying
the potential dedupe savings as observed in this analysis, the HP 3PAR StoreServ
system would be capable of storing 90.01TiB of usable capacity due to the reduction in
duplicated data. In comparison, the NetApp FAS system was less efficient, resulting in
an effective usable capacity of only 71.65TiB.
The figures are represented in Figure 1 and Table 2 below and have been adjusted to
ensure the RAID overhead is comparable between both systems.
Figure 1 – HP/NetApp Savings Comparison, 100TiB
Table 1 - HP/NetApp Savings Comparison 100TiB
In conclusion, we see that the HP 3PAR StoreServ 7200 with File Controller is both more
efficient at initial allocation and once data is loaded onto the system compared to the
equivalent technology from NetApp, making it the clear choice from an efficiency
perspective.
Metric
Adjusted RAID-6 Values
HP 3PAR StoreServ NetApp FAS
Physical (TiB) 100.00 100.00
System Overhead (TiB) 96.91 89.07
RAID-6 64.61 59.38
Usable after Dedupe 90.01 71.65
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 3
Lab Configuration and Evaluation Methodology
To perform the testing for this report, Edison installed two systems into the company’s
New York City Lab. HP provided HP 3PAR StoreServ 7200 Storage with 3PAR StoreServ
File Controller and Edison rented a NetApp FAS3250 Storage System.
Both systems were configured with SMB (Server Message Block) / CIFS (Common
Internet File System) shares and mapped to an HP ProLiant staging server that held the
test data on local storage. The dataset file sizes, drive capacity and utilization levels were
obtained from the respective management user interfaces, normalized to binary for
precision and consistency. See “System Capacity Metrics” on page 12 for more
information.
The dataset used for the efficiency comparison was comprised of actual Edison
documents copied to the staging server. See “Datasets” on page 13 for details.
Test Process
The test process was performed over a number of steps.
Measure System Capacity/Efficiency at Initialization. The available capacity of 1.
both systems was measured prior to any data being copied into the SMB shares.
Copy Data to the Test Systems. Data stored locally on the ProLiant management 2.
server was copied to the SMB shares.
Measure. The utilization of each system and respective savings were measured 3.
both from the array management user interfaces, and from the host file system
view, as reported by Windows Server 2012 installed on the ProLiant
management server.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 4
Test Results
The data presented below has been taken from the management user interfaces of the
two storage systems in the test comparison. As the platforms use different metrics to
display capacity, the data has been normalized to the binary format utilized by the HP
3PAR StoreServ system (see “System Capacity Metrics” on page 12 for details).
Capacity at Initialization
Table 2 compares the reported and calculated capacities for each of the 2TB drives
installed into the two systems under test (SUT) after system initialization and before
hosts/clients were attached or shares created. HP 3PAR StoreServ systems show just
over 3% overhead, whereas the equivalent figure on NetApp FAS was almost 11%. The
NetApp value derives from a combination of the SATA drive block formatting overhead
for error correction and the rounding used by NetApp to ensure all drives report the
same capacity for similar drive models.
Metric HP 3PAR
StoreServ & File
Controller
NetApp FAS HP 3PAR
Advantage
Reported Disk Capacity 1,805GiB 1.62TiB
(1,659GiB)
146GiB (8.8%)
Normalized Capacity
(MiB)
1,848,320 1,698,816 149,504GiB (8.8%)
Standard 2TB Drive
Capacity (MiB)
1,907,349 1,907,349 N/A
System Overhead (MiB) 59,029 208,533 28.3% of NetApp
overhead
System Overhead (%) 3.09% 10.93% N/A
Table 2 - Physical Presented Drive Capacities
Table 3 shows the figures on disk capacities extrapolated up to cover the 24 drives
inserted into a standard disk shelf.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 5
Metric (Figures in MiB unless
Specified)
HP 3PAR StoreServ
with File Controller
NetApp FAS
24 Drive Shelf Capacity (MiB) 44,359,680 40,771,584
System Overhead at Provisioning
(MiB)2 0% 5%
Capacity After System Overhead
(MiB) 44,359,680 38,733,001
3PAR Capacity Advantage (MiB) 5,626,679
3PAR Advantage as Drive Count
(Difference/Normalized Capacity)
3.06
Table 3 - Savings per 24-drive Disk Shelf
At initialization, NetApp systems reserve 5% of capacity for Snapshot Reserve, although
the customer may amend this figure. Over a standard 24-drive shelf, the relative saving
of HP 3PAR StorServ translates to approximately three 2TB drives. Alternatively,
expressing these figures on the basis of a 100TiB allocation of storage we see overheads
as detailed in Table 4. Customers can expect to receive almost 8TiB of extra capacity per
100TiB of storage over the equivalent NetApp FAS array when using HP 3PAR
StoreServ systems. Alternatively, expressing this in terms of the requirement to provide
100TiB of usable capacity, NetApp FAS systems require an additional 9.1TiB of physical
disk capacity to be installed.
Metric HP 3PAR StoreServ
with File Controller
NetApp FAS
Usable Capacity 96.91TiB 89.07TiB
3PAR Savings Advantage 7.84TiB
Required Capacity for 100TiB 103.2TiB 112.3TiB
NetApp uplift required for 100TiB
of usable capacity
9.1TiB
Table 4 – Savings per 100TiB of Storage
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 6
Share Provisioning Efficiency
Testing was performed on 100GB (93GiB) SMB shares provisioned through the
management interface of each system.
As discussed, NetApp’s Data ONTAP file layout (known as WAFL) requires a 5%
reserve for snapshots, further reducing the space available per share by that amount.
Neither the user interface on the storage system nor the client mounting the share shows
this overhead directly; it can only be viewed when managing snapshots —usually when
the reserve has been consumed — and snapshots need to be removed to allow the
feature to continue functioning. The 3PAR StoreServ array requires no reservation space
for snapshots.
Deduplication and Compression Efficiency
Table 5 presents the data from deduplication and compression of the unstructured data
set. The raw data capacity was 10.8GiB, with NetApp reporting a reduction of 2GB
(normalized to 1.86GiB) for a saving of 17%. The HP 3PAR StorServ system with the File
Controller reported a reduction of 3.06 GiB for a saving of 24% (based on sub-file level
deduplication performed by the File Controller), showing greater savings than NetApp
FAS. In this direct comparison, HP 3PAR StoreServ with File Controller is more than
64% more efficient in optimizing unstructured data than NetApp FAS.
Metric HP 3PAR StoreServ 7200
File Controller
NetApp FAS3250
Space Saving (GiB) 3.0600 1.86265
Space Used (GiB) 7.7527 8.95005
Space Efficiency – UI Figures (%) 24% 17% from 6%
(dedupe) and 11%
(compression)
Relative Advantage (%) 64.3%1
Table 5 - Data Reduction Efficiency Comparison 10.8GiB Data
The screen shots in Figure 2 and Figure 3 on page 7 show the deduplication results as
reported on the two systems. HP 3PAR StoreServ performs compression as part of the
same process. Compression is a manual, post-write process on the NetApp FAS system3.
1 Calculated as the percentage improvement of physical space saved by HP 3PAR StoreServ over NetApp FAS
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 7
This post-process compression consumes processor cycles, often affecting system
performance, when large numbers of files need to be compressed or rehydrated when
files need to be read.
Figure 2: Deduplication Results for HP 3PAR
Figure 3: Deduplication Results for NetApp FAS
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 8
Conclusion
The growing amount of unstructured data is one of the fastest consumers of enterprise
storage capacity. Storing this data is a major cost factor for most organizations, and
managing these costs is critical in enabling IT groups to achieve business goals within
tight budget restraints.
Hardware vendors have incorporated several approaches to accommodating the conflict
between storage growth and customer needs to control capital expenditures. Among
these approaches are thin provisioning, compression and deduplication.
Thin provisioning, as the name implies, allows a storage administrator to define storage
space capacity for a server without actually setting aside that space until actually
needed.
Deduplication reduces capacity requirements by applying algorithms that eliminate the
need to physically store multiple copies of identical data. This allows an organization to
acquire less physical capacity at the outset and delay acquisition of additional capacity.
(Data is still growing; deduplication and thin technologies merely slow the need to
acquire additional physical capacity.)
Edison’s testing showed that HP 3PAR StoreServ systems have much less physical
storage capacity overhead to manage system features such as snapshots, thin
provisioning and data integrity. The overhead on HP 3PAR StoreServ in this comparison
was approximately 3% compared to NetApp FAS at approximately 11%. NetApp
systems show a further 5% overhead to accommodate space reserved for snapshots. HP
3PAR StoreServ systems have no snapshot reserve.
The testing performed shows that HP 3PAR StoreServ with File Controller
deduplication and compression on the test unstructured dataset was 64% more for the
NetApp system; HP 3PAR StoreServ reduced the dataset by 24% compared to NetApp
FAS at 17%.
Given 100TiB of raw storage, the HP 3PAR StoreServ solution would provide 64.61TiB of
usable capacity after system overhead and RAID-6 (4+2). In comparison the NetApp
FAS system provides only 59.38TiB after overheads and RAID-6 (3+2). When applying
the potential dedupe savings as observed in this analysis, the HP 3PAR StoreServ
system would be capable of storing 90.01TiB of usable capacity due to the reduction in
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 9
duplicated data. In comparison, the NetApp FAS system was less efficient, resulting in
an effective usable capacity of only 71.65TiB.
The figures are represented in the following table and have been adjusted to ensure the
RAID overhead is comparable between both systems. These figures are also illustrated
in Figure 1 on page 2.
Metric
Adjusted RAID-6 Values
HP 3PAR StoreServ NetApp FAS
Physical (TiB) 100.00 100.00
System Overhead (TiB) 96.91 89.07
RAID-6 64.61 59.38
Usable after Dedupe 90.01 71.65
Table 6 – HP/NetApp Savings Comparison, 100TiB
Overall, HP 3PAR StoreServ systems incorporating the File Controller offer significantly
higher rates of data efficiency for unstructured data when compared to the equivalent
NetApp FAS platform.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 10
Appendix
SUT Configurations
The tables below show the specifications for the two SUT and how the two systems were configured for testing. Table 7 on page 10 provides
an overview of the system specifications
SUT Component HP 3PAR NetApp FAS
Model HP 3PAR StoreServ 7200 with
HP 3PAR StoreServ File Controller
NetApp FAS 3250
Storage Software HP 3PAR OS v3.1.2
Data ONTAP version 8.2.1 (CDOT)
Number of Controllers 2 block and 2 file – 10U Rack space 2 – 10U rack space
Number of Drives 24 24
Drive Type 2 TB NL-SAS (7K RPM) 2 TB SATA drives in SAS enclosure (7K RPM)
Table 7: SUT General Specifications
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 11
Table 8 on page 11 provides an overview of the test environment including an overview of array nodes, the networking components and the
staging server used for file copying and as a management station.
ILO/Node Management
Device/Component Label Vendor Device/Component Label Vendor
3PAR StoreServ File Controller v2 (node1) ILO HP TestCluster_01 MGMT NetApp
3PAR StoreServ File Controller v2 (node2) ILO HP TestCluster_02 MGMT NetApp
Storage Management
Device/Component Label Vendor Device/Component Label Vendor
StoreServ Node 1 HP TestCluster MGMT NetApp
3PAR StoreServ File Controller v2 (node1) HP
3PAR StoreServ File Controller v2 (node2) HP
StoreServ Node 2 HP
Shared Resources
Switches
8/24 SAN Switch Not used
6600ml-24G-4XG Network Switch 10 GB Ports used for NAS
Server for DC/Control
HP ProLiant DL 360 G7
Table 8: Test Environment Comparison
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 12
System Capacity Metrics4
HP and NetApp use different measuring systems for storage capacity and utilization in the management user interfaces of their products,
which can cause confusion when comparing the relative storage efficiency of the two systems. HP uses the binary Gibibyte (GiB) and
Tebibyte (TiB) formats while NetApp uses decimal Gigabyte (GB) and Terabyte (TB) respectively5. Most of the technology industry
(including hard disk drive vendors) uses the decimal format, mainly out of habit and convenience, as the mental math is easier. However
binary is much more precise, especially when used for comparisons as in this study. Table 9 on page 12 illustrates how the two metrics
differ.
Label Bytes Power Label Bytes Power
1 GB or 1 Gigabyte 1,000,000,000 10003 1 GiB or 1 Gibibyte 1,073,741,824 10243
1 TB or 1 Terabyte 1,000,000,000,000 10004 1 TiB or 1 Tebibyte 1,099,511,627,776 10244
Table 9: Binary versus Decimal Byte Comparisons
The difference between representations may seem insignificant especially at the lower capacity levels, but when comparing large capacity
the greater precision can be significant. For example, the difference between 1 GB and 1 GiB is 7.37 percent, whereas the difference between
1 TB and 1 TiB is 9.95 percent, demonstrating increasing disparity between the two measuring systems as capacities scale.
To limit confusion and increase precision, the comparisons in this study have been normalized to binary values and represented at the
Gibibyte level.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 13
Datasets
The reference file-set for the product comparisons in this test are a selection of actual content from Edison's own IT systems. The files
include Microsoft Office files (DOC, DOCX, XLS, XLSX, PPT, PPTX), Open Office files (ODT, ODS), Adobe Acrobat files (PDF), web pages
generated from SPEC benchmarks, graphics files (typically BMP, JPG, PNG), source and compiled code (JS, PHP, ASP). In general, these
represent the type of file content typically used in most businesses. These files were copied to the arrays creating identical copies of the file
content on each device. Figure 4 on page 14 and the legend list that follows shows a breakdown of the file set of the number of files by file
extension. Figure 5 on page 17 and the legend list illustrates the breakdown by file size and by extension.
Some of these files, such as the Microsoft Office, web pages and source code are highly compressible so the space savings from compression
mostly comes from these files. The other file types, such as PDF, JPG, PNG, audio and of course ZIP, are already compressed and generally
resulted in little or no deduplication or compression savings.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 14
Figure 4: Data Set Breakdown – Distribution of Files by Extension
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 15
Distribution by Extension Details
DOCX Files: 11.72%
TXT Files: 11.63%
HTML Files: 11.53%
PNG Files: 11.28%
LOG Files: 9.37%
RESULTS Files: 9.29%
ASPX Files: 7.81%
RAW Files: 6.81%
JPG Files: 4.75%
DOC Files: 4.10%
PDF Files: 3.08%
XLSX Files: 1.40%
XLS Files: 1.08%
DOTX Files: 0.83%
ZIP Files: 0.82%
WMA Files: 0.80%
PPT Files: 0.65%
PPTX Files: 0.60%
CSV Files: 0.48%
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 16
AVI Files: 0.29%
CSS Files: 0.26%
MP3 Files: 0.26%
VCP Files: 0.15%
BMP Files: 0.15%
MMP Files: 0.13%
JS Files: 0.08%
SWF Files: 0.08%
VSD Files: 0.07%
GIF Files: 0.05%
RTF Files: 0.05%
ODT Files: 0.03%
ODS Files: 0.03%
Other Files: 0.36%
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 17
Figure 5: Data Set Breakdown – Disk Space by Extension
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 18
Disk Space Utilization by Extension Details
WMA Files: 32.7% (3.80 GB)
MP3 Files: 12.69% (1.47 GB)
AVI Files: 8.46% (0.98 GB)
DOC Files: 8.23% (0.96 GB)
DOCX Files: 7.71% (915.74 MB)
PDF Files: 6.29% (747.65 MB)
PPT Files: 4.50% (534.37 MB)
JPG Files: 3.76% (446.92 MB)
PPTX Files: 2.34% (278.59 MB)
VCP Files: 1.83% (217.07 MB)
BMP Files: 1.31% (156.19 MB)
ZIP Files: 1.13% (133.98 MB)
HTML Files: 0.81% (95.79 MB)
RAW Files: 0.80% (95.35 MB)
WAV Files: 0.76% (90.57 MB)
SWF Files: 0.64% (75.62 MB)
XLS Files: 0.60% (71.50 MB)
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 19
RESULTS Files: 0.58% (69.13 MB)
CSV Files: 0.57% (68.28 MB)
CAMREC Files: 0.55% (64.83 MB)
PNG Files: 0.46% (55.08 MB)
MP4 Files: 0.41% (48.65 MB)
RAR Files: 0.36% (43.31 MB)
XLSX Files: 0.36% (43.03 MB)
TXT Files: 0.31% (36.36 MB)
LOG Files: 0.27% (32.07 MB)
WMV Files: 0.27% (31.99 MB)
IBOOKS Files: 0.14% (16.81 MB)
MOV Files: 0.12% (14.21 MB)
ASPX Files: 0.12% (13.89 MB)
MPP Files: 0.11% (13.59 MB)
ODT Files: 0.10% (11.91 MB)
Other: 0.64% (75.48 MB)
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 20
Reference Screen Shots
The screen shots on the following pages provide source references for the data used in this study.
System Capacities
Figure 6 on page 21 shows the post-initialization capacity of the drives in the 3PAR StoreServ array. The cage0 drives were not used in
this study.
Figure 7 on page 22 shows similar information for the NetApp array. Only those drives used in the test aggregate show their effective
capacity.
Figure 8; page 23, Figure 9; page 24, Figure 10; page 25, and Figure 11; page 26 show provisioning data from the 3PAR StoreServ 7200
system. The standard NetApp management software does not provide the same amount of configuration detail; additional software is
required.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 21
Figure 6: HP 3PAR Results
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 22
NetApp reserves physical disk space for checksums on all SATA drives, which results in additional reserve capacity and overall lower
usable capacity (typically around an 11% capacity loss).
Figure 7: NetApp FAS Results
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 23
Storage Provisioning Reserve
HP 3PAR Provisioning Reserve
Figure 8: HP 3PAR CPG has no system reserve and by default grows into the overall system capacity.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 24
Figure 9: Thin Provisioned Virtual Volumes system overhead is negligible (<1%).
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 25
Figure 10: Once the virtual volume is created, CPG maps the space for the virtual volumes, which is 1% of the overall system capacity.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 26
Figure 11: Virtual Copy (snapshots) by default sets a 1% overhead. A 100GB virtual volume has a total overhead of 1.2GB. Copy overhead for fully provisioned and
thin provisioned virtual volumes is the same. There is no additional overhead for the clones.
Edison: HP 3PAR 7200 versus NetApp FAS 3250: Usable Capacity Analysis Page 27
Endnotes
1 The Digital Universe in 2020, December 2012. 2 NetApp sets aside three drives for a root aggregate at provisioning time. Root contains the array operating system and configuration data. 3PAR utilizes
custom ASICs and other embedded technology for this purpose requiring minimal identity tagging information be written to the array drives. 3 In a past study for IBM, Edison tested compression on a NetApp filer. Although the NetApp system tested at that time was a different model than the
one used in the current study, the nature of the results would be the same today:
A Comparative Evaluation of Block Storage Compression Technology for Multipurpose Storage Environments, July 2012. 4 Historically, the computer industry used Kilobytes (KB) and Megabytes (MB) for storage and memory capacity. These terms were used as a compromise
for the byte multiples that needed to be expressed as power of 2 but lacked a convenient name. In 1998, standards for binary prefixes were established by
the International Electrotechnical Commission (IEC) giving birth to Kibibytes, Mebibytes, Gibibytes and so forth. This standard has been embraced by
some fields, especially the memory standards by the JEDEC or Joint Electron Device Engineering Council, but the use of Kilo, Mega, Giga and so forth is
still a permitted usage, but the binary prefixes are the endorsed standard. 5 To add to the confusion, HP uses GB and TB when describing the size of the drives in product specifications and non-technical documents.
4AA5-6485ENW