Presentation of SAP's latest in-memory technology Hana, presentation to School of Information and Service Economy of Aalto University Helsinki, Prof. Matti Rossi, presentation includes links to demo systems and explains how to apply for access to a real SAP Hana system.
2. AGENDA
management|consulting
Self Introduction
Trends in the Global IT industry
The Pretense of Knowledge
The Journey towards In Memory Computing
Introducing SAP Hana - In Memory DB
SAP Hana - Live Demonstrations
Q&A
3. COMPANY INTRODUCTION
management|consulting
Founded in 2008 by Dr. Markus Alsleben, Alsleben Ltd. provides
management consulting and professional services critical for companies
engaging in the high velocity Chinese marketplace.
At Alsleben Ltd. we believe that quality advisory in the context of high
velocity environments can only be successful through a solid scientific
foundation. Management research projects are therefore an integral part of
Dr. Markus Alsleben our practice incorporating latest research into unique client solutions. Our
CEO Alsleben Ltd.
affiliation with prestigious research institutions and corporations enables us
to utilize the latest knowledge base for your management consulting projects
with Alsleben. Ltd. implementing next practice today.
Affiliations Selected Clients Our services include:
•Management Consulting and Training Services: Since 2008 Alsleben Ltd. has
worked together with leading multinational companies across various industries in
China and around the world to design and implement strategies, change
organizations and conduct training services that deliver results.
•Information Technology Advisory: Business without powerful IT support is
impossible in today's hyper competition. Designing and implementing IT Strategies
and ERP Systems provides the competitive edge sustainable success for your China
operations.
•Human Resources: World-class talent acquisition and management are key
capabilities of successful enterprises in China. Alsleben Ltd. provides talent
management solutions that let you win the war for talent in China.
4. BIO
management|consulting
Alsleben Ltd.
2008 - today CEO Management Consulting,
Hong Kong
Lead Management Consultant
Location Strategy & Management Project SAP AG
2008 - 2010 Designing and Implementing SAP's global Location Strategy.
Germany
Spatial reorganization and optimization of R&D at SAP.
Vice President SAP Labs China
2006 - 2008 Corporate development and execution of growth strategy for
Shanghai
development locations in China.
Vice President - Consulting Director North Asia SAP China
2000 - 2006 Consulting head for Greater China with more than 150
Beijing
consultant, delivering SAP implementations.
KPMG Consulting
1997 - 1999 Senior SAP Consultant for Logistics now Bearing Point &
o.tel.o Telecom, Germany
5. PUBLICATIONS
management|consulting
Creating Dynamic Capabilities
R&D Network Management
for Globally Distributed
Research and Development
in the Software Industry SAP: Establishing a Research Centre in China
Harvard Business Publishing - Case Study
7. GLOBAL IT TRENDS - HYPE CURVE
management|consulting
Big Data
Cloud
Mobile
Source: Gartner, 2012.
8. GLOBAL IT TRENDS
management|consulting
CLOUD
COMPUTING BIG DATA
The exponential growth
in data across all
Cloud computing provides “convenient industries requires new
on-demand technologies for:
network access to a shared
pool of configurable computing
resources that can be quickly
provisioned and released with minimal • Data Sourcing
management effort or service provider and Storage
interaction.”1 The various subsets of • Data Integration and
could computing as SaaS, PaaS, Iaas Transformation
or more generic XaaS provide cost to generate new insights
effective and high available computing
• Data Analysis and
and opportunities.
Classification
resources with near to unlimited
scalability.
MOBILE COMPUTING
The increasing penetration of connected mobile phones and tablet computers allows
new context based services as e.g. location based services, augmented reality and
rapid data collection e.g. for traffic analysis. Always on mobile devices allow quick
communication and collaboration. By 2013, more than 15 billion devices will be
connected to the Internet using a mobile device.
Source: Mell, p. and Grance, t. the nIst definition of cloud computing. Special Publication 800-145, 2011; http:// csrc.nist.gov/publications/nistpubs/800-145/sp800-
1
10. 10 n
management|consulting
Prefix 10n Decimal Scale
0 1 one
deca 110 ten
4k Memory
hecto 2100 hundred Apollo
Guidance
kilo 31,000 thousand Computer
1 Terabyte
mega 61,000,000 million equals 210
giga 91,000,000,000 billion single sided
DVDs
tera 121,000,000,000,000 trillion 2.5 Petabyte
Wallmart’s
peta 151,000,000,000,000,000 quadrillion annual Data
295Growth
Exabyte
exa 181,000,000,000,000,000,000 quintillion
estimated
zetta 211,000,000,000,000,000,000,000 sextillion complete
human
yotta 241,000,000,000,000,000,000,000,000 septillion knowledge in
2007
880 Yottameter 7.9 Zetabyte
diameter of est. amount of
observable universe digital data by
2015
11. BIG DATA IS NOT ONLY BIG...
management|consulting
Source: SAP 2012.
12. Business Rational of Mobile Enterprise
Computing management|consulting
Source: SAP 2012.
13. A day in the life with mobile analytics suite
management|consulting
Source: SAP 2012.
14. management|consulting
THE PRETENSE OF
KNOWLEDGE
Friedrich August Hayek Herbert A. Simon Nassim Nicholas Taleb
15. SOCIAL SCIENCE ≠ PHYSICAL SCIENCE
management|consulting
“It seems to me that this failure of the economists to guide policy more
successfully is closely connected with their propensity to imitate as closely as
possible the procedures of the brilliantly successful physical sciences - an
Friedrich August Hayek
Noble Laureate in Economics 1974
attempt which in our field may lead to outright error. [...]
Unlike the position that exists in the physical sciences, in economics and other
disciplines that deal with essentially complex phenomena, the aspects of the events
QUANTITATIVE to be accounted for about which we can get quantitative data are necessarily
RESEARCH limited and may not include the important ones.
While in the physical sciences it is generally assumed, probably with good reason,
QUALITATIVE
RESEARCH v that any important factor which determines the observed events will itself be
directly observable and measurable, in the study of such complex phenomena as
the market, which depend on the actions of many individuals, all the
MIXED-METHODS circumstances which will determine the outcome of a process, for reasons which I
RESEARCH shall explain later, will hardly ever be fully known or measurable. [...]
[Using Mathematical techniques] has led to the illusion, however, that we can
use this technique for the determination and prediction of the numerical values
of those magnitudes; and this has led to a vain search for quantitative or
numerical constants.”
SOURCE: http://www.nobelprize.org/nobel_prizes/economics/laureates/1974/hayek-lecture.html
16. BOUNDED RATIONALITY:
“I KNOW THAT I DON’T KNOW”
management|consulting
In Economics the so called Neoclassical school postulated rational decision
making of the “homo oeconomicus” with perfect information available.
Herbert A. Simon Uncertainty about the future and costs in acquiring information in the present
were not considered part of rational decision theory. However do uncertainty
and costs limit the extent to which agents can make a fully rational decision,
thus they possess only “bounded rationality” and must make decisions by
BOUNDED “satisficing,” or choosing that which might not be optimal but which will make
RATIONALITY
them happy enough.
v
SATISFYICING
The internal organization of firms and the external business decisions thereof
did not conform to the Neoclassical theories of “rational” decision-making.
POLITICAL Bounded rationality is used to designate rational choice that takes into
BEHAVIOR
account the cognitive limitations of both knowledge and cognitive capacity.
Bounded rationality is a central theme in behavioral economics. It is
concerned with the ways in which the actual decision-making process
influences decisions. Theories of bounded rationality relax one or more
assumptions of standard expected utility theory”.
SOURCE: WIKIPEDIA.ORG
17. SH.... HAPPENS
management|consulting
Nassim Nicholas Taleb
LUCID
FALLACY
v
HINDSIGHT BIAS SURPRISE
DON’T BE THE
TURKEY
18. SH.... HAPPENS
management|consulting
Until 1697 all known Swans were white, so that the existence of a black swan was
considered impossible, until the discovery of Australia and with it the discovery of black
swans.
Nassim Nicholas Taleb Nasim Nicholas Taleb defines a black swan event as a surprise (to the observer), one
that has a major effect, and after the fact is often inappropriately rationalized with the
benefit of hindsight explaining:
LUCID
FALLACY
•The disproportionate role of high-profile, hard-to-predict, and rare events that are beyond
the realm of normal expectations in history, science, finance, and technology
HINDSIGHT BIAS •The non-computability of the probability of the consequential rare events using scientific
v methods (owing to the very nature of small probabilities)
1.The psychological biases that make people individually and collectively blind to
DON’T BE THE
TURKEY uncertainty and unaware of the massive role of the rare event in historical affairs
“Fat Tail Mitigation strategies
Distributions”
• Built robustness agains black swan events, exploit white swan events
• Avoid modeling based on normal distributions as risk is typically NOT normal distributed !
• “Avoid being the Turkey” - turn around black swan into white swan events.
SOURCE: WIKIPEDIA.ORG
19. SH.... HAPPENS
management|consulting
Then one morning Deadalus said to Icarus:
Nassim Nicholas Taleb
“Now Son, we are ready to leave this island for good. We
LUCID shall fly home to Athens. But although you are now quite
FALLACY good at flying, you must not forget that it can be very
dangerous. So listen to my instructions and be sure to
HINDSIGHT BIAS follow them to the letter. At all times follow me, for I will
v find the way home. Do not veer off on a different flight
DON’T BE THE
TURKEY
path, or you will soon be lost. And do not fly too low, or
your wings will fill with moisture from the waves and they
“Fat Tail
Distributions”
will become too heavy you will sink down. Nor should you
fly too high, or the sun will heat the wax and your wings
will fall apart. Have you understood all that I have said?”
20. SOLID DATA IS NOT EVERYTHING
management|consulting
Political behavior is an important contingency in enterprises. Strategic
Management is not a mere planning problem as intended strategies are
often not implemented as planned and deliberate strategies emerge
over time.
Kathleen Eisenhardt Clay Christensen
Preconditions of political processes:
POLITICAL •diverging interests among organizational members
BEHAVIOR
•limited amount of resources available to satisfy all such
interests.
RESOURCE
ALLOCATION v •Decisions with non-determined outcome
PROCESS •The larger the available decision space the more political
decisions tend to become, as outcomes require coalitions,
DYNAMISM negotiations and tactics between participants in the political
process.
While political processes typically negatively correlate with
profitability in high velocity environments, they can be a
source of corporate renewal that leads to higher
profitability. Resource Allocation Process
SOURCE: Christensen, C. M. & Dann, J. B. (1999). Process of strategy definition and implementation. Harvard Business Publishing.
Eisenhardt, K. M. & Bourgeois, L. J. B. (1988). Politics of strategic decision making in high-velocity environments: Toward a midrange theory. Academy of Management Journal, 31(4), 737-770.
Schreyögg (2008). Organisation - Grundlagen moderner Organisationsgestaltung [Organization - Foundations of modern organizational design] (5th Edition ed.). Wiesbaden: Gabler.
22. THE ROAD TO IN-MEMORY COMPUTING
management|consulting
George E. Moore
SOURCE: SINGULARITY.COM
23. ORIGINS OF OLTP AND OLAP
management|consulting
“Relational database systems have been the backbone of business
applications for more than 20 years. We promised to provide
companies with a management information system that covers
the core applications, including financials, sales, order fulfillment,
manufacturing, as well as human resources, which run from
planning through business processes to individually defined
Hasso Plattner analytics.
However, we fell short of achieving this goal. The more complex business
requirements became, the more we focused on the so-called transactional
processing part and designed the database structures accordingly. These systems
are called OLTP (Online Transactional Processing) system. Analytical and financial
planning applications were increasingly moved out to separate systems for more
flexibility and better performance. These systems are called OLAP (Online
Analytical Processing) systems.”
Plattner, H. (2009). A common database approach for oltp and olap using an in-memory column
database. In Proceedings of the 35th sigmod international conference on management of data.
24. SAP’s product landscape circa 2000 - 2005
management|consulting
Advanced Business
Planner & Warehouse
Optimizer (APO) (BW)
Supplier Customer
Relationship Relationship
Management ERP Management
(SRM) (CRM)
Logistics
Mobile Platform
Execution
25. OLTP AND OLAP ARCHITECTURES
management|consulting
OLTP - THREE TIER ERP SYSTEM OLAP - DATA WAREHOUSE SYSTEM
Data Cubes
Architectural Benefits Architectural Challenges
(+) Performance due to dedicated system (-) More Expensive through additional hardware
(+) Independent / No single point of failure (-) Double work for data cleansing, uploading, cube design, report writing
(-) Upload Windows often not sufficient in large scale installations.
Adopted from: Plattner, H. & Zeier, A. (2012). In-Memory data management: Technology and applications. Springer
26. USER EXPECTATIONS HAVE CHANGED
management|consulting
“At the University of Potsdam, I got bored with the presentation of traditional enterprise
v
software and the students didn't like it much, either; they wanted something more
modern, more like Google.” Hasso Plattner
Traditional Business Analytics In-Memory Business Analytics
Source: google-classic.com
28. WHY DO WE NORMALIZE AT ALL ?
management|consulting
Normalized Database Form (De-)Normalized Database Form
Flat File
SOURCE: http://www.codinghorror.com/blog/2008/07/maybe-normalizing-isnt-normal.html
29. SAP HANA - HIGH LEVEL ARCHITECTURE
management|consulting
Plattner, H. & Zeier, A. (2012). In-Memory data management: Technology and applications. Springer
30. COLUMNAR VS. ROW BASED STORAGE
management|consulting
Source: Plattner, H. & Zeier, A. (2012). In-Memory data management: Technology and applications. Springer
31. TECHNOLOGIES BEHING IMDB
management|consulting
Source: Plattner, H. & Zeier, A. (2012). In-Memory data management: Technology and applications. Springer
32. IMDB: RADICALLY SIMPLIFYING ENTERPRISE
APPLICATIONS (e.g. SAP ERP FINANCIALS)
management|consulting
Accounting Document
Accounting Document Items
Header
Future Table Structure in SAP
Current Table Structure in SAP ERP Finance
ERP Finance (Vision)
Source: Plattner, H. & Zeier, A. (2012). In-Memory data management: Technology and applications. Springer
SOURCE: Plattner, H. (2009). A common database approach for oltp and olap using an in-memory co lumn database. In Proceedings of the 35th sigmod international conference on management o f data.
33. BUSINESS BENEFITS (TCO)
management|consulting
On the fly financial aggregation, e.g. closing according to different
accounting standards (US-GAAP, IAS, etc), financial applications
faster and less complex. Provision of on-demand scenarios and
analytics allow frequent run of simulations and establish higher
business agility.
Simplification of overall IT landscape (one application server
instead of server farm with dedicated application servers)
resulting in less power consumption, cooling etc. - The solution
is easier to setup, scale and change.
Less complex software, through reduction of software layers
resulting in less maintenance and administration costs.
Allows the creation of innovative business solution for on the
spot decision making that were previously not feasible - online
personalised discounts.
34. DYNAMIC CAPABILITIES
management|consulting
Competitive Advantage based on organizational resources or capabilities
is not sustainable in high velocity environments, Dynamic Capabilities
thus become a critical differentiator for successful global enterprises.
Micro-foundations of Dynamic Capabilities (Teece, 2009, p. 49)
Source: Teece, D. J. (2009). Dynamic capabilities and strategic management. Oxford: Oxford University Press.
35. Case Study: SAP Location Strategy & Management
management|consulting
36. THE FUTURE OF DATA DRIVEN MANAGEMENT:
THE MANAGEMENT COCKPIT
management|consulting
support@v2softlogic.com
SOURCE: Controlling - Zeitschrift für die erfolgsorientierte Unternehmensführung, Vol. 18, June 2006, p. 311-318
40. YOUR PERSONAL SAP HANA CLOUD
DEMO management|consulting
SAP HANA VISUAL INTELLIGENCE
HANA Studio
http://www.saphana.com/welcome
41. YOUR PERSONAL SAP HANA CLOUD
DEMO management|consulting
How to get access to your personal SAP Hana Test Drive System?
1) Sign up with the SAP Community Network (SCN) at
http://scn.sap.com/welcome
42. YOUR PERSONAL SAP HANA CLOUD
DEMO management|consulting
2) Navigate to http://scn.sap.com/docs/DOC-28191, read the document and sign
up via the link at the bottom of the page
3) Accept the T&Cs
4) Confirm you data
5) Follow the instructions you
have received in your email
43. Now it’s your turn... SAP HANA Web
access management|consulting
PROFITABILITY ANALYSIS SALES COCKPIT
CENSUS DATA WITH GIS INTEGRATION
http://www.saphana.com/welcome
44. Use Case: Profitability Analysis
management|consulting
PROFITABILITY ANALYSIS Profitability Reports in the SAP ERP
Controlling Module (CO-PA) are what
managers are most interested in to
analyze profitability, over time, by
region, product group and customer
segments.
Traditionally these reports have a very
long run time in large enterprises.
This web based example shows the CO-
PA Accelerator in which CO-PA data
structures are copied into Hana.
This web based example with a real backend Hana
system
allows account manager, regional sales manager and sales
director to review critical profitability information.
http://www.saphana.com/welcome
45. Use Case: Sales Cockpit
management|consulting
SALES COCKPIT
Regular reviews of the Sales Pipeline and
analysis of sales performance are critical
for Sales Executives to safeguard revenue
generation for the enterprise.
Recent data is critical for territory
planning, account reviews and definition
and implementation of marketing
strategies.
Traditionally this data resides in SAP CRM
and reports have a very long run time in
large enterprises.
This web based example with a real backend Hana system
allows to assume the roles of senior sales director and vice
president of sales reviewing sales pipeline and sold revenue.
http://www.saphana.com/welcome
46. Now it’s your turn... SAP HANA Web
access management|consulting
Governments all around the world need
accurate data for provision of public
services, benefits, taxation and
infrastructure.
This SAP Hana application combines the
power of in-memory computing with a
Geographical Information System to
immediately visualize census data with
changes of the map. It also allows the
analysis and breakdown of census data by
CENSUS DATA WITH GIS INTEGRATION various dimensions.
This web based example with a real backend Hana system
allows to analyze annonymised real US Census data in a
geographical context.
http://www.saphana.com/welcome
Good morning everyone, my Name is Markus Alsleben and it ’ s a great pleasure to be here at Aalto University to talk about In Memory Computing for agile business intelligence. We should have plenty of time, so if you have any question along the way, feel free to ask.
This is today ’ s agenda, we have approximately three hours and I believe that we should cover most of the theoretical foundation in the first 1.5 hrs, then have a short break and continue afterwards with the introduction of SAP Hana, SAP ’ s in memory data base and several live demonstrations.
Every product must contain mobile access & strategy
An ever increasing amount of data is People talk about “ long data ” not only manage big amounts of data, but also to ensure longetivity.
There is clearly a trend to use mobile devices, even to a point that they are replacing conventional desktops. In SAP every product has to have mobile incorporated and must be mobile enabled. Mobile solutions enhance productivity and allow access to information anytime and anywhere.
One of the classic mobile business scenarios is sales force automation, in which the sales executive goes out to the customer and: - takes customer orders directly on a mobile device - checks availability to promise, or replacements for discontinued items - records his sales pipeline (opportunities, prospects, etc.) - plans a visit - etc. Here we see a typical day of a mobile enabled worker.
Neoclassical economics and decision theory often give the impression that decisions are always rational and consistent. As you as experts of business intelligence will be major decision supporter, I'd like to provide a word of caution. I therefore have selected five researchers that provide different perspectives on decision theories. I hope that you keep this in mind when preparing the next business case or business model for your company.
This is from an interview with Hasso Plattner one of the founders of SAP. Working for IBM in the early 1970s, they moved from customer to customer to always develop the same finanical accounting application, so that one day they thought about developing a standard product that could be used at every customer with minimal customisation effort instead of complete redevelopment. - The birth of the standard business application software product. However as Hasso points out, many of the initial design ambitions had to abandoned along the way.
Starting in the early 1990, SAP R/3 was designed in a three tier structure, where all software modules of the ERP system would run on a database server, application server and client computer to allow scalability for larger installations with 1000s of users. Around the year 2000, however, the internet boom made it necessary for SAP to open the ERP system to the internet and provide internet based functionality as e.g. catalog based buying via webbrowser, web stores as sales channel etc. SAP at that time couldn't accommodate the additional functions in the existing ERP product due to different release cycles and data structures, thus leading to separate systems that were connected to the core ERP system via interfaces/data replicators. Overtime additional products were developed outside the ERP system as standalone or connected systems, leading to ever complex landscapes.
Here we see a typical three tier ERP System with an attached data warehouse system. As you see data replication is required to ensure that the data warehouse system has the most recent data available. After data import, the data is stored in data cubes that have the reporting dimensions and characteristics that the queries that are run by the business user (e.g. CFO, CEO CxO) require. As you can see this design has several advantages and disadvantages. At a time when processing power and memory was expensive surely a feasible design. BUT.... technology and especially user expectations have changed !!!!!!!!
Hard to imaging that you would send a post card to google, as your parents did to wait 30 days for the printout of the search result. So the expectation of users these days is instant search results, as provided by Google and others. Allowing a "Trail and error" approach to find the right answers. Web searches however, sift through indexed data with only relevant data being presented. Business Applications however require complex aggregations, data thus needs to be prepared before it can be presented to the user. Data also originates from different sources (Finance/Materials Management/HR, external Data.)
When we look at traditional Business Intelligence Architectures, multiple steps are required before the business user can run analytics. Data need to be copied from the operational OLTP system, limitations: copy windows getting smaller, Data cleansing is required, as often data quality of OLTP systems is limited and often leading to inconsistencies in an analysis. (COO-CDM example cleansing of SAP's CRM System - conflicting definitions.) Data Cubes need to be predefined by developers to ensure that the right dimensions/characteristics of the data can be stores. Time intensive, requires IT specialists - often the bottleneck in running flexible analytics. Queries need to be pre-written, similar to a small IT project with User requirements, development, testing etc. - making ad-hoc queries and simulations difficult if not impossible.
Typical Database Structures of OLTP applications have highly normalized formats, to avoid redundancies, use less memory and reduce dependencies and speed up inserts. They however need percalculated aggregate views for performance reasons. Basically a design constraint from a time where both processing power and memory were expensive. Columnar Storage does not normalise data, as agregates are calculated on the flight, normalization is not required as memory is available in abundance and effective compression reduces the required memory by 30-50%.
- Unified Location Data Cube contained the 12 defined internal and external KPIS with location information to allow “ slice and dice ” by single locations or groups of locations. - Data cube provided SAP Location Dashboard (Webbased tool), the Excel Based Evaluation Tool (DSS) and the ARCGIS Desktop Version with data for analysis and decision making. - The web based tool was received enthusiastically by stakeholders including the worker ’ s council as an intuitive way to visualize and analyze the global setup of SAP. - The Evaluation Tool allows managers to compare different locations through a scoring model based on the data provided by the unified location data cube while the professional grade ARCGIS Desktop Software allows more comprehensive geospatial analysis and creation of specific maps.