Smarter Data Storage.PDF

Mesabi Group
Commentary
© 2012 Mesabi Group LLC Telephone: 781 326 0038
26 Country Lane Fax: 781 326 0038
Westwood, Massachusetts 02090 www.mesabigroup.com
July 5, 2012
IBM Smarter Storage: What a Smart Idea
Getting smarter is an imperative for all components of the information infrastructure including
storage. That imperative is so because meeting the challenges such as ongoing explosive data
growth with always tight IT budgets is not possible otherwise. Increasing software-based intelli-
gence, i.e., getting smarter, is the way out of the trap. On the storage side, IBM calls this Smart-
er Storage. IBM exemplifies Smarter Storage through three pillars: efficient by design to manage
cost and capacity growth, self-optimizing for improved performance and productivity, and cloud
agile to use storage effectively in the new cloud models.
What Smarter Means
The physical world, both the live (such
as people and agricultural plants) as
well as the inanimate (including natural
resources, electricity, and buildings)
face innumerable ongoing challenges,
such as improved health care, in-
creased crop yields, and more efficient
(including green and economic cost
savings) for the resources we use.
On top of that, the logical world, such
as the content-rich Internet and the
explosion of connected devices,
prevents a number of challenges of
which keeping up with change is one.
IBM refers to this as the need for a
Smarter Planet that has an infrastruc-
ture that is instrumented, intercon-
nected, and intelligent.
While this infrastructure has many
components (such as highways as part
of a mobility and factories as part of a
production infrastructure), central and
indispensable to all this is the infor-
mation infrastructure. While technolo-
gies, such as biotechnology, materials
science and computer science are
invaluable and requisite to the process,
the information infrastructure is the
integrating glue that permeates
everything symbiotically
And for the information infrastructure
to deliver on the countless possibilities,
promises, and opportunities that
present themselves, it has to become
more intelligent, which simply means
that the infrastructure has to make
software-based decisions and take
actions that enable the infrastructure
to deliver desired outcomes better and
more efficiently than can be done
manually alone. The process of getting
more intelligent is called smarter.
Smarter is a process that has been
going on for some time and will
continue on into the future.
Why Smarter Matters
IT budgets are likely to remain flat (or
nearly so) for some time because of

IBM Smarter Storage: What a Smarter Idea
Page 2
Commentary
uncertainties about world economic
conditions. This means that IT adminis-
trator headcounts are also likely to
remain flat. Yet continued explosive
data growth as well as the number of
producers of data (including sensor-
based) as well as increased number of
consumers (those who use, i.e.
consume, the newly-minted infor-
mation) leads to the famous do more
with less dictum.
But that is not all. The possibilities for
taking advantage of the multitude of
benefits that the infrastructure can
deliver remain grossly underdeveloped
for two reasons: 1) more IT resources
need to be diverted to innovation from
operations, but that can only happen if
the smarter infrastructure can save
more administrator time than just
having them tread water so to speak
and 2) while innovation provides
direction, the ability to leverage and
make real those directions requires
software intelligence that can enable
the infrastructure to take the necessary
actions. Smarter matters therefore if
the information infrastructure is to be
exploited to the needed degree.
Enter Smarter Storage
So how does Smarter Storage fit into
the pantheon of smarter and why does
it matter? Now IBM views Smarter
Computing as providing the IT infra-
structure that is needed for a Smarter
Planet and that Smarter Storage is
encapsulated within Smarter Compu-
ting. IBM has three basic tenets for
Smarter Storage: efficient by design,
self-optimizing, and cloud agile. But
before those three tenets are explored
in depth, a little context will help to
envision what Smarter Storage is all
about.
Among other things, Smarter Storage
had the following characteristics: data
centric and not media-centric, have a
confluence of both storage management
and data management, and focused on
the storage information infrastructure
itself.
Data Centric, Not Media Centric
Smarter Storage leads to storage
services — such as snapshotting and
auto-tiering — being viewed as data-
centric storage services rather than
media-centric storage services. The
traditionally media-centric worldview is
limited to the physical characteristics of
disks, tapes, and arrays (i.e., a feeds
and speeds view) whereas the new
data-centric paradigm recognizes the
rapidly changing requirements of
applications/information in a world
that is increasingly virtualized.
One result is the ability to encapsulate
a service level agreement (SLA) with the
data in a virtual volume using IBM’s
SmartCloud Virtual Storage Center (aka
Storage Hypervisor). With storage
virtualization, data is not locked into a
single piece of storage media, but is
mobile. Think of storage virtualization
as an honest “shell game” where the
data “pea” may move — such as from
one array to another or within tiers of a
single array — based upon business
requirements. Think data centric, not
media centric.
Confluence of Both Storage Management and Data
Management
Storage management (aka storage
resource management) discovers,
monitors, and controls storage assets
whereas data management is the non-
data-path control of data itself from
creation to deletion, such as migration,

Page 3
Commentary
replication, and backup/restore
processes. That is the traditional way of
distinguishing the two (for example a
storage administrator would deal with
provisioning an array whereas a
backup administrator would configure
backups). However, that distinction has
been blurred with the fluidity of data,
such as the auto-tiering within tiers of
storage within an array (when data is
“hot” it is on one tier and when it is
“cold” it is put on another tier).
Think instead of range (vertical and
horizontal) and breadth (category of
storage services). Range has two
dimensions — horizontal (the hierar-
chal range of tiers that say IBM’s Easy
Tier can manage) and vertical (e.g., a
single logical pool of storage that spans
multiple arrays that IBM SmartCloud
Virtual Storage Center manages).
Breadth refers to the type of storage
services delivered. These can include
capacity services (such as Real-time
Compression of primary storage and
thin provisioning), data protection
services (such as those related to
backup/recovery and disaster recov-
ery), performance services (such as
ensuring that data gets the IOPS that it
needs now), and data management
services (such as those involving data
movement and data deletion).
Focused on Information Infrastructure Itself
Providing a smarter storage infrastruc-
ture is an IT internally-focused
responsibility. Think of IT as building
multiple, better, and distributed storage
engines and storage “fuel” tanks.
Although the business benefits
indirectly from how well this is done
and the end result should be IT as a
service, IT has the responsibility for
making it happen and continuing to
manage the storage infrastructure.
But what about self-service portals?
Now “consumers,” which here could
mean business users as well as IT
professionals, such as DBAs. can
choose from services provided in a
service catalog (as in a “cloud”).
However, designing, configuring, and
exposing these services, is the respon-
sibility of IT specialists who have the
needed skill sets. The heavy lifting has
been done so that the consumer can
select from a controlled and limited
pallet of services, such as gold, silver,
and bronze levels of backup/recovery
data protection or for self-provisioning
different levels of storage.
Taking a Holistic View of Smarter Storage
IT has to think with a big picture or
holistic mindset about smarter storage.
That starts with taking a data centric
rather than media centric point of view.
And although point products and
services are essential, all of the pieces
of the puzzle require IT to think of
range (vertical and horizontal) and
breadth of services. This is a broad and
not a narrow IT responsibility so it has
to be a cross-discipline effort within IT
and not just given to a few specialists.
With that as a little context, let’s
explore the three tenets (or principles)
that IBM uses to flesh out the concepts
of Smarter Storage.
Efficient by Design
This tenet reflects IBM’s ongoing
commitment and deliberate focus to
extract more and more efficiency from
the storage infrastructure through
hardware/software product design.

Page 4
Commentary
One view of efficiency is technical and
one is business-focused. The technical
revolves about the least waste of
resources or, said slightly differently,
making the most of a resource (such as
storage space) without having to
expend more of other resources (such
as time or money) to achieve the
improved (i.e., more efficient results).
The business view and impact of a
more efficient design is that the new
storage platform is more economical to
use than the previous version.
Improving storage utilization is a
classic illustration of employing the
tenet of efficient by design as a part of
Smarter Computing. Thin provisioning
as provided by the IBM System Storage
SAN Volume Controller (SVC) as a
foundational product enables the
provisioning of storage on a logical
rather than a physical basis. The net
result is that planning and sizing disk
storage is simpler and less expensive as
opposed to the traditional fixed
allocation of disk space by applications
which leads to never used disk space.
An important emerging example for
improving storage utilization is Real-
time Compression for primary storage,
which squeezes data so that an
enterprise can double or more the
effective usable capacity from its
nominal (i.e., stated) capacity. IBM has
just made Real-time Compression
available on the IBM Storwize V7000
and SAN Volume Controller.
Now if storage utilization techniques
are used on an existing storage
investment, there is no immediate
economic benefit. However, because of
the more efficient management of
storage growth, the economic benefit is
deferred expenditures for storage and
reduced expenditures for future growth.
But here are also a couple of other
business benefits. Each storage
administrator has to do less work (thin
provisioning is a lot less time consum-
ing and easier task than traditional
“thick” provisioning) and can manage
more storage with the same amount of
effort. The ratio of administrator cost to
physical storage costs has improved
significantly and, over time, this
reduces operating expenses.
Now these savings are not likely to
reduce next year’s IT budget, but rather
free up money that can be applied to
the important innovative tasks that the
business needs, but IT could not do.
Illustrating Efficient by Design
 Real-time Compression — the
Storwize V7000 and SAN Volume Control-
ler can now squeeze out the redundant data
“water” in an array. Real-time Compression
is unique in the market in that it is designed
for active primary workloads, compressing
and decompressing data in real time. This
makes real-time compression practical and
would seem to give IBM an important
competitive differentiator.
 Tivoli Productivity Center (TPC)
Enhanced Reporting — TPC now inte-
grates with Cognos reporting and modeling.
This is important as Smarter Storage not
only uses software intelligence on storage
itself, but also uses software smarts to
provide better information to IT administra-
tors to manage the storage service infra-
structure.
Self-optimizing
During its lifecycle, a piece of data may
have a brief burst of time where it is
latency-sensitive (i.e. “hot”) as users

Page 5
Commentary
access it and manipulate the data. The
IOPS that are necessary to make this
piece of data (and many of its closest
friends) available for timely-use is likely
to be high. However, for the majority of
the life of most data, access to data can
range from periodic, to ad hoc, or to
never. The challenge is to make sure
that data during its latency-sensitive
time is available on high relative IOPS
delivering storage media, but that when
the latency-sensitive need has passed,
the data resides on less IOPS delivering
media that are more relatively cost-
efficient than high IOPS media.
IBM does this through the use of tiers
of storage that can include tier 0 SSD
(i.e. NAND flash memory), tier 1 FC or
SAS drives, and tier 2 SATA drives; IBM
also achieves this by using advanced
algorithms to evenly and automatically
distribute data across the array with
the XIV Storage System.
Moving the data manually is an
impossible task so the task has to be
software-driven auto-tiering. The
process has to be self-optimizing.
Although an administrator prescribes
some parameters, essentially software
uses algorithms and heuristics to make
the decisions that place data with the
best balance that it can achieve
between meeting the need for high
performance when it is needed with the
cost of the overall storage solution.
Illustrating Self-optimizing
 Auto-tiering using Easy Tier — this is
available now on the IBM System Storage
DS8000, Storwize V7000, and SVC. Easy
Tier will also be able to access flash cache
on servers as part of IBM’s “flash is every-
where” focus. This capability was an-
nounced as a formal Statement of Direction
(SoD) along with plans for Easy Tier to
become application-aware. Applications are
likely to know more about what they will be
needing in terms of data than can be derived
by caching algorithms alone. The bottom
line is better balancing performance needs
with the need to control costs on a single
array (remember horizontal).
 IBM SmartCloud Virtual Storage
Center — in a statement of direction, IBM
promises to introduce software that enables
automated movement and placement of data
across (remember vertical) storage systems
to better optimize price, performance, and
access needs.
Cloud Agile
Cloud is shorthand for transforming
the IT information infrastructure so
that the infrastructure (for example,
storage) is presented as a series of
services that the consumer can choose
from (i.e. IT-as-a-service). To the user
the services (such as for storage) scale
on demand with service level manage-
ment that includes reporting on what is
consumed and how much it costs.
However, the consumer only sees the
surface of the cloud “duck” and not the
providers (either IT or a third party)
paddling like mad under the surface
out of sight. The providers must deliver
a cloud that is adaptable to change (as
in a virtualized world the data may be
highly mobile for example), automated
(so the provider doesn’t have to work
(paddle) as hard, which means
increased productivity), and robust
service level metrics (uptime, response
time performance, and data accessibil-
ity). The code word for managing all
this dynamism and flexibility is “agile;”
hence cloud agile.
IBM broadly defines the clouds that it
supports as compute clouds or storage

IBM Smarter Storage: Shat a Smarter Idea
Page 6
Analyst Name: David Hill
Topic Area: Storage
Mesabi Group LLC
26 Country Lane
Westwood, MA 02090
www.mesabigroup.com
This document was developed with IBM funding. Although the
document may utilize publicly available material from various
vendors, including IBM, it does not necessarily reflect the
positions of such vendors on the issues addressed in this
document.
Phone: (781) 326-0038
email the author: davidhill@mesabigroup.com
The information contained in this publication has been obtained from sources Mesabi Group LLC believes to be reliable, but is not
warranted by Mesabi Group LLC. Commentary opinions reflect the analyst’s judgment at the time and are subject to change without
notice. Unless otherwise noted, the entire contents of this publication are copyrighted by Mesabi Group LLC, and may not be
reproduced, stored in a retrieval system, or transmitted in any form or by any means without prior written consent by Mesabi Group
LLC
Commentary
clouds. A compute cloud integrates
servers, networking, and storage
along with the applications. A
storage cloud’s sole purpose is to
provide storage. One variant is an
application centric storage cloud
where files and objects are adminis-
tered, secured, and self-provisioned
by an application, A standalone
storage cloud is one where the
consumer self-provisions his/her
requirements and is billed for
services rendered.
Illustrating Cloud Agile
 Compute Cloud — Tight integration
with hypervisors such as VMware,
PowerVM, Hyper-V, or KVM delivers an
integrated server, storage, and network-
ing platform that can serve as the heart of
a compute cloud. Mid-range customers
can achieve their cloud goals by combin-
ing BladeCenter for Cloud Flash with
Storwize V7000 and Easy Tier, while
enterprises can achieve theirs with XIV
and Flash Cache.
 Application-centric Storage Cloud
— IBM offers Scale Out Network At-
tached Storage (SONAS) and Storwize
V7000 Unified to serve as the foundation
for this type of storage cloud.
 Stand-alone self-service storage
cloud portal — IBM has announced
intentions for SONAS and Storwize
V7000 Unified capability to be available
to support this type of storage cloud in
the fourth quarter of 2012 (and is now in
Beta).
Conclusions
Business as usual is not possible for
the information infrastructure,
including storage. On the storage
side, the continued explosion of data
accompanied by tight IT budgets
cannot be totally offset by the impact
of Moore’s Law on the server side and
similar effects on the storage side.
From a storage perspective therefore,
IBM’s Smarter Storage offers a
perspective on how to deal with this
situation that not only manages the
entire storage services infrastructure
more efficiently and productively, but
may even be able to free up some
scarce IT resources to devote to
much need innovative activities.
But IT has to take a deep breath and
view the storage infrastructure as a
whole rather than as a task specific
exercise (such as storage provision-
ing and backup recovery). That
requires putting together a storage-
services perspective of the infrastruc-
ture that is data centric and not
media centric as well as covering a
spectrum of services on not only a
single array, but also spanning the
entire storage infrastructure. At that
point IT can apply the three IBM
tenets for Smarter Storage — efficient
by design, self-optimizing, and cloud
agile — and start reaping the benefits
of the Smarter Storage approach.
David Hill

Smarter Data Storage.PDF

More Related Content

What's hot (20)

Viewers also liked (10)

Similar to Smarter Data Storage.PDF (20)

Smarter Data Storage.PDF