1. SAP Thought Leadership
Business intelligence
SAP® In-MeMory CoMPutIng teChnology
Changing the Way Business intelligenCe is Managed

2. saP in-Memory Computing technology enables real-time computing by bringing together
online transaction processing applications and online analytical processing applications at
a low total cost. Combining the advances in hardware technology with saP in-Memory
Computing empowers the entire business – from shop floor to boardroom – by giving
real-time business processes instantaneous access to data. the alliance of these two
technologies can eliminate today’s information lag for your business.

3. Content
3 about the authors 12 The Potential of Metadata
and Master Data
4 Executive Summary 12 Cost impact for Metadata and
Master data
5 Introduction 13 Cost-effective deployment
5 Where does in-Memory Come of in-Memory Computing
into the Bi Picture?
5 Business examples that leverage 14 People, Processes, and
in-Memory Computing Structures in Business and IT
14 Bi governance
7 In-Memory Computing 15 Cost impact for People,
Technology Processes, structure,
and technology
8 Business Intelligence
16 BI Data Lifecycle Management
9 Project Portfolio Management 16 Cost impact for Bi data
9 Cost impact for lifecycle Management
Portfolio Management
17 Comparing Architectures
10 Enterprise Architecture
10 a Clean, spare architecture 18 Outlook
11 Cost-effective impact of 18 For More information
enterprise architecture
About the Authors
Erich Schneider is principal enterprise architect in the Business transformation
services group of saP® Consulting and has been implementing saP solutions
for over 20 years. he was chief architect in the internal saP program to enable
data and business intelligence (Bi) architecture using saP in-Memory Computing
technology across all saP ag lines of business worldwide.
Raghav Jandhyala is a solution expert in solution management for banking ana-
lytics. he is an saP in-Memory appliance (hana) software champion for the
americas and provides guidance and support for hana-related activities in finan-
cial services. as a development architect for banking and retail analytic applica-
tions at saP, Jandhyala guided the integration of saP BusinessObjects™ solutions
with price optimization software and developed the Bi strategy with respect to
saP in-Memory Computing.

4. exeCutIve SuMMAry
enaBling Real-tiMe COMPuting
saP® in-Memory Computing technology enables real-time
computing by bringing together online transaction process-
ing (OltP) applications and online analytical processing
(OlaP) applications at a low total cost. Combining the
advances in hardware technology with saP in-Memory
Computing empowers the entire business – from shop
floor to boardroom – by giving real-time business processes
instantaneous access to data. the alliance of these two
technologies can eliminate today’s information lag for
your business.
With the revolution of in-memory While non-saP customers can inte-
computing already under way, the grate in-memory technology into their
question isn’t if this revolution will existing software environments, saP
impact businesses but when and, customers can transform their busi-
more importantly, how. similar to the nesses into real-time enterprises with
advance of enterprise resource plan- game-changing potential without dis-
ning (eRP) software in the 1990s, rupting their existing it landscapes.
in-memory computing won’t be intro- the Business transformation services
duced because a company can afford group of saP Consulting is available
the technology. it will be brought on to help you assess the impact this
board because a business cannot technology can have on your it envi-
afford to allow its competitors to adopt ronment and develop a strategy that
the technology first. leverages it specifically to differentiate
your business.
this paper details how in-memory
computing can change the way you the first product to leverage saP in-
manage business intelligence and the Memory Computing technology is saP
value your business can derive from in-Memory appliance (saP hana™)
the technology. For business and it software, which is currently in use at
executives, the paper furnishes sub- saP and with saP customers.
stantial information and business exam-
ples about what changes they can look
forward to and how those changes can
catalyze their strategic initiatives.
4 SAP Thought Leadership – saP in-Memory Computing technology

5. IntroduCtIon
WhiCh halF OF My Budget is WORKing?
in order to establish an effective Where Does In-Memory Come Business Examples That Leverage
business intelligence (Bi) strategy, it into the BI Picture? In-Memory Computing
professionals must ask themselves
which half of the money they spend on in 2010 saP co-founder hasso Plattner to show how compelling this technolo-
Bi investments is working. Within the announced in his keynote speech at gy can be, we have included a handful
marketing organization, you probably the saPPhiRe® nOW conference that of examples showing how in-memory
use focus groups to determine what saP in-Memory Computing technology computing works in real-life, real-time
works and what does not. By using was an it revolution that was putting situations to make business more
saP BusinessObjects™ business intelli- the “r” for “real time” back into enter- responsive, more effective, and more
gence (Bi) solutions, you can get help prise resource planning solutions. saP successful.
from powerful statistics tools to focus started its pursuit of real-time data pro-
your efforts quickly. statistics can help cessing with saP R/2® and saP R/3® in manufacturing enterprises, in-memory
you determine, for example, which 20% software. computing technology will connect the
of the data in your data warehouse is shop floor to the boardroom, and the
used by 80% of your users. When you saP co-CeO Bill Mcdermott suggests shop floor associate will have instant
know this, you can focus 80% of your that the new technology will eliminate access to the same data as the board
investment on that fraction of the data. the Bi middleman – the employee who member. the technology supports this
reports data. the reason for this is by integrating on-premise, on-demand,
Our experience shows that Bi statistics there will no longer be any need to and on-device architectures. Once the
are a good starting point for optimizing report up or down the key performance appropriate business processes are in
Bi spend. they can help you manage indicators (KPis). place, empowered shop floor staff can
your Bi graveyards – the reports and take immediate action based on real-
data that are simply never used – and here is just a sampling of what in- time data to make whatever adjust-
maximize the usefulness of business memory computing technology can ments on the shop floor are necessary.
intelligence that is in high demand. a do for you: they will then see the results of their
great complementary tool for optimiz- • enable mixed workloads of analytics, actions reflected immediately in the
ing your Bi investment is one or more operations, and performance man- relevant KPis. you could call this true
Bi focus groups that include end users. agement in a single software 360-degree Bi, as it eliminates the mid-
landscape dleman as well as the need to create
although using system statistics to • support smarter business decisions any reports other than whatever statu-
manage Bi spend is available to all cus- by providing increased visibility of tory or legal reports may be required.
tomers, we find that some companies very large volumes of business
don’t perform even this basic step. information saP BusinessObjects event insight
as a result, they remain unaware of • enable users to react to business software is key. in what used to be
what Bi is available to them through events more quickly through real-time called exception reporting, the soft-
their software and don’t leverage it. analysis and reporting of operational ware deals with huge amounts of real-
they certainly are in no position to take data time data to determine immediate and
full advantage of the new in-memory • Provide greater flexibility by delivering appropriate action for a real-time situa-
computing technology. even the best innovative real-time analysis and tion. Product managers will still look at
technology will not help if you lack an reporting functions inventory and point-of-sale data, but in
effective Bi strategy. • support the deployment of innovative the future they will also receive, for
new business applications example, notifications when customers
• help streamline the it landscape and broadcast their dissatisfaction with a
reduce total cost of ownership (tCO) product to the masses over twitter.
SAP Thought Leadership – saP in-Memory Computing technology 5

6. in order to establish an effective Bi strategy, it profes-
sionals must ask themselves which half of the money
they spend on Bi investments is working. Within the
marketing organization, you probably use focus
groups to determine what works and what does not.
By using saP BusinessObjects Bi solutions, you can
get help from powerful statistics tools to focus your
efforts quickly.
Or they might be alerted to a negative reels, changes in technology – both dis- the following few hours. Consumers
product review released online that ruptive and accelerated – will surge then have the option to save money by
highlights some unpleasant product through other industries besides it. modifying their immediate consumption
features requiring immediate action. patterns, perhaps by switching off the
From the other side, small businesses another example worth mentioning power at their residence and going
running real-time inventory reports will harks back to Mcdermott’s middleman: to a movie. By giving consumers an
be able to announce to their Facebook it will no longer be good enough to informed choice and an incentive, utili-
and twitter communities that a high- have the weekend numbers ready ties companies have a chance to mod-
demand product is available, how to for executives on Monday morning. erate peaks in energy consumption.
order it, and where to pick it up. executives will run their own reports on this advantage will become much more
revenue, twitter their reviews over the dramatic when we switch to electric
an excellent example comes from weekend, and by Monday morning cars; predictably, those cars are going
today’s entertainment companies. Con- have acted on their decisions. to be recharged the minute the owners
ventionally, bad movies have been able return home from work, which could be
to enjoy a great opening weekend Our final example is from the utilities within a very short period of time.
before crashing the second weekend industry: the most expensive energy
when negative word-of-mouth feedback that a utilities company provides is
has cooled off the initial enthusiasm. energy to meet unexpected demand
that week-long grace period is about to during peak periods of consumption. in
disappear for silver screen flops. in the those cases, the provider may have to
future, consumer feedback won’t take a buy additional energy to support the
week, a day, or an hour. the very sec- power grid, which can get expensive.
ond showing of a movie could suffer however, if the company could analyze
from a noticeable falloff in attendance trends in electrical power consumption
due to consumer criticism piped instan- based on real-time meter reading, it
taneously through the new technolo- could offer its consumers – in real time
gies. since such rapid response is not – extra low rates for the week or month
possible with old-fashioned movie if they reduce their consumption during
6 SAP Thought Leadership – saP in-Memory Computing technology

7. In-MeMory CoMPutIng
teChnology
RePlaCing tRaditiOnal dataBases
in-memory computing technology leverages in-memory computing tech-
combines hardware and software tech- nology. the accelerated version of
nology innovations. hardware innova- saP BusinessObjects explorer™
tions include blade servers and CPus software makes use of the technol-
with multicore architecture and memory ogy for data provided by the saP
capacities measured in terabytes for netWeaver Business Warehouse
massive parallel scaling. software inno- (saP netWeaver BW) component and
vations include an in-memory database saP BusinessObjects data services
with highly compressible row and col- software. however, in-memory comput-
umn storage specifically designed by ing will not eliminate the need for data
saP to maximize in-memory computing warehousing. a real-time reporting
technology. Parallel processing takes function will solve old challenges and
place in the database layer rather than create new opportunities, but new chal-
in the application layer as we know it lenges will arise.
from the client-server architecture. in manufacturing enterprises,
total cost is expected to be 30% lower saP hana 1.0 software supports real-
than traditional relational database time database access to data from the
in-memory computing tech-
technology due to: saP applications that support OltP. in nology will connect the
• leaner hardware and less system a parallel environment, updates of data
capacity required, as mixed work- in real time utilize database replication
shop floor to the board-
loads of analytics, operations, and developed by sybase. the first saP room, and the shop floor
performance management are han- application to run on an in-memory data-
dled within a single system, which base is saP netWeaver BW 7.30 using
associate will have instant
also reduces redundant data storage saP hana 1.5. this advance has elimi- access to the same data as
• Reduced extract, transform, and load nated the need for separate hardware to
(etl) processes between systems run saP netWeaver BW accelerator.
the board member. the tech-
and fewer prebuilt reports, reducing nology supports this by
the support effort required to run the Both versions of the appliance soft-
software ware can be accessed with the saP
integrating on-premise,
BusinessObjects Business intelligence on-demand, and on-device
Replacing traditional databases in saP platform. the platform provides a
applications with in-memory computing shared semantic data layer for saP
architectures. Once the
technology resulted in report runtime BusinessObjects Bi solutions and saP appropriate business pro-
improvements of up to a factor of 1000 BusinessObjects enterprise perfor-
and compression rates of up to a factor mance management (ePM) solutions. it
cesses are in place, em-
of 10. Performance improvements are allows optimized Bi applications to take powered shop-floor staff
expected to be even higher in saP full advantage of the in-memory com-
applications natively developed for in- puting technology.
can take immediate action
memory databases, with initial results based on real-time data to
showing a reduction of computing time
from several hours to a few seconds.
make whatever adjustments
Currently saP netWeaver® Business on the shop floor are
Warehouse accelerator (saP
netWeaver BW accelerator) software
necessary.
SAP Thought Leadership – saP in-Memory Computing technology 7

8. BuSIneSS IntellIgenCe
PROviding MORe than FanCy dashBOaRds,
But hOW MuCh MORe?
Business intelligence is more than software, and cost management simula-
fancy dashboards – at least most Bi tion tools. add to the Bi category statis-
professionals agree with that. at first, tical applications that perform predictive
computer-generated data was simply analysis and saP BusinessObjects
referred to as reports. Over time new event insight – which is based on infor-
terminology was introduced: executive mation streams issuing from an event
information system (eis) in the early grid – and two things happen. it be-
1990s, enterprise information manage- comes almost impossible to agree on
ment (eiM), and strategic information an industry-wide definition of business
management (siM, see Figure 1), to intelligence, and it becomes impossible
name a few. to separate Bi from other it data
processing.
Business intelligence can be complex
and, depending on its definition, could For that reason, we do not focus on
be understood to cover functionality traditional Bi definitions in this thought
offered by saP BusinessObjects ePM leadership paper. We focus our atten-
solutions. this would include financial tion instead on analyzing the overall
planning and consolidation applications, impact of in-memory computing
strategy management and profitability technology.
On premise
Competitive
differentiation
Fact-based Organizational
decision making readiness
On demand
Value
On device
Governance Strategic management
information
management
Organizational
Enterprise data change
management management
Predictive
analytics
and business
intelligence
Enterprise Enterprise
architecture performance
alignment management
Figure 1: Strategic Information Management
8 SAP Thought Leadership – saP in-Memory Computing technology

9. ProjeCt PortfolIo MAnAgeMent
aligning it With CORPORate-Wide
Business stRategy
Our world is characterized by rapid zations, achieve business objectives, that transaction processing functionality
changes, disruptive technologies, and keep lines of business profitable. will have to be aligned much more closely
and strong competition. Clearly, no with the integrated Bi functionality.
business can afford to waste its limited the objectives for it portfolio manage-
resources on duplicate efforts; on ment include: saP BusinessObjects event insight,
freewheeling short-term, throwaway • establishing clear guidelines for which uses in-memory computing,
development; or on shadow it efforts – gathering and prioritizing require- requires a tightly woven network of
stopgap, maverick it work performed ments, executing project work, and data processes and data exploration to
outside the official corporate it stra- maintaining budget accountability identify specific thresholds of perfor-
tegy. nor will the biggest corporate • aligning existing and planned projects mance measures. Reaching a threshold
spenders necessarily be tomorrow’s with these guidelines then triggers additional steps in one or
leaders. that role will go to the compa- • Prohibiting all it and business initia- more business processes. in order to
nies with it strategies that enable their tives not aligned with the guidelines implement an end-to-end business pro-
enterprise architects to align the it of the it portfolio cess with embedded Bi, both business-
infrastructure with business strategy process effort and it effort must be
across business units. if a company intends to leverage in- carefully orchestrated.
memory computing technology, its it
that is the reason why it portfolio portfolio management becomes even Cost Impact for
management is a top priority for com- more important. We give two examples Portfolio Management
panies today. however, to put that in that illustrate this.
place, you must understand your busi- the lion’s share of it spend is tied up in
ness strategy, then establish an it Formerly, operational reporting function- operations and maintenance efforts,
portfolio that reinforces that strategy ality was transferred from OltP appli- with less it spend available for busi-
with every it decision made (see cations to a data warehouse. With ness innovation. if potential for innova-
Figure 2). in doing so, your it can help in-memory computing technology, this tion is to be significantly increased,
align company priorities across lines of functionality is integrated back into the organizations need mature portfolio
business and corporate support organi- transaction system. the consequence is management processes and capabili-
ties to optimize spend without creating
too many parallel developments, which
Strategy Strategic direction could entail even higher it support
• Sets strategic direction for the spend. For most companies this will
organization mean new governance structures that
• Defines objectives Strategy and objectives span silos.
• Drives all further organizational actions
Planning and management Operations Project portfolio
establishes appropriate actions required
to meet strategic objectives
Execution Ongoing Authorized
• Manages the organization’s operations operations programs
and portfolios and projects
• Enables efficient and effective execution
Organizational resources
Figure 2: Portfolio Management
SAP Thought Leadership – saP in-Memory Computing technology 9

10. enterPrISe ArChIteCture
designing the FutuRe
enterprise architecture emerges where metadata and master data across busi- implemented. the technical specifica-
business requirements are formally and ness processes. some of the most tions of the hardware making up the
rigorously sustained by it. in advanced salient shifts and trends in future enter- landscape also play a role. another fac-
enterprise architectural processes, no prise architectures will be: tor is whether a company’s data center
technology is implemented that has not • a shift to Bi self-service application- is committed to a single vendor’s tech-
been vetted and approved by the enter- like data exploration, instead of rolling nology or is prepared to incorporate
prise architecture office with regard to out static report solutions for struc- technology from a range of vendors.
strategic viability and medium- to long- tured and unstructured data
term benefit for the enterprise. how- • Full integration of planning business it is most likely that future deployments
ever, those organizations wanting a processes with instant Bi provision- of saP netWeaver BW will not require
head-start in in-memory computing ing from the source applications, separate hardware to run saP
must deploy at least one in-memory replacing data warehouses for opera- netWeaver BW accelerator. however,
computing project as soon as possible tional data and near-real-time data to what extent existing hardware com-
to develop know-how, resources, and a transfers ponents can be reused for saP
feel for how the new technology will • substitution of traditional etl archi- netWeaver BW accelerator will
impact their unique situation. tecture and expansive data cleansing depend on how that hardware exploits
and harmonization processes with the in-memory computing technology
Obviously, it will take some time before real-time data validation during all of the application.
the first applications developed specifi- manual and automated data input
cally to exploit in-memory computing processes Real-time in-memory computing technol-
affect the enterprise application archi- • Central metadata and master-data ogy will most probably cause a decline
tecture. yet in-memory computing tech- repositories that define the data in sheer numbers of structured Query
nology will have a major impact on Bi architecture, allowing data stewards language (sQl) satellite databases.
and data warehouse application archi- to work effectively across all busi- the purpose of those databases as flex-
tecture as well as on OltP applications. ness units and all architecture ible, ad hoc, more business-oriented,
this is something enterprise architects platforms less it-static tools might still be required,
should take stock of earlier rather than • instantaneous analysis of real-time but their offline status will be too much
later. Real-time data access to Bi infor- trending algorithms with direct impact of a disadvantage and will delay data
mation and the repurposing of data on live execution of business updates. some might argue that satellite
warehouses will be key to orchestrating processes systems equipped with in-memory com-
their on-premise, on-device, and on- • Offline long-term historic trending puting technology will take over from
demand architecture successfully. and that can impact future execution of satellite sQl databases. For limited
they will find that in-memory computing business processes sandbox purposes, that is a possibility.
will open up new avenues and change • Construction of an event insight grid however, because in-memory comput-
the way they regard three-tier architec- architecture combining live business ing technology can process massive
ture for data processing. applications across on-premise, on- quantities of real-time data to provide
device, and on-demand architectures instantaneous results, traditional satel-
A Clean, Spare Architecture for proactive use of Bi instead of ana- lite architectures will always be at least
lyzing historic events after the fact one step behind. they are also likely to
adopting in-memory computing results inherit undesired transformations made
in an uncluttered architecture based on What specific changes are introduced during the etl process.
a few, tightly aligned core systems to existing landscapes depend on how
enabled by service-oriented architec- functional requirements, such as high
ture (sOa) to provide harmonized, valid availability or disaster recovery, were
10 SAP Thought Leadership – saP in-Memory Computing technology

11. in-memory computing technology combines hard-
ware and software technology innovations. hard-
ware innovations include blade servers and CPus
with multicore architecture and memory capacities
measured in terabytes for massive parallel scaling.
software innovations include an in-memory data-
base with highly compressible row and column stor-
age specifically designed by saP to maximize in-
memory computing technology. Parallel processing
takes place in the database layer rather than in the
application layer.
Cost-Effective Impact of incrementally, and compatibility with cur-
Enterprise Architecture rent and legacy landscapes. in-memory
computing technology provides scaling
When enterprise architecture plays a and flexibility of hardware for higher per-
role in setting up and maintaining Bi proj- formance. On-the-fly aggregations
ects, it helps reduce costs by consider- relieve it staff from manual query tuning
ing the overall context of the project. it and data aggregation tasks. in cases
examines business needs, it landscape, where an enterprise data warehouse is
performance requirements, data model not in place, saP hana provides
complexity, and the tools and software instant access to real-time data via repli-
required to meet reporting demands. in cation from eRP software from saP
seeing that these requirements are met, with no need for complex etl
it helps establish a unified Bi platform processes.
that supports administration, thereby
significantly trimming the tCO for a Bi By contrast, in traditional data ware-
project. house environments, ever higher per-
formance and functional requirements
analytics based on in-memory comput- lead to the acquisition of additional
ing also contribute to lower tCO of hardware, software, and performance-
both new and existing landscapes. tuning tasks. in highly heterogeneous
some huge cost benefits for it include environments, multiple Bi solution sets
reduced hardware costs, higher perfor- require additional independent lifecycle
mance and business agility, faster management, which adds to solution
deployment, the opportunity to adopt maintenance efforts.
SAP Thought Leadership – saP in-Memory Computing technology 11

12. the PotentIAl of MetAdAtA
And MASter dAtA
PROFiting FROM intelligent data
if a Bi solution is to provide relevant ments, strict data governance must be in-memory computing tech-
information, high data quality is essen- in place to maintain the high quality of
tial. But real-time analytics allow no real-time data based on company-wide nology provides scaling and
time for data cleansing processes. to data standards and controls. if current flexibility of hardware for
tackle this issue, it is sound strategy to data management processes are in-
establish the formal role of chief data sufficient, in-memory computing tech- higher performance. On-
officer in a company, as saP has done. nology will only make existing issues the-fly aggregations relieve
this step is an effective way to stan- worse.
dardize data processes across the it staff from manual query
enterprise and achieve the highest pos- Cost Impact for Metadata and tuning and data aggregation
sible data quality at the time the data is Master Data
created. tasks. in cases where an
Reporting in real time based on in- enterprise data warehouse
Considering the number of data fields memory computing requires reading
in saP Business suite software, it directly from analytics data pools in is not in place, saP hana
would be a herculean task to define the original source system. Because provides instant access to
processes and standards for them all. the data is consumed before traditional
since the functionality in saP Business data cleansing processes can be real-time data via replication
suite is designed to take care of this, performed, data validation must be from eRP software from
there is no need to. this is not the case performed upstream. From a tCO
in a heterogeneous environment that perspective, that might sound like more saP with no need for com-
brings together legacy systems from effort than conventional reporting plex etl processes.
different vendors, systems inherited methods require. We would like to sug-
from mergers and acquisitions, and gest that every dollar a company
business processes reflecting different spends on data quality upstream elimi-
levels of maturity. For such environ- nates two dollars spent downstream.
Negative Business Impact from Poorly Managed Master Data
Line of business
Financials IT Sales Supply chain
Chief financial officer Chief information officer head of sales and marketing head of supply chain
or manufacturing
Challenges
Compliance TCO and Flexibility Opportunities Optimization
• Manual efforts to update • high master-data • lack of consolidated view • lack of transparency relating
financial master data in local maintenance costs of customer data across all to suppliers and products
financial systems • need for multiple systems channels • unused benefits from
• slow group closes to manage master data in • Compromise of customer discounts, conditions,
• Risk of not complying different domains transactions due to limited and central contracts
with international Financial • lack of flexibility in dynamic view of the facts • limited availability of up-to-date
Reporting standards or business environments master data in business network
the sarbanes-Oxley act
12 SAP Thought Leadership – saP in-Memory Computing technology

13. Most companies spend a significant
amount of time and effort manually
Incomplete address cleansing master data after transac-
in new account... tions have been created in OltP appli-
cations. excessive effort mounts for
other reasons: data validation must
be developed in the etl processes.
Multiple sources of truth must be har-
monized for multiple Bi applications.
Meetings must be held in which staff
members discuss which data is the
right data. Often, these meetings kick
off another round of data validation
...is automatically efforts. By contrast, in-memory com-
filled in. puting technology provides support for
improved data storage, search, and
analysis to enhance data consistency.
Cost-effective deployment
Figure 3: Real-Time Data Entry Quality Check of in-Memory Computing
to deploy in-memory computing cost-
effectively, it is critical to link data-quality
tools such as saP BusinessObjects
data services software directly to both
manual and automatic data entry (see
Figures 3 and 4). these tools help
curb invalid data coming from OltP
applications into the data warehouse
through traditional etl processes or as
periodic batch processes. Other ways
to reduce tCO include identifying the
right data sources with proper data gov-
User has option ernance processes and choosing the
to resolve the
cleansing case
right integration platform, such as saP
by merging now BusinessObjects data services.
or later.
Figure 4: Real-Time Data Entry Duplication Check
SAP Thought Leadership – saP in-Memory Computing technology 13

14. PeoPle, ProCeSSeS, And StruCtureS
In BuSIneSS And It
ReQuesting Just the data, Please, nO RePORts
to fully enable Bi governance, support
of business processes end to end is
today, saP netWeaver BW can store high-demand required. the saP Business Process
data using saP netWeaver BW accelerator, medium- Management service offers this kind of
support based on a business process–
demand data with relational database management centric approach. On the it side, a pro-
systems, and low-demand data with near-line storage cess information owner is responsible
for an entire business process. From
solutions on optical discs. these different storage the business side, a business process
strategies remain invisible to users. using in-memory owner is given responsibility for the
business process.
computing, the database can manage cold, lukewarm,
and hot spaces and move data automatically between an example from forecasting – a typical
cross-function business process –
the areas based on parameter values. illustrates how this works. the parties
involved could be marketing, sales, and
the finance and controlling function.
Frustrated by limited Bi, today’s gener- nance is ineffective. in these contexts, depending on the type of industry, a
ation of users wants to bypass prefab- under-the-desk sQl databases and delivery business unit like manufacturing
ricated reports and get right to the data freewheeling Microsoft excel spread- or mining or a supply chain unit could be
itself. We hear it often during our Bi sheets compete for executive attention. involved. From the business side, the
review sessions with customers, and business process owner of forecasting
we have an answer: Bi tools in saP BI Governance is responsible for aligning the forecast-
software can satisfy this need. the ing processes across business units.
tools reduce the dependence of busi- Bi governance as we understand the how well this is done can be measured
ness users on it and increase Bi term does not strive to control the num- in process performance indicators
self-service. ber of new reports created by users, but (PPis). the business process owner
it strives to base data interpretation on a also serves as the central point of con-
saP Business explorer tools that use single source of truth. true business- tact between business and it and is
in-memory computing technology rep- user empowerment should not restrict explicitly invested by the board with
resent a paradigm shift. instead of wait- freedom to use available resources but decision power. (it should be noted that
ing for it to work on a long queue of provide a framework for: the business process owner is not
support tickets to create new reports, • standards to achieve consistent qual- responsible for getting actual revenue
business users can access large data ity data to meet the forecast data. instead, the
sets for data exploration and defining • access to controls to minimize risk business process owner’s role is to con-
reports on the fly. and support compliance trol, optimize, or reinforce the process.)
• Bi architecture where analytics does
the key to success lies not in the tech- not impair business process to maximize the effect of the business
nological support the Bi tools provide, performance process owner’s efforts, the company
but in the effective governance sup- • training and education to maximize should designate an it counterpart: a
porting the Bi self-service with regard user potential single point of contact responsible for
to processes, structures, and people. supporting the forecasting business
in-memory computing technology only process from the it side. to realize
increases negative impact when gover- this, it must be able to adjust its opera-
14 SAP Thought Leadership – saP in-Memory Computing technology

15. tions to provide single-person, end-to-
end process support, thus eliminating
the need to hand off it responsibilities
from one it group to another for a sin-
gle business process. this dramatically
increases the likelihood of handling
business processes effectively.
We saw a case where a number of it
teams were assigned to perform differ-
ent stages in handing off an external
data load-interface task to the saP
eRP Central Component component.
One team was responsible for the
etl stage from saP eRP Central
Component to the data warehouse,
data harmonization, and data validation.
another team took over data modeling
in the data warehouse. a third team
was responsible for the Bi front-end
applications. Other it teams supported
sQl satellite data warehouses for
other Bi front-end applications. al-
though splitting up support resources
this way is cumbersome and costly, we
are aware of cases where seven or
more handovers occurred between it
groups, at times across continents. business savvy, they can provide ever reports. unfortunately, this is precisely
better support for business processes the time in the project when the it
in-memory computing technology pro- end to end. group members must focus on stabiliz-
vides a major opportunity for it to ing the new environment. their tasks
realign itself as a business process– Cost Impact for People, Processes, may include fine-tuning the data ware-
centric organization based on orches- Structure, and Technology house design, preparing and deploying
tration of an on-premise, on-demand, prebuilt reports running on predefined
on-device architecture. this would do at the start of a Bi implementation aggregates, or defining processes to
away with a fragmented support model, project, implementation teams may be warm up limited memory caches. With
which only intensifies shadow it efforts presented with thousands of legacy valuable it resources tied up with over-
by the business. as the organization reports they are to rebuild in the new head functions, the business cannot
changes, so will the people. the more environment. it is not uncommon, how- wait for its new reports, so it starts
tech-savvy business users become and ever, that once the teams have rebuilt building sQl databases. this situation
the easier Bi products are to use, the those reports and after the new soft- presents a tremendous opportunity to
less business will have to rely on it to ware goes live, they are inundated with reduce capital and operational expendi-
produce reports and dashboards. simi- new report requests because nobody tures while increasing user experience.
larly, as it personnel become more in the user community used the old
SAP Thought Leadership – saP in-Memory Computing technology 15

16. BI dAtA lIfeCyCle MAnAgeMent
KeePing data hOt OR COld?
We all know that the amount of data answers to such questions as: Why spaces. Only high-demand data should
packing this planet is growing. Just keep pushing marketing dollars for a be stored in the “hot” spaces. “hot”
reflecting on how much information from product version when its reviews are refers to storing data in a dynamic
social media networks we sift through under par? should we accelerate the state, which allows frequent data loads
every day can make us feel slightly ill at release of a new version and start pro- for high availability and split-second
ease about it. We also feel quite keenly motions to push the old product out of response time despite access by many
that the rate at which that data is accu- the retail shelves and out of the internet users. “Cold” refers to the storage of
mulating is increasing. What’s worse, warehouses instead of replacing the nonoperational static data infrequently
we neither can nor should ignore it. inventory? there’s a very real possibility accessed and causing no additional
that unstructured data might render tra- data loads. data for long-term analytics
Companies are in a far worse quandary: ditional forecasting methods obsolete. or multiyear trending is an example of
their business depends on it. One day cold data.
unstructured data might render tradi- Cost Impact for BI Data
tional forecast methods obsolete. Prod- Lifecycle Management today, saP netWeaver BW can
uct reviews posted on the internet may store high-demand data using saP
have greater impact on forecasts, sales the majority of databases and data netWeaver BW accelerator, medium-
pipelines, profit margins, and product marts could be consolidated into a demand data with relational database
lifecycles than actual sales from point- single, integrated enterprise data management systems, and low-demand
of-sale data. it is no mystery why more warehouse, and it would make sense data with near-line storage solutions
saP customers spend as much money to do so. Most of the data in data ware- on optical discs. these different stor-
and time analyzing unstructured data as houses is hardly used, with only a frac- age strategies remain invisible to users.
they do analyzing traditional structured tion accessed on a daily basis. data using in-memory computing, the
Bi data. volume maintenance is one of the database can manage cold, lukewarm,
major recurring costs and activities and hot spaces and move data auto-
analysis of unstructured data may of it. to minimize these costs, infre- matically between the areas based on
become essential in determining prod- quently used data should be off-loaded parameter values.
uct lifecycles by giving companies to “cold” hardware and database
Unstructured enterprise content... ...is changing in two fundamental ways... ...creating strategic business challenges
Growth of data and documents Business process inefficiency
accelerates while regulatory oversight • Breakdown of ad hoc processes
continually increases • Lost productivity searching for content
• Emergence of siloed solutions
tRi PsR
Paper/Physical; Graphics; Video; E-mail sOX euROPa Regulatory risk exposure
hiPPa eC1085 • Legal requirement of storing complete
s-296 FisMa record of processes
• Unmanaged content “lost in the
Corporate glBa uCC-1 RCRa haystack”
documents Osha eeO eC 2003 • Longer-than-necessary retention
of content
1990 2000 2010 • Growth of e-discovery costs
Figure 5: Business Challenges of Unstructured Data
16 SAP Thought Leadership – saP in-Memory Computing technology

17. CoMPArIng ArChIteCtureS
tRaditiOnal veRsus in-MeMORy COMPuting
Traditional architecture In-memory computing
Portfolio management Parallel programs to address process and Overarching portfolio management for
analytical requirements in separate projects redundant data and business intelligence
(Bi) efforts
Enterprise architecture • layered approach: transactional system layer, • streamlined it landscape with analytics,
enterprise data warehouse layer, and regional operations, and performance management
and business satellite data warehouse layers combined in one system
• Bi based on 24-hour-old data • Real-time enterprise
Metadata and master-data support • Fragmented data governance • enterprise-wide data governance
• downstream data management, such as • upstream data validation, with data
harmonization and validation consumption occurring before data
management processes can be performed
People, processes, and structures in • high dependence of business on it • high degree of Bi self-service
business and IT • shadow it in competition with enterprise it • enterprise-wide governance, with business
• Business silos process owners crossing business units in
partnership with process-centric it
BI data lifecycle management • Focus on structured data • structured and unstructured data
• growing data volume is not proactively • growing data volume distributed across
managed by actual usage and only segre- different storage media determined by user
gated by multiple online and offline solutions needs and managed by a single database
SAP Thought Leadership – saP in-Memory Computing technology 17

18. outlook
aCCeleRating intO the FutuRe
We do not proclaim to know the future on how you can shape your business
and cannot know for certain where the opportunities with saP in-Memory Com-
in-memory computing revolution will puting. the Business transformation
lead us. What we can say is that the services group from saP Consulting is
revolution is under way, and for a busi- skilled at bringing business and it into
ness to succeed, it must consider the alignment. the saP experts can work
possibilities opening up and prepare for with you to create a strategy to leverage
the transformation now. technology, deliver value, and differenti-
ate your business in the marketplace.
in-memory computing technology will
alter existing it processes more quickly For More Information
than ever before. at the same time,
business is pursuing disruptive busi- to learn more about the software and
ness ideas at an unprecedented rate. tools that can help you manage tCO
in-memory computing will be able to for Bi, call an saP representative today
connect the dots between existing or go to www.sap.com. For more infor-
application platforms, new mobile mation on the Business transformation
community technologies, and nascent services group, see www.sap.com/usa
technologies. /services/consulting/bts/index.epx.
For the latest information on the saP
although some of the business exam- product road map and release informa-
ples provided in this paper might not tion, visit the saP developer network
relate to your unique situation, we trust site at www.sdn.sap.com.
we have provided you with some ideas
18 SAP Thought Leadership – saP in-Memory Computing technology

20. 50 104 231 (11/03)
©2011 saP ag. all rights reserved.
saP, R/3, saP netWeaver, duet, Partneredge, Bydesign,
saP BusinessObjects explorer, streamWork, and other saP products
and services mentioned herein as well as their respective logos are
trademarks or registered trademarks of saP ag in germany and
other countries.
Business Objects and the Business Objects logo, BusinessObjects,
Crystal Reports, Crystal decisions, Web intelligence, Xcelsius, and
other Business Objects products and services mentioned herein as well
as their respective logos are trademarks or registered trademarks of
Business Objects software ltd. Business Objects is an saP company.
sybase and adaptive server, ianywhere, sybase 365, sQl anywhere,
and other sybase products and services mentioned herein as well
as their respective logos are trademarks or registered trademarks of
sybase, inc. sybase is an saP company.
all other product and service names mentioned are the trademarks
of their respective companies. data contained in this document serves
informational purposes only. national product specifications may vary.
these materials are subject to change without notice. these materials
are provided by saP ag and its affiliated companies (“saP group”)
for informational purposes only, without representation or warranty of
any kind, and saP group shall not be liable for errors or omissions with
respect to the materials. the only warranties for saP group products
and services are those that are set forth in the express warranty
statements accompanying such products and services, if any. nothing
herein should be construed as constituting an additional warranty.
www.sap.com /contactsap