SAP HANA and Real Time Analytics_BI
-
Upload
aslamharys -
Category
Documents
-
view
38 -
download
0
Transcript of SAP HANA and Real Time Analytics_BI
Version Feb 2, 2011
A BVA Knowledge Share Paper BVA = Business Value Attainment
SAP HANA
& Real Time Analytics
By Hari Guleria
BI Solution Manager
Ask how HCL AXON can assist - Meet Business Expectations
SAP HANA
& Real Time Analytics
Hari Guleria
BI Solution Manager
Meet Business Expectations with your HANA initiative
2 | P a g e
Contents
Introduction 2
Problem Statement 2
Alternatives 2
HCL Solution 2
Introduction to Hana 3
Is Hana the Silver bullet to all SAP BI? 3
Will HANA eliminate the need for BW? 3
What Berlin said about HANA 4
Why Data Quality? 4
Data quality issues in In-Memory solutions 5
HANA, BW, BWA & Accelerated Explorer? 5
HANA 1.0 - Where to next 6
Business Case - Utilities 6
Conclusion 8
About the author 8
Introduction
In Memory and Real-Time Informatics are the big growth area is
global Performance Management and Key Decision Enablement.
With the introduction of HANA SAP takes a leap into real time
Informatics as a global leader in In-Memory & Real-Time
Informatics
Problem Statement
How should a customer decide when HANA becomes a business
asset for a corporation?
Alternatives
BW Accelerator was a previous alternative. However, BW
Accelerator could only connect to the SAP BW Accelerator or
BWA.
HANA on the other hand connects directly to the operational
system and provides critical real-time analytics for decision
enhancement..
HCL Solution
The HCL Solutions is based on the following:
Benefit 1
Review and deploy a ‘Business Value Attainment’ Solution. This
will ensure Hana is deployed for business reasons and not as a
technology deployment
Benefit 2
Global SAP Services Partner. This assures customers that all HCL
AXON solutions are pre-aligned with SAP’s Best Business Practices
Benefit 3
High experience and leadership in SAP BI Alternatives
deployment; Business Solutions and ‘Meet Business Expectations
in BI’ methodology.
Introduction to Hana
Both at Las Vegas and Berlin SAP TechEd’s were
dominated by HANA. Interspersed with the brand-new
launch of NetWeaver 7.3 and BusinessObjects 4.0, HANA
was the main push at Berlin. HANA for High Performance
Analytic Appliance takes SAP from BW Accelerator as an
BW performance appliance to HANA as a bigger
operational appliance.
The current version of HANA is designed as version 1.0
and will enable high performance, in memory processing
from your SAP ECC system, and other disparate source
systems.
HANA is the next big thing even as the dust on BW
Accelerator and BusinessObjects Explorer has not yet
settled.
It is certain HANA will bring new frontiers in high-
performance analytical processing and new opportunities
to SAP and their customers; however its introduction is
raising a lot of important questions about things like data
quality, data warehouse or BW, and the appliance
promotion from SAP. In recent presentations and
communications SAP has started communicating an old
theme when asked how things work inside a
BusinessObjects Universe, WebI, BusinessObjects
Explorer, BW Accelerator and now HANA – their answer is
a constant; These are appliances developed by SAP and
they will work perfectly. You have to see how you can
leverage the appliance and not waste too much time in
trying to figure its innards.
On one side of the equation an appliance is excellent,
however on the other side of BVA (Business Value
Attainment) it becomes necessary to know how and why
things work so one can leverage the tools capabilities by
understanding what the tool can and more importantly
cannot do.
Is Hana the Silver bullet to all SAP BI?
The simple answer is No.
Hana is to ECC what BW Accelerator is to SAP BW. Both
are expensive technologies and must be deployed for
business reasons in order to meet, and exceed, business
expectations. For more years than one can remember we
have heard of the ultimate solution and then realized that
there is a gap between marketing and capabilities. In 1998 SAP
BW was proposed to answer any question an executive could ask
and yet twelve years down the line we are still struggling at a 50%
success factor. The same can be applied to BW Accelerator and
BusinessObjects Explorer. So let us understand HANA and how it
is both similar and different in its BI capabilities.
Will HANA eliminate the need for BW?
This is a question that surprisingly keeps creeping up from all
sides and customers. However, we can only see so far ahead. My
current statement will be not for the next five to eight years
based on two basic myths that must be bust right now:
1. If we take all the ERP fields into a data warehouse, or an
appliance, we should then be able to answer every
question.
a. This is a myth that is still prevailing in the data
warehouse world. It is most applicable to the
Accelerated Explorer and now to Hana. Not
considering this will result in a costly mistake. All data
can answer is simple questions. However, large
organizations do not live by simple answers alone.
Data filled appliances cannot answer complex
questions that make modern enterprises run ‘Give me
the stock position for the next week considering all
operational variables like production schedules, stock,
returns, customer bookings, requisitions, PO orders in
process, etc..
Release 1.0 of HANA is for non-disruptive and very fast
Operational Reporting. Operational reporting normally
consists of simple reports and to meet day-to-day
reporting needs.
b. HANA version 1.0 is for ECC data access only
c. HANA leverages existing technologies of BW
Accelerator, BusinessObjects Explorer and
BusinessObjects WebI along with Sybase database to
deliver real-time informatics in critical areas. The key
here is critical areas and how business defines this
singularity for using HANA.
d. Data warehouses enable complex transformations,
summarizations, exceptions and a host of other
filtering that enables a ‘SPOT’ (Single Point of Truth)
answer to critical and complex questions. So far as
executives need answers to complex questions data
warehouses will remain – unless of course we get a self
modeling InfoCube that remodels itself on information
demand and consumption.
What Berlin said about HANA
Vishal Sikka, SAP CTO, outlined [1] on-demand, [2]
mobility and [3] in-memory as the three technology
priorities across the global enterprises. All three belong to
the domain of Business Intelligence and analytics.
HANA was aligned with ‘Non-Disruptive’, just like
BusinessObjects Explorer was aligned with CLEAR by Hasso
Plattner and Leo Apothekar (now HP) a year or so ago. The
focus at Las Vegas and Berlin has been HANA. Vishal Sikka
quoted a meeting he had with Hasso Plattner prior to the
Berlin TechEd in his keynote speech.
HANA’s key value proposition is it’s ability to support
operational BI in a non-disruptive manner. It manages this
by leveraging SAP new acquisition Sybase. HANA uses the
Sybase’s replication Server for real-time replication and
synchronization of a SAP ERP system by loading all the
data into In-Memory appliances for real-time analytics
capabilities.
By running parallel to the source SAP ERP application,
HANA enables business users to query
extremely large volumes of
operational data in near real time.
Just like in the BusinessObjects
explorer there will be no need to pre-
aggregate or filter any data. At the
same time HANA will use SAP’s acceleration technology
for query performance and exceptional speed. HANA’s
support for standard SQL and MDX means it can also work
with a BusinessObjects client such as WebI or Crystal
Reports, however it must not be forgotten that MDX
cannot read OLAP so hierarchies from BW will still be a
problem until BICS becomes a standard read for HANA in
version 1.5 when HANA will integrate with SAP BW.
With the introduction of BusinessObjects 4.0 the MDX
issue will be replaced with a BW compatible BICS
technology. This will make reading BW, ECC and OLAP
hierarchies very efficient in BWA and HANA.
SAP is focusing all their energy into In-Memory computing.
From BW Accelerator, to BusinessObjects Explorer and
now HANA. The HANA platform comprises of SAP BAE
(Business Analytics Engine); an in-memory columnar data
store with compression technology very much like the BW
Accelerator with current compression ratios of 10x. So a 1
terabyte data source can now be compressed to 100
gigabytes, and with a 50% space for computation it would
mean a 200GB space requirement. All this combined with
partners like HP and IBM in the near future.
So as the world progresses with better and better technology
solutions, in-memory computing, multi-core chips, parallel
processing, true 64 bit computing, columnar databases I still need
to ask one simple question – What steps are we taking to assure
that all this technology helps your enterprise meet business
expectations?
Very much like a BW implementation that does not meet business
expectations and recommends installation of a BW Accelerator to
solve their problems. All the group succeeds in doing is delay the
inevitable, i.e. not meet business expectations – only this time
faster. So the BI team goes and installs a BW accelerator and gets
their reports in 30 seconds instead of 500 seconds. The net result
of all their investments all that what the users got faster were
reports that they could not use initially – only this time faster.
The goal of every BI initiative must start with ‘Meet Business
Expectations’ HANA is no exception.
The goal must not simply be lightening fast analytics but first
meeting business expectations and
then taking that to lightening speeds.
Business participation thus becomes
imperative in HANA as it does in other
BI initiatives.
Why Data Quality?
HANA’s sweet spot is high performance reports and analytics from
operational data. As BI solution architects we understand that not
all data from every operational systems is squeaky clean from the
start. This issue becomes all the more complex as we merge more
than one source system. HANA will soon be accepting data from
more than ECC and we have to plan for that point in time. It also
becomes more complex as we find different systems having
different ID’s for the same master data or the same ID’s for
different Master Data. By simply replicating all the data from an
ERP, and other operational, systems we also stand to replicate
any data quality issues that may reside in the ERP system/s.
Also if HANA is simply deployed as a Technology solution, on a lets
wait and see principle then it may turn out to be a very expensive
investment. However, if we work with business and identify
critical, real-time business needs and then further prioritize then
HANA has the potential to become an informatics asset as no
other technology has ever been before.
HANA has the potential to become
an informatics asset as no other
technology has ever been before
SAP may have inadvertently taken the ‘Technology as the
final solution’ and business stakeholders need to review
the operational and technical capabilities before
approving the HANA project as a BI alternative. This is just
brought up as a point of reference and a check-list point.
Data quality issues in In-Memory solutions
While SAP points out the benefits of using HANA they
also identify data quality issues. Their current
recommendation is to correct this at the source system
and this may not be a practical or perfect solution right
now. With Data Governance kicking in between the HANA
deliverables and source system data issues we identified a
lag in data quality, i.e. what HANA has delivered vs. what
needs to be corrected at the source systems to rectify the
data issues. While correcting data issues in the source
system may eradicate some of the problems the time and
complexity factor of making changes in any source system
are all too familiar to the data warehouse world. That is
the reason we have the Transform as part of the ETL
sequence in BI. At the same time it is critical to realize that
HANA may not turn out to be the preferred application for
resolving data quality issues since it leverages the
columnar centric data store approach which is not ideally
suited for data transformation or corrections. Its sole
purpose is performance.
A possible scenario might be to leverage the use of a data
warehouse and an intermediary point to resolve data
quality issues and then push the data into HANA for
performance. This tends to align with SAP’s current
thinking around the ‘Future State’ use of the data
warehouse.
HANA, BW, BWA & Accelerated Explorer?
One of the questions customer have been asking, and we
do not currently have a clear answer to this one is that -
despite all the HANA and Explorer extolling on
performance how does HANA fit into the overall picture of
companies that have invested into BW, BW Accelerator
and BusinessObjects Explorer. How does HANA assist in
‘Meet Business Expectations’ in the overall enterprise BI
Equation?
One thing that is certain is that the role of the data
warehouse is becoming more and more solidified. When
executives need integrated, harmonized, cleansed and a
consolidated view of their business Analytics or global
operations the current data warehouse is their only
option.
While customers continue to ask whether HANA will replace BW
or the BW accelerator we are witnessing a clear pattern that SAP
is continuing to invest heavily into the BW (their data warehouse
platform) that already has a 10,000 plus customer base. In this
sense BW remains the company’s go forward strategy for Reports,
Analytics and Informatics. It remains the option where companies
need a persistent data layer.
Thus, while HANA is designed to meet the gap between the last
extract and the current reality with high-performance virtual view
of the ERP data, it will complement and not supplement the BW
data warehouse and is not planned to replace the functionalities
of a Federated Enterprise Data warehouse or the old time EDW
(Enterprise Data Warehouse)
However, the question that still requires clarification still hangs in
the air. What is the relationship between HANA and BW
Accelerator, and SAP’s other In Memory products like
BusinessObjects Explorer.
The current answer is that BW Accelerator is an accelerator for
pre-existing BW Queries. It can do little more than that. The BWA
can also be used to accelerate BOBJ (BusinessObjects) Web,
Crystal and other analytics if the source of the data is from BW.
The BWA can only report on data in BW. By the way, SAP has
already announced that HANA v.5 will support BW, Explorer and
ECC blades. Whether the current BWA can be upgraded to HANA
v.5 is still to be confirmed. The current assumption, going on
SAP’s history should be a yes, but this is an unqualified statement.
HANA, on the other hand, is exactly like the BW Accelerator
except it connects directly to the ECC environment and will run
real-time analytics from data in the ECC.
Scenario 1- BW Accelerator: Is the BWA scenario where BWA is
simply used to accelerate existing query performance of SAPBW
analytics. This can assist BW, Business Objects and the Explorer all
of which can use the standard versions and the accelerated
options. This is done by loading the entire BW InfoCube (SAP’s
extended star Schema format) into ‘In-Memory’ and thereby
leveraging the BWA’s capabilities to index for faster response to
the same queries.
Scenario 2- BOBJ Explorer: Is the accelerated Explorer where data
is fed to the BIBJ Explorer via the BW Accelerator for search and
Analyze in fixed format Analytics.
Scenario 3: HANA on the other hand is currently targeted at the
SAP ERP/ECC environment. HANA is for customers who want
extremely high performance operational reporting straight from
the source systems, i.e. real-time POS (Point of Sale) data for a
retail organization. SAP is positioning HANA as an evolution of
BWA and as an option to open high-performance, in-
memory options to non BW users.
The challenge for most architects and business
stakeholders right now is how to articulate where HANA
fits into the overall ‘Enterprise Wide Business Intelligence
Landscape’ especially when considering investments made
into BW, BWA, accelerated explorer – without further
confusing the customers.
HANA 1.0 - Where to next
The current HANA release is HANA 1.0 and HANA is
gearing up for greater ‘Future State’ options. SAP has
ambitious plans for HANA. SAP plans to use In- Memory
computing for financial planning, simulation scenarios and
real-time inventory management.
Likewise the core in-memory computing engine used
within HANA has the potential to power newer versions of
SAP NetWeaver, allowing in-memory technology to be
used across all of its product development roadmaps for
both transactional and analytical applications. SAP is
betting heavily on HANA and in-memory computing and
combining this with cloud computing, appliances and peak
performance.
Business Case - Utilities
Example 1: DRILL THROUGH
Utilities are investing into smart meters that are
dispatching data every 15 minutes, or so depending on
ETL plans and settings, from each household. The data
here is a representation of monthly smart meter peaks
and troughs based on smart meter readings
What we see here is 1.8 billion records of data from smart
meter being viewed in real-time.
Action 1: By Simply passing your mouse over a selected
view of data the user can drill through into the detailed
data from the selection. In order to do this the user can
continue to drill into more and more detailed data.
The example here is about drilling down into a few months and
then further drilling into a few weeks or even a few days in a
matter of seconds.
The screen below shows the user selecting a range of visual
months
Here we then start to see week
.. and further drill down into days and hours
In this example 1 we demonstrate how a business stakeholder
can view very large volumes of data and drill down from years
into months to weeks and then from days to hours within a
matter of seconds
Example 2: MARGINS & Profitability Analytics
On the left of the HANA Informatics we have user
selections as shown below and the user moves from usage
to ‘Margins’
Now when the user selects Margins, the HANA engine
runs through our 1.8 billion records and in a matter of 3 to
4 seconds returns with a Margins analysis for business
decisions. We immediately see the Margins view as shown
below
Now the user decides to add the financial data with the
margin data for the period being reviewed. Once again
within seconds the HANA engine displays all the financial
data along with Margins
So in the above chart we see the Profitability, which is the
green line below, we see the Revenue Generated for the
same period which is represented by the Blue line and
then we look at the Cost of providing required energy shown by
the Brown line.
Now margins in the Utility industry are extremely critical, as they
are for all industries, but HANA gives the ability for key
stakeholders to commence with different aspects and how it
impacts the overall margins for the company.
So what we find in this chart is very interesting where the
profitability at certain times goes below ‘Zero’. When we look at
costs we find that during the same period the costs have gone
extremely high. Further analysis shows that during peak hours the
utility buys power, often at up to a hundred (‘00) times higher to
buy at the spot market than what it would cost to produce
internally.
Suddenly this report becomes a critical piece of evidence for the
business stakeholders to review adding capacity or continuing to
erode profitability by routinely buying energy from external
suppliers.
It is excellent to see when the utility is procuring energy in the
spot market.
At the same time the user can also take any single month and
review customers to see which customers were profitable –
represented by green, who were marginally profitable-
represented by yellow, and who was unprofitable- represented by
red, during any given period. The3 graph below shows this
analysis
The user can now click on the unprofitable customers and
instantly get a list of the entire unprofitable customer base (for
the selected period) as to names, how much they consume and so
on and so forth by different categories
This type of analysis can now assist the marketing
department to plan promotions, pricing conditions or
special offers to change this paradigm and convert them
into profitable customers.
Conclusion
SAP HANA has exceptional potential if planned and
leveraged for its business potential and not as another
technology deployment.
Like most BI initiatives HANA too needs business
comprehension, business capability analysis followed by
identifying what true benefits can be realized with HANA,
that traditional BI cannot provide. The key here is real-
time and not simply reports, analytics or informatics.
About the author
Hari Guleria is a BI Solutions Manager with HCL AXON and
routinely works with Fortune 1000 companies and their BI
initiatives. He is the author of BI Valuenomics – the story
of meeting business expectations in BI. Prior to this Hari
worked with SAP America and Anderson Consulting. Hari
comes with over 9 years of senior business management
experience, 15 plus years of SAP and 13 plus years of SAP
BI. Hari has implemented projects in BW, BW Accelerator,
BusinessObjects and Oracle DW. He may be contacted at