HP Hewlett Packard AutoRAID 12H User Manual

hp virtual array:  
double your operating efficiency  
table of contents  
customer quotes  
Amazon.com  
Room & Board  
State of New Mexico  
i
i
i
ii  
ii  
ViaSat  
Santa Clara  
section 1: hp storage white paper  
November 2001  
1
virtualization, simplification and storage  
1.0  
1.0  
1.1  
1.1  
1.2  
1.2  
1.3  
1.4  
1.4  
1.4  
1.5  
1.5  
1.5  
1.6  
1.8  
1.8  
array virtualization defined  
the reasons you need more virtualization in the array  
key features of hp’s patented virtual technology  
why now is the time for array virtualization  
managing traditional storage  
configuring an array for a database  
the system administrator’s dilemma  
managing the hp virtual array  
adding capacity with hp virtual arrays  
time to implementation: formatting the array  
automating the cache parameters  
performance  
faqs  
summary  
for more information  
section 2: “hp extends virtualization to  
the array”  
—by D.H. Brown Associates, Inc. November 15, 2001  
2
virtual value in hp’s va7000 series  
super redundancy enhances RAID  
reliability characteristics of hp’s va7000 series  
software extends virtual functions  
2.0  
2.1  
2.2  
2.3  
customer quotes  
After seven years and 20,000 sales, we hardly consider  
virtualization new. Many companies have optimized their performance with HP, including:  
Amazon.com  
State of New Mexico  
“In consolidating from eight controllers to the HP  
Surestore VA7100 we expected to see some  
performance degradation in running our Children,  
Youth, and Family Department databases. We  
haven't seen any degradation and are pleased  
with the performance. We've found it very easy  
to allocate space with the VA7100 and have  
saved time with reduced administrative overhead.  
With the AutoRAID feature, we've eliminated  
worrying about mirroring. We let AutoRAID go  
to work and don't even have to think about it.”  
We look at three things when selecting data  
warehouse disks. Price per terabyte, throughput  
capabilities, and availability features. With the  
HP Virtual Array 7100, the price per terabyte  
was extremely competitive and the throughput  
wasphenomenal. Our current Superdome VA7100  
configuration is capable of driving three to four  
gigabytes of I/O per second, and the autoraid  
features of the VA7100 exceed our availability  
requirements.”  
Mark Dunlap  
Director of Data Warehousing  
Amazon.com  
Tom Elder  
DBA 3 Children, Youth and Family  
Department  
State of New Mexico  
Room & Board  
“After reviewing various competitive scenarios  
for our SAP environment, we selected HP as our  
single vendor with rp5450 servers and a VA7100  
for our SAN. We found the set-up to be fast and  
easy, and the performance excellent. It was an  
easy decision to add a second virtual array, a  
VA7400, to accommodate our growth and  
position us for the future.”  
John Focht  
Systems Administrator  
Room & Board  
i
customer quotes cont’d  
Santa Clara  
ViaSat  
We needed to move to the next generation of  
technology as we merged data centers. We had  
outgrown the large server farm scenario and  
wanted to focus on consolidation. In discussing our  
current and future needs for our NT environment,  
the VA7400 was the best solution. We have  
transformed to an enterprise environment with  
great expandability for the future. And as a hospital  
environment with needs for an always-on, 24x7  
operation, we've found a reliable solution with HP.”  
We selected the HP Surestore Virtual Array  
7400 storage based on the number of Oracle  
ERP instances that would be needed. We found  
the Virtual Array has many more features and  
capabilities than the Sun solution that was being  
proposed. We felt it would scale the way we needed,  
provided strong price/performance, and offered  
excellent operating efficiencies. It has saved many  
hours and long nights for our DBA’s and system  
administrators. Plus, we're using the HP Surestore  
tape library for efficient tape backup data protection.”  
Gary Davis  
Hospital Information Systems Manager II  
Santa Clara  
Mike Johnson  
IT Manager  
ViaSat  
ii  
section 1  
hp storage white paper  
virtualization, simplification  
and storage  
These are not trivial benefits. Data centers are  
already complex, and with the continued explosion  
1
in storage capacity they will only become more so.  
HP’s Array Virtualization is evolutionary, not  
revolutionary. It is a logical progression in array  
technology. It is proven. It is necessary. It is the  
perfect storage area network (SAN) technology.  
And it is offered only by Hewlett-Packard!  
In 2001, HP announced two new disk arrays—  
the HP Surestore Virtual Array 7400 and the HP  
Surestore Virtual Array 7100. This paper will focus  
entirely on the virtualization technology that makes  
these the easiest arrays to manage and the most  
intelligently simple arrays on the market. This paper  
answers the question, “What is the value  
of HP’s Array Virtualization?”  
array virtualization defined  
The purpose of virtualization in any technology  
is to hide complexity from the user, or in the case  
of disk arrays, to hide complexity from the storage  
administrator and provide a standard environment  
for application development and increased  
price/performance.  
For those familiar with data center environments,  
virtualization is not a new concept. Virtualization  
already exists to some extent at every point in the  
solution stack. For example, servers, operating  
systems, databases, file systems, volume managers,  
drivers, switches, and storage devices all require  
virtualization to achieve their purposes.  
Disk arrays are complex devices designed for  
complex tasks. A disk array with 50 disk drives  
is more complex to manage than a disk array  
with one drive. Armed with virtualization, an  
array could potentially allow the 50 drives to be  
perceived and managed as one big drive or as  
one big pool of storage. The power of virtualization  
is the power of simplification.  
At the array level, HP’s Virtual Architecture simply  
expands on already familiar ground, and it  
cleanly fulfills the promise of virtualization—  
it hides complexity from the administrator and  
can have a dramatic and positive effect on real-  
world performance.  
Now for the definition: Virtualization in arrays is  
about creating and managing virtual storage  
devices. It is about taking blocks of storage on  
the disk drives and presenting them as LUNs  
(logical units of storage). What system administrators  
see, then, are not the actual physical disk drives  
but rather a created, simplified “virtual” view of  
the actual physical storage, i.e., the LUNs.  
This reduction in complexity greatly simplifies and  
streamlines the data center environment.  
hps array virtualization:  
reduces the time spent managing individual  
arrays  
allows storage administrators to manage  
more storage with less effort  
HP’s Virtual Array works with LUNs just as traditional  
arrays do. However, the virtual array doesn’t  
stop there. The virtual array actually manages  
the disks down to the level of the smallest available  
“cluster” or “chunk.” Further, while traditional  
arrays utilize static address-translation algorithms  
for managing the chunks, the virtual array uses  
a dynamic mapping system.  
reduces the opportunities for human error  
frees up precious IT resources to work on  
revenue-generating projects  
self-manages the RAID configuration for  
optimum performance  
1.0  
hp storage white paper  
array virtualization defined  
This allows the array to dynamically allocate and  
de-allocate clusters of any RAID type without  
affecting the logical view as seen from the server.  
In other words, clusters can be moved and new  
RAID 5 and RAID 1+0 stripes can be created or  
extended dynamically without the host view being  
changed and without system administrator  
intervention.  
achieve a balanced workload and eliminates  
“hot spots”  
1
automatically adds new disk drives into  
existing RAID groups  
Benefit: capacity upgrades require no  
IT training, are faster, and hot spots are  
virtually eliminated  
Just as a file system on the host presents a virtualized  
view of the storage to the application or user,  
the virtual array presents a virtualized view of  
the storage to the file system. For example, just  
as you can create and delete files in a file system,  
you can create and delete LUNs within the virtual  
array. This degree of virtualization is what gives  
the virtual array its unique ability to relieve the  
system administrator of many of his/her mundane  
storage management responsibilities.  
dynamically, and without any human  
intervention, optimizes the RAID level to  
the application workload  
Benefit: performance tunes the array 24  
hours per day and eliminates downtime for  
reconfigurations  
allows LUNs to be created in seconds  
Benefit: reduces setup times  
allows immediate use of the array after  
LUN configuration  
Benefit: speeds implementation time  
the reasons you need more  
virtualization in the array  
capacity requirements will always grow  
allows hundreds of LUNs to be created of  
virtually any size—from a single megabyte  
to over 2TB  
Benefit: simplifies configurations, improves  
application performance, and greatly  
decreases management complexity and  
potential for security errors  
IT departments will always have a limited  
budget and limited human resources for  
managing storage  
virtualization improves the efficiency of  
storage administrators  
virtualization reduces the training required  
for people to manage storage  
data always grows, but human head count  
doesn’t—virtualization reduces the number  
of people required to manage storage  
allows any combination of disk drive  
capacities and speeds within the array  
Benefit: simplifies disk drive inventory  
control and capacity management  
key features of hps patented  
virtual technology  
allows disk drives and controllers to be  
moved to any slot in the array  
Benefit: reduces the chance for human error  
automatically stripes every LUN across all  
disks in a very large redundancy group  
Benefit: simplifies overall management;  
reduces the number of LUNs required to  
1.1  
hp storage white paper  
reasons for more virtualization in the array  
why now is the time for array  
virtualization  
managing traditional storage  
1
“The mistakes are all there waiting to be made.”  
—Chessmaster Savielly Grigorievitch  
Tartakower on the game’s opening  
position  
Data centers have become enormously complex.  
Interestingly, new technologies that initially held  
the promise of simplifying the data center have,  
in fact, added to the complexity—at least for  
now. Fibre Channel, SANs, and the low cost  
and ease of adding storage have enabled data  
centers to expand both in size and complexity.  
Configuring and managing traditional storage  
is time-consuming and challenging even for  
experienced system administrators. When  
configuring storage, a wide variety of factors  
must be accounted for, judgments must be  
made, and steps repeated over and over again.  
The world is moving from SCSI to Fibre Channel.  
Fibre Channel and SANs allow for more devices  
to be connected together, with greater performance,  
and at far greater distances. All of these are  
wonderful benefits, but along with those benefits  
comes an ability to now create configurations  
that are larger, more complex, further apart, and  
harder to manage than anything envisioned even  
a few years ago. Couple this new reality with the  
exponential increase in storage, fueled by its low  
cost, and you have a recipe for losing control of  
the data center.  
The potential for error is high. The cost of configuring  
an array improperly is also high. The administrator  
must consider the following factors:  
capacity, cost, performance and availability  
requirements  
requirements for future additional capacity  
and performance  
number of disk drives and their capacity  
performance characteristics of the disks  
which RAID level will meet desired capacity,  
cost, and availability needs  
Consider this: while human resources remain static,  
environments grow larger and more complex.  
Simplification is the only realistic answer.  
Adding newer and slicker management software  
oftentimes can help, but in the long term, adding  
software tends to have the opposite effect of  
reducing complexity. Software may give you  
a central place from which to manage your  
hardware and it may simplify processes, but  
it doesn’t necessarily eliminate those processes.  
Armed with better management software,  
environments are encouraged to add more  
hardware and eventually the environment  
becomes as complex as it was before, but for  
different reasons. And when the human resources  
are already stretched to the breaking point, this is a  
recipe for complexity, stress, long hours, and human  
error. Virtualization is the answer. It solves the basic  
underlying problem. It permanently simplifies the  
environment for the system administrator.  
number and type of RAID groups  
number of LUNs based on application,  
performance, and array configuration needs  
size of LUNs  
configuration of the server volume manager  
cache configuration options  
stripe depth configuration  
implementation plan: who, what, and when  
time to bind LUNs  
1.2  
section 1  
hp storage white paper  
In addition, the same processes must be followed  
whenever disk drives are added or the environment  
changes. Plus, these additional factors must be  
considered:  
administrators oftentimes don’t have all the critical  
information. They don’t know the precise database  
performance requirements for each of the pieces,  
and they don’t know the performance behavior  
of the array in its multitude of configurations.  
1
current configuration  
desired additional capacity, performance,  
and availability  
In these real-world environments, it is typically far  
too time-consuming to try a number of different  
storage configurations, so database administrators  
typically apply rules from previous installations.  
The changing characteristics of newer versions  
of the database typically result in an unbalanced  
configuration that has “hot spots” that limit the  
performance of the system.  
whether the new disks will be stripe extensions  
of existing disks or be independent groups  
here is a typical process for setting up  
a traditional array:  
1. Determine number of disks, number of RAID  
groups, disks and disk type per RAID  
group, RAID level of each group, total  
LUNs, LUNs per RAID group, stripe depth.  
2. Determine volume manager configuration,  
stripe size and depth, LUNs per logical  
volume.  
3. Using the command station, set up the  
LUNs and their RAID levels and assign  
them to particular disks.  
4. Set up the cache page size depending  
on the size of the I/Os coming in from  
the host.  
This entire process can take from a few hours  
to several days, depending on the skill of the  
administrator and the number and size of the  
LUNs. During much of this time the array is  
either unusable or must operate in a degraded  
performance mode. In other words, LUNs  
cannot be utilized until they have been formatted.  
This formatting takes up a lot of the array’s  
internal resources and bandwidth. After a LUN  
has been formatted, it can be used; but as long  
as other LUNs in the array are also going  
through their format process, the entire array  
will suffer from degraded performance.  
5. Finally, before the new LUNs can be  
used, disks must be formatted, which can  
take many hours per array.  
Now, a short word about human error. Every step  
of this process has the potential for human error.  
Except in the grossest cases, errors would probably  
not result in data loss, but every miscalculation  
in this process would easily result in a decline in  
performance. Some of these declines could be  
huge. For example, miscalculating the RAID levels  
or the cache page size could severely degrade  
the array’s performance.  
configuring an array for  
a database  
Properly configuring an array for a database  
typically involves a large problem set with many  
variables. Many database administrators have  
been taught to isolate different pieces of the  
database in an attempt to optimize performance,  
availability, and recovery. This process, although  
based on sound objectives, is far too error-prone.  
This typically involves a large problem set with  
many, many variables. Unfortunately, database  
1.3  
section 1  
hp storage white paper  
the system  
administrators dilemma  
determining the number of disks required  
for the necessary performance  
creating the LUNs—note: creating LUNs  
takes only a few seconds per LUN  
1
Change is the issue. Many environments change  
over time and this makes their initial configurations  
progressively less and less optimal. The dilemma  
is that an optimum configuration today eventually  
becomes out-of-date, and typically over time,  
performance degrades in traditional arrays, but  
the administrator usually has neither the time nor  
the window of opportunity to bring the system  
down and perform the necessary reconfiguration.  
Step away from the array; the configuration is  
now complete. Every other step is automatic.  
RAID levels are automatic. The different capacity,  
number, and speeds of disk drives are automatically  
accounted for. The cache page size is automatically  
set. Even the disk formatting is performed  
automatically. Moreover, after the LUNs are set,  
the array is immediately available to accept data.  
The array does the work, not the administrator.  
Access patterns change. Different areas of the  
database become more highly used. Sometimes  
at the end of the month certain data that normally  
lies unused now becomes highly used and requires  
higher performance. All of this poses problems  
for the conscientious system administrator.  
Not only is this the initial configuration process,  
but the process is just as simple for any subsequent  
reconfigurations. In other words, if LUNs have to  
be deleted and new ones created, the process is  
just as simple. Note: As with any array, if you  
wanted to delete the LUNs but save the data,  
you would have to do a backup and restore.  
At first, the administrator notices a change in  
performance, but it is not enough to justify bringing  
the system down. However, over time, the  
performance continues to degrade until eventually  
the problem is so severe that the administrator is  
forced to bring the system down over the protests  
of the users. He must then go through all 13  
configuration steps, including binding all the  
LUNs and waiting for the reformat to complete.  
After that the cycle begins again: Performance  
initially is great, but over time it degrades  
until finally the pain increases to the point that  
a reconfiguration again becomes justifiable.  
adding capacity with  
hp virtual arrays  
HP’s Virtual Array Architecture also simplifies the  
process of adding capacity to an array. Today,  
many traditional arrays allow the administrator  
to add disks on-the-fly—in other words, to add a  
disk drive when the array is up and running and  
accepting I/Os.  
However, when the disk drive is added to a  
traditional array, it is not part of any LUN. It is  
not formatted, and it is not able to accept data.  
An administrator must go in and manually perform  
those functions. If the disk drive is to be added  
to an already existing RAID group, then the data  
in that group must first be backed-up and later  
restored to the newly created LUN that now  
contains the new disk drive. And except for the  
backup, the steps are the same as when an  
array is first configured.  
managing the hp virtual array  
Configuring an HP Virtual Array is much simpler  
than the process for configuring a traditional  
array. Remember the complex steps involved in  
configuring a traditional array? The steps involved  
in configuring the HP Virtual Array consist of:  
determining the total capacity and  
performance requirements  
for each application, determining the  
number and size of LUNs  
1.4  
section 1  
hp storage white paper  
The HP Virtual Array accepts new disks while  
the array is up and running and accepting I/Os  
as with some higher-end traditional arrays.  
However, the HP Virtual Array takes it one step  
further. Once the disk is inserted, the array  
automatically includes that disk into the existing  
disk space and stripes all LUNs across that disk.  
This means that even without the creation of any  
additional LUNs, the array performance will  
improve because of the additional available  
spindle. Only the HP Virtual Array automatically  
adds the new disks to existing LUNs. Further,  
any newly created LUNs are also automatically  
spread across all the disks in the array, including  
the additional disk.  
automating the cache  
parameters  
1
Configuring a traditional array typically requires  
setting the cache parameters such as the percentage  
of read and write cache, the size of the cache  
pages, and, in some cases, the allocation of cache  
to specific LUNs. In making these determinations,  
there is ample opportunity for error.  
With HP’s Virtual Arrays, all of this is preset and  
automatic. And this means that all the parameters  
within the array are tuned to work in unison with  
the stripe size and the array hardware. First, the  
cache is set at 80% read and 20% write, is  
shared between controllers, and is treated as a  
“pool.” Second, the cache page size is set at  
64K and is set to automatically destage to disk  
every 4 seconds whether the page is full or not.  
The 64K size minimizes the number of I/Os to  
the back-end in sequential environments and  
provides a carefully calculated balance within the  
array between the number of cache pages and the  
speed of the back-end in random environments.  
time to implementation:  
formatting the array  
As mentioned earlier, after new disks are added  
to a traditional array, it then takes several hours  
to complete the formatting of the RAID group.  
During this format phase, no data can be written  
to the new LUNs. With some implementations,  
the array is offline until all the LUNs have been  
formatted. In other implementations, I/Os can  
be written to already formatted LUNs even while  
other LUNs are going through the format  
process, although performance is very slow.  
performance  
Traditional arrays are susceptible to “hot spots” and  
to changes in the environment that make the initial  
configuration obsolete. The HP Virtual Array virtually  
eliminates these critical performance issues.  
Because executing the disk format command  
uses up so much of the array’s internal bandwidth,  
array performance is greatly reduced until all  
of the disk formatting has been completed.  
With HP’s Virtual Array Technology, the array is  
immediately available as soon as the LUNs have  
been configured. The disk formatting is done as  
the writes are done. In other words, as writes  
are sent to disk, the formatting is accomplished  
for only those blocks being written to. This means  
that while there is a small hit to performance for  
that individual write, there is very, very little  
impact on overall array performance.  
First, the HP Virtual Array is far less likely to  
experience a hot spot—in other words, it will  
almost never experience a condition where a  
few disk drives become a performance bottleneck  
in the array. Here’s why: the virtual array  
always (and automatically) stripes all of the  
LUNs across all of the disks in the RAID group.  
For example: assume a virtual array loaded with  
a total of 60 disks had 30 disks in each of its  
two RAID/redundancy groups. Every LUN in  
that group would be spread across all 30 disks.  
1.5  
section 1  
hp storage white paper  
Every LUN would have 30 spindles at its disposal.  
And don’t forget, in the virtual array the spare  
disk capacity is also spread across all of the  
spindles, i.e., there are no unused spindles in an  
HP Virtual Array.  
for more details.  
Question: Shouldn’t a good system administrator  
know which RAID level the data is in? With the  
virtual array, I have no idea if it is in RAID 5 or  
RAID 1+0.  
1
Short answer: Both the HP Surestore Virtual  
Array 7100 and the HP Surestore Virtual Array  
7400 can be configured to run in either fixed  
RAID 1+0 or AutoRAID modes. And to help with  
performance analysis, the controller can provide  
data on the actual usage of each RAID level.  
Long answer: With the HP Virtual Array, if  
your data is frequently used, it WILL be in RAID  
1+0 and will have the best performance. The  
virtual array RAID level policies were developed  
after researching the decision patterns of  
Second, the virtual array automatically performance  
tunes the array 24 hours per day, 365 days per  
year. The RAID level is matched with the workload.  
Data with access patterns that would benefit  
from RAID 1+0 storage are automatically directed  
to a RAID 1+0 section of the array. Infrequently  
used data, or data whose access patterns match  
RAID 5DP performance characteristics, are  
directed to the RAID 5DP section of the array.  
These are the same rules that a storage expert  
would use to optimize an application’s performance.  
In addition, like an expert storage administrator,  
any changes to the configuration are made only  
during low-usage periods or when disks are  
added. Note: When disks are added, the  
existing LUNs are automatically extended across  
the new disks. This normally requires data  
movement within the array. However,  
administrators have the option to postpone this  
data movement by simply turning off the array’s  
“Auto-Include” feature.  
experienced  
system  
and  
database  
administrators. Almost always, technology  
progresses from highly manual to highly  
automated operations. In almost every case, the  
developers of automation simply replicate the best  
of the already developed manual processes.  
This is exactly what HP did with the HP Virtual  
Array. Also please note: The Virtual Array  
policies are improved over those of the Model  
12H. The virtual array really does strive in almost  
every instance to do any background data  
movement during periods of low array activity.  
faqs  
Question: Isn’t this virtualization technology  
new and doesn’t that make it risky and untried?  
Question: On which disks is my data kept? I  
suppose the virtual array can be trusted to handle  
the RAID level decisions, but I also need to know  
on which disks my indexes and redo logs are  
kept. The way the virtual array moves data and  
stripes across all the disks means I have no idea  
where my data is kept.  
Answer: It would be new for HP’s competitors,  
but HP has been virtualizing arrays since 1995,  
and since then HP has sold over 20,000 virtualized  
arrays. The technology is neither new nor risky  
and it offers a simple and compelling value  
proposition, i.e., great ease of management and  
great data protection. For arrays, management  
costs are far more costly than the initial purchase  
price. The HP virtual array is the easiest array in the  
world to configure, add capacity, and manage—  
the administration savings are significant. It also  
has the best availability of any mid-range array  
on the market. See the white paper titled  
VA7100 Hardware High Availability Features”  
Answer: First, if you are worried about the  
integrity of your data, the HP Virtual Array does  
a combination of things to protect your data that  
no other array does. First, the HP Virtual Array  
offers end-to-end checksum, ECC protected  
memory, parity coherence, disk scrubbing, and  
RAID 5DP which, in a typical configuration,  
1.6  
section 1  
hp storage white paper  
gives 100X the data protection of traditional  
RAID 5 and 10X the data protection of RAID 1.  
Second: The HP Virtual Array stripes across all  
the disks in the redundancy group. This involves  
more disks but is not so different from any other  
RAID 1+0 implementation. Third: array striping  
is the way of the future. It’s faster and safer.  
environment changes. HP’s Virtual Array  
Architecture tunes the array automatically, 24  
1
hours per day. It is the hands-down winner in  
real-world performance. A better analogy would  
be to compare the multiple manual processes  
required to set the type, load the paper, and  
actually print books on the old-style printing  
presses with those of the automated printing  
systems of our generation. Obviously, automation  
in printing presses adds to greater performance.  
It works the same for arrays.  
Question: Doesn’t all this “behind the scenes”  
movement of data require a huge performance  
hit? Surely, the trade-off for virtual technology is  
slow performance.  
Question: Wouldn't striping the data across  
such large RAID groups make the disk rebuild  
times very long and take up a high percentage  
of the array resources?  
Answer: Not true. For years, HP’s original  
virtual array, the AutoRAID Model 12H, was  
used for HP’s V-class TPC-C benchmark tests.  
And today, HP Virtual Array products have been  
used for the new rp8400 benchmark testing and  
are scheduled to be used for the Superdome  
TPC-C tests. Remember, the HP Virtual Array  
mimics the policies of experienced system  
administrators. Would an experienced system  
administrator do a reconfiguration of the array  
during a period of high workloads? No! And  
neither would the HP Virtual Array. The HP Virtual  
Array policy is to NOT perform the background  
tuning operations when the array is under a  
greater than 60% workload.  
Answer: Because of HP’s RAID 5DP this is  
practically a non-issue for HP Virtual Array  
products. The HP Virtual Array’s RAID 5DP gives  
each redundancy group 10x the protection of  
RAID 1. RAID 5DP requires that 3 drives would  
have to fail before there would be data loss.  
Thus, even AFTER a drive fails, the data is STILL  
protected with the same degree of protection as  
standard RAID 5. Finally, this means that a single  
drive failure does not put the data at risk and  
therefore does not require an emergency  
rebuild. The HP Virtual Array can take the time  
and do the rebuild in the background without  
impacting incoming I/Os. Also, the HP Virtual  
Array will rebuild the RAID 1+0 data first since  
that is the most vulnerable after a failure. In all  
cases, data integrity is ensured and performance  
is preserved. No other array can make this claim.  
Question: Isn’t manual always better than  
automatic? In cars, manual transmissions give  
better performance than automatic transmissions  
because they give the driver more control over  
performance; likewise wouldn’t manual RAID  
configurations be better than HP’s Virtual Array  
Technology and its automatic RAID configurations?  
Question: If your environment is totally stable,  
wouldn’t a manual configuration by an experienced  
system administrator result in better performance  
than one derived from the virtual array’s policies?  
Answer: Manual transmissions in cars would  
not give better performance if drivers were only  
allowed to shift gears once and could never  
change them after that. In essence, this is what  
you have with traditional disk arrays. You are stuck  
with the initial configuration unless you bring the  
array down and go through a time-consuming  
and complex reconfiguration every time the  
Answer: Let’s first admit it: Totally stable  
environments are rare. But the answer is that if  
the environment were totally stable, and if the  
administrator configured the array absolutely  
1.7  
section 1  
hp storage white paper  
accurately, the traditional array and the HP Virtual  
Array would have similar performance.  
However, the HP Virtual Array would still have  
the performance advantage because of its ability  
to efficiently load balance across all the disks in  
the array both at the initial configuration and  
after capacity growth. In both stable and changing  
environments, the HP Virtual Array is the right  
choice. Now, if the environment is not totally stable,  
then the HP virtual array is the hands-down winner.  
When both arrays are initially configured, the  
traditional array should be faster at least for  
a day or so, but after that the HP Virtual Array  
will catch up and continue to operate at peak  
performance for as long as the array is plugged  
in, while the traditional array will get slower  
over time. This is how it works. When the HP Virtual  
Array is first turned on, it doesn’t know which  
data needs to be in RAID 1+0 and which data  
needs to be in RAID 5. After a day or so of reading  
the access patterns, it will figure out which RAID  
level is best suited for which data. Once that  
happens, the performance will be as good as in  
any manually configured array. And, don’t forget,  
if the environment should change after that, even  
slightly, the HP Virtual Array will adjust while a  
manually configured array can do nothing but  
keep plugging along in a degraded mode.  
Consider: It is always easier to manage a  
smaller number of things than a larger number of  
1
things. At first, bits were grouped into bytes.  
When there were too many bytes to keep track  
of, they had to be grouped into blocks. When  
there were too many blocks to manage, someone  
invented volumes and volume managers. What do  
you think the next logical step is when there are too  
many volumes?  
Answer: HP’s Virtual Array.  
summary  
HP’s Virtual Array with HP’s patented Virtual  
Technology is the industry’s most intelligent disk  
array. Because of its unique ability to greatly  
simplify storage management, it significantly  
reduces overall IT management costs and practically  
eliminates mistakes caused by human error. HP’s  
Virtual Technology also has a significant positive  
impact on real-world performance by automatically  
eliminating “hot spots,” and by performance tuning  
the array 24 hours per day, 365 days per year.  
for more information  
For additional information on HP Virtual Arrays  
and other HP storage products and solutions,  
please call your local HP sales representative or  
Question: Are you saying that this technology  
is for everyone? I’m not looking for anything new.  
Answer: If it’s new to you, then it is understandable  
that HP’s virtual technology might seem unnecessary.  
After all, you’ve succeeded without it up till now.  
But sometimes improvements do come along that  
really do hold out the promise of a widespread  
advantage. The dilemma of our rapidly advancing  
technological age is to know which new technologies  
should be ignored and which are the ones that  
need to be grasped.  
1.8  
section 2  
“hp extends virtualization to the array”  
—by D.H. Brown Associates, Inc.  
virtual value in hps va7000 series  
2
Virtualization is a key to solving complexities associated with storage configurations. As Information  
Technology continues its non-stop evolution, tools and devices that improve manageability while increasing  
business efficiency will command a premium in the marketplace. HP has taken a significant step forward in  
delivering its Virtual Array (VA) 7000 series with a full complement of virtualization functions. Enterprises that  
value a return on investment in areas such as data-sharing, optimized self-tuning performance, and data-access  
flexibility must evaluate HP’s Virtual Array offerings.  
SAN virtualization is in effect the aggregation of multiple storage devices in a centralized  
configuration, a storage pool. This simplifies the storage administrator’s role by enabling all resources to be seen  
as a single entity. SAN virtualization by itself, however, offers little assistance in efficiently managing requirements  
of single applications. It pays scant attention to the capacity and performance dynamics of individual,  
heterogeneous devices within the storage pool. HP, in its Federated Storage Area Management (FSAM) strategy,  
acknowledges the importance of SAN virtualization. In its VA7000 Series, HP extends these concepts and  
benefits to the device-array level. Together, they demonstrate the full advantage of interdependencies between  
managing SAN virtual storage pools and managing virtual storage at the device level.  
HPs VA7000 Series Characteristics  
Feature  
VA7100  
VA7400  
Capacity  
1 TB (raw) in 3U enclosure;  
14 TB in 2M rack  
7.6 TB (raw) using the Disk System (DS)  
2400 (up to six DS2400s chained);  
14 TB in a 2M rack of 73 GB disks  
LUNs (max)  
128  
1,024  
Disk Drives Supported  
15 per 3U enclosure;  
same disk support as 7400  
Up to 105 drives with DS2400; 18 GB 15 K rpm;  
36 GB 10 K and 15 K rpm and 73 GB 10 K rpm  
I/O ports  
1 Gb/sec. FC host and disk ports  
Two 2 Gb/sec. FC host ports;  
two 1 Gb/sec. FC disk ports  
Cache  
256/512/1024/ MB mirrored  
512/1024 MB mirrored per controller  
Transfer Rate  
90 MB/sec. 12,000 I/Os per sec. from  
cache; 3000 I/Os per sec. back-end  
160 MB/sec.; 28,000 I/Os per sec. from cache;  
7500 I/Os per sec. back-end  
In the HP Surestore VA7000 family, “virtual array” refers to treating the disks within the array as a pool of  
storage blocks rather than as physical disks. With this technology, the VA7000 family emerges as a second  
tier of storage virtualization. Extending virtualization to the device level, the SAN’s view of storage becomes  
independent of physical disk attributes. This characteristic reduces the cost associated with SAN virtualization  
by simplifying storage management, enabling more efficient use of available device capacity, and decreasing  
the amount of required data movement at the device level.  
D.H. Brown Associates, Inc.  
Our research program in Enterprise Storage Solutions (ESS) makes this Technology Trends available to all our  
subscribers. Those interested in this program should contact [email protected] or 914-937-4302, ext. 281  
2.0  
section 2  
“hp extends virtualization to the array”  
—by D.H. Brown Associates, Inc.  
VA7000 virtualization, managed at the array  
level, offers the advantage of providing greater  
granularity when managing a pool of logical  
storage. Logical Unit Name (LUN) capacity is  
definable as a logical set of storage blocks rather  
than as a specific set of predefined physical disks.  
Each LUN becomes scalable from a few  
megabytes to several terabytes by adding  
capacity without operator intervention (the array  
automatically controls the addition of new  
capacity to the storage pool).  
This “self-tuning” capability automatically  
restripes data across new disks added to the  
array, minimizing any potential imbalances of  
data distribution within the VA7000 series.  
Redistribution occurs in the background with no  
impact on SAN bandwidth or server performance.  
This function significantly reduces the efforts  
required by the system administrator to keep  
storage resources in balance.  
super redundancy  
enhances RAID  
2
As alternatives to the “self-tuning” LUN-management  
capability described above, the VA7000 series  
also enhances LUN performance through its use  
of RAID configurations “redundancy groups.”  
Many variations to RAID configurations have  
surfaced over the past 10 years. Most provide  
moderate variations on the standard RAID 1  
(mirroring) and RAID 5 (striping) techniques.  
In its VA7000 Series, HP delivers a variation that  
substantially adds to its availability characteristics.  
RAID 5DP (Double Parity) approaches a RAID 6  
solution in that it enables recovery from simultaneous  
failures in two disks without loss of any data.  
While this might seem superfluous in RAID arrays  
of only a few disks, it becomes critical in  
configurations supporting a large number of  
disks in a virtual array. The potential for disruption  
due to disk failure grows with the number of  
disks involved in an array. The goal of virtualization  
is to remove the physical constraints, which  
enables dozens of disks to be used in a single  
virtual array. HP has addressed this concern  
with its RAID 5DP solution, increasing data  
availability by two orders of magnitude over  
traditional RAID 5 implementations. Additionally,  
use of storage is made more efficient as very  
large RAID groups (up to dozens of disks) are now  
manageable using only two parity disks rather  
than a parity disk for each group of five disks.  
Also with VA7000 virtualization, the number of  
LUNs is no longer restricted by the amount of  
physical storage. Administrators can configure  
LUNs up to the limit permitted by the architecture  
without concern for physical disk groupings  
within arrays.  
Disk arrays without virtualization support LUNs  
only from contiguous free space, requiring allocation  
of large storage chunks even if only a small portion  
is necessary. As changes occur over time, data  
must be moved or storage reconfigured to recover  
unused space.  
Within the VA7000, fragmentation of physical  
space is managed at the RAID (Redundant  
Arrays of Inexpensive Disks) block level rather  
than in disk-sized chunks. The impact of free-space  
management and “garbage collection” is  
absorbed at the array level, without consuming  
SAN or server resources.  
Technology Trends  
Copyright © 2001 by D.H. Brown Associates, Inc.  
2.1  
section 2  
“hp extends virtualization to the array”  
—by D.H. Brown Associates, Inc.  
reliability characteristics of hps  
va7000 series  
(consisting of all the even-numbered disk slots). Data  
movement and failover occur in the same manner  
as in the VA7100.  
2
redundant, hot swappable controllers, fans,  
power supplies, and an internal fibre-channel  
hub eliminates planned downtime  
The variation in data management at the array level  
provides great flexibility in matching performance  
to application need. Self-tuning enables immediate  
response to data and application variables  
ensuring more work completed per transaction  
and maintaining more consistent service-level  
performance. RAID 1+0 uses a large number of  
disks in a single configuration to deliver strong  
performance to selected LUNs. It does so, however,  
at the expense of additional space for mirroring  
the data. RAID 5DP incurs a performance penalty  
for write operations and can affect system  
performance if frequent updates to large volumes  
of data are required. Its advantage is in the data  
redundancy provided and the improved cost  
efficiency vs. mirroring. Dual-redundancy groups  
in the VA7400 enhance the efficiency of data  
movement through the controllers, improving  
system performance.  
online firmware and capacity upgrade  
using mixed capacity disk drives  
Checksum algorithm used for end-to-end  
data and data-path protection against  
data corruption  
double parity supports recovery from two  
simultaneous disk failures (RAID 5DP)  
error Correction Code (ECC) protects mirrored  
cache from corruption before writing to disk  
three-day battery backup protects user data  
in cache from power outages and disasters  
phone-home capability is provided  
through hp’s Predictive Support programs  
and Event Monitoring Services (EMS)  
“Redundancy groups” are physical disks that contain  
redundant (copies of) data, defined by the RAID  
type assigned to the data. Each redundancy  
group is divided into LUNs, addressable by a host.  
The VA7100 supports only one “redundancy  
group” with each of the 15 disk drives in the  
enclosure addressable by both controllers with  
no physical restrictions. The two fibre-channel  
loops (one to each controller) include a transmit  
line and a receive line for the transfer of data to  
and from the controller. Controller-to-controller  
communication and loop failover occur via an  
internal N-way bus.  
The VA7400 supports two redundancy groups in  
which one controller manages access to one redundancy  
group (consisting of all the odd-numbered disk  
slots) and the other controller to the second group  
Technology Trends  
Copyright © 2001 by D.H. Brown Associates, Inc.  
2.2  
section 2  
“hp extends virtualization to the array”  
—by D.H. Brown Associates, Inc.  
software extends virtual functions  
2
While hardware design and functionality are critical to enabling performance, software function can maximize  
the efficiency and flexibility of the device. HP’s software packages extend the management and flexibility of  
the VA7000 family while exploiting the virtualization capability.  
HPs Software for VA7000 Series  
Feature  
VA7100  
VA7400  
CommandView SDM  
HP-UX with all PA-RISC computers;  
Windows and Red Hat Linux with Intel  
HP-UX with all PA-RISC computers;  
Windows and Red Hat Linux with Intel  
Pentium III/500 MHz computers or above Pentium III/500 MHz computers or above  
Enterprise Management  
Smart Plug-ins  
HP-UX with all PA-RISC computers;  
Windows and Red Hat Linux with Intel  
HP-UX with all PA-RISC computers;  
Windows and Red Hat Linux with Intel  
Pentium III/500 MHz computers or above Pentium III/500 MHz computers or above  
HP-UX with all PA-RISC computers; HP-UX with all PA-RISC computers;  
Windows and Linux with Intel Pentium III/500 Windows and Linux with Intel Pentium III/500  
MHz computers or above; supports 128 MHz computers or above; Managed by  
World-Wide Names and 128 secure LUNs Command View SDM; supports 128 World-  
Wide Names and 1,024 secure LUNs  
Secure Manager VA  
Business Copy VA  
Auto Path VA  
HP-UX with all PA-RISC computers;  
HP-UX with all PA-RISC computers;  
Windows and Linux with Intel Pentium III/500 Windows and Linux with Intel Pentium III/500  
MHz computers or above; supports 127  
Business Copy LUNs  
MHz computers or above; Managed by  
Command View SDM; supports 1,023  
Business Copy LUNs  
Windows 2000 with Intel Pentium III/500  
MHz computers or above  
HP-UX with all PA-RISC computers;  
Windows and Red Hat Linux  
Network Management Tools Not Applicable  
HP OpenView NNM, HP Toptolls, CA  
Unicenter, Tivoli NetView, BMC Patrol  
Operating Systems  
HP-UX 11.0; Windows NT/2000; Red Hat  
Linux 6.2  
HP-UX 10.2, 11.0, 11i; Windows  
NT/2000; Solaris 2.6, 7.0, 8.0; AIX  
4.3.3; NetWare 5.0/5.1; MPE/iX 6.5,  
7.0; Red Hat Linux 6.2, 7.0  
Technology Trends  
Copyright © 2001 by D.H. Brown Associates, Inc.  
2.3  
section 2  
“hp extends virtualization to the array”  
1
The major HP Surestore software packages (identified on page 2.2) include,  
2
Command View SDM speeds and eases local and remote management, including install, configure,  
monitor, and control via a web browser, direct host attach, or SAN. HP Predictive Support programs  
and Event Monitoring Service (EMS) provide phone-home capability.  
Smart Plug-ins allow HP Command View SDM to integrate and process storage events via SNMP with  
HP OpenView Network Node Manager (NNM).  
Secure Manager Virtual Array (VA) controls server access to data on a LUN-by-LUN basis, even in  
simultaneous heterogeneous array sharing environments.  
Business Copy Virtual Array (VA) supports LUN replication for development, testing, or backup situations.  
Data replication takes place at a specified point in time while the system is operating and entirely within  
the storage array, reducing the potential for I/O bottlenecks.  
Auto Path Virtual Array (VA) routes data around any path’s failure to maximize data availability. It provides  
the ability to self-configure and to automatically manage multiple I/O paths. It also provides dynamic  
load balancing to ensure peak performance.  
This document is copyrighted © by D.H. Brown Associates, Inc. (DHBA) and is protected by U.S.  
and international copyright laws and conventions. This document may not be copied, reproduced,  
stored in a retrieval system, transmitted in any form, posted on a public or private website or  
bulletin board, or sublicensed to a third party without the written consent of DHBA. NO copyright  
may be obscured or removed from the paper. All trademarks and registered marks of products and  
companies referred to in this paper are protected.  
This document was developed on the basis of information and sources believed to be reliable. This  
document is to be used “as is.” DHBA makes no guarantees or representations regarding, and shall  
have no liability for the accuracy of, data, subject matter, quality, or timeliness of the content.  
1
CA Unicenter TNG and BMC Patrol are supported in addition to the HP OpenView Storage Node Manager, OpenView Allocator, Builder,  
Optimizer, and Accountant.  
Technology Trends  
Copyright © 2001 by D.H. Brown Associates, Inc.  
2.4  
06/02  
5981-2067ENUC  

National Instruments 6211 User Manual
Magnavox AZ 1103 User Manual
LG Electronics VX4500 User Manual
LG Electronics 5450 User Manual
HP Hewlett Packard 302400 User Manual
Grindmaster 70753 C User Manual
Fujitsu ServerView Resource Orchestrator Cloud Edition J2X1 7611 03ENZ0 User Manual
Freestyle FS80934 User Manual
Dell PowerVault MVT01 User Manual
Black & Decker DCM1350 User Manual