1310 success stories_and_lessons_learned_implementing_sap_hana_solutions
DESCRIPTION
hana successesTRANSCRIPT
September 10-13, 2012
Orlando, Florida
Success Stories and Lessons Learned Implementing SAP HANA Solutions
Session 1310
2
About Me
� Jonathan Haun Consulting Manager with Decision First Technologies.
http://www.decisionfirst.com
� Over 12 years implementing BI solutions utilizing Crystal Reports and
BusinessObjects
� Experience with multiple ETL tools and RDMS
� Certified in multiple versions of BusinesObjects Enterprise, BusinessObjects
Data Services and BusinessObjects Reporting tools
� Certified SAP Trainer
� Over one year of experience implementing BOBJ and BW solutions on SAP
HANA
� Manager of the “All Things BOBJ BI blog”. http://bobj.sapbiblog.com
� Twitter Feeds @jdh2n http://twitter.com/jdh2n
3
Learning Points
� What is SAP HANA ?
� Solutions Currently Available based on SAP HANA
� Tips and Tricks implementing solutions on SAP
HANA standalone
� Performance Improvements
� ROI for Solutions Based on SAP HANA
What is SAP HANA ?
5
What is SAP HANA ?
� SAP HANA is a true next generation database. It is a fusion of both software and
hardware that is fully engineered to provide deep and rich access to billions of rows of
data.
� Data on SAP HANA is stored and accessed in RAM which eliminates the traditional
limitations of spinning magnetic disks.
� Data is stored physically close to the CPU allowing faster and more direct access to the
data.
� The HANA software is the first of it’s kind to be built from the ground up to support
multi-core CPU’s and in-memory direct access to data. Memory First, Persistent Storage
Second.
� Data can be stored using either Row or Column based storage providing flexibility in
supporting both Front End Applications and Business Intelligence on a single platform.
� SAP HANA has built in Multi Dimensional Modeling Views that can mimic the capabilities
of OLAP without the need to duplicate data into Cubes. This reduces the TCO of storing
the data while providing the flexibility to perform both MDAS or Operational reporting.
6
Solutions Currently Available based on SAP HANA
7
Solutions Currently Available based on SAP HANA
� SAP HANA Standalone
� Leverage the power of SAP HANA with data from any source.
� Move data into SAP HANA in real-time or in batch using replication or Data Services
4.0
� Highly customizable for any data analysis requirements
� Supports Analytics, MDAS, Forecasting and Operational reporting in a single solution
� Analyze billions of records at amazing speeds using SAP BusinessObjects 4.0
� Similar to the Traditional EIM strategy of building Data Marts
� SAP Netweaver BW 7.3 powered by SAP HANA
� Run BW directly on SAP HANA by replacing your existing RDMS with SAP HANA
� Reduce the overall data footprint of BW with direct integration with SAP HANA
� Easy solution to adopt for existing SAP BW and SAP ECC customers
� Load 3rd Party Data in BW utilizing Data Services
� Analyze billions of records at amazing speeds using SAP BusinessObjects 4.0
8
Solutions Currently Available based on SAP HANA
� Multiple rapid solutions based on your Industry or Line of Business
� COPA, CRM, Finance, ERP, Cash Forecasting and many others. See more at
sap.com
� Over 20 prebuilt solutions and growing
� Prebuilt solutions that are easy to deploy while reducing development costs
� Ability to manage billions of transactions with sub-second response times
� Analyze billions of records at amazing speeds using SAP BusinessObjects 4.0
9
� Data Modeling vs Analytic Modeling� Understand the difference before you begin developing solutions on SAP HANA standalone.
Modeling in HANA centers around creating rich MDAS views, forecasting , and complex cross
fact calculations.
� When loading data in batch, model your data into a traditional Star Schema or Fully De-
normalized Table before loading SAP HANA. Joining hundreds of physical normalized tables
will reduce the performance of queries significantly. Also Attribute Views that contain millions
of records should be modeled into the Analytic Foundation tables to increase performance.
� Project Experience:
� 100% Analytic Modeling: Normalized Data loaded into HANA, containing multiple tables and
joins and modeled completely in SAP HANA: 16 sec response times.
� 90% Data Services Data Modeling: With the same data, data modeled into a highly
deformalized table, with Attribute Views containing less then 10,000 records and fewer then 10
total joins produced results in < 1 sec.
� Data Quality must be addressed via business processes and Data Services
� Merging Dimensional Data from multiple sources should be addressed using Data Services
� Reverse Pivots to de-normalize data should be addressed via Data Services
� Hierarchy Parsing should be addressed via Data Services.
Tips and Tricks for SAP HANA standalone
10
� Using More SAP HANA “Analytic Modeling”
Tips and Tricks for SAP HANA standalone
11
� Using More Data Services “Data Modeling”
Tips and Tricks for SAP HANA standalone
12
� Analytic Modeling Tips� Derived Attributes allow you to make copies of an existing Attribute View. The copies can not
be modified and will always represent the configuration of their master. They are great for
Date and Time based Attributes that are needed to create joins between multiple date
columns found in the Analytic Foundation. When you update the master Attribute all child
attributes will inherit the master’s columns. Similar to creating an alias of a table.
� Joins between an Attribute View and Analytic Foundation can not be defined on Decimal(x,y)
columns.
� Joins on varchar or text columns do not perform as well as joins between Integer columns.
� SAP HANA has a built in Date / Time Dimension (Attribute View). There is no need to use Data
Services to create a calendar date dimension.
� Joining multiple tables in the Analytic Foundation will greatly reduce performance. Better to
use Data Services to combine the tables or Calculation Views to merge two independent
Analytic Views.
� Every Column in an Analytic View must have a unique name. It is best to implement a naming
standard before you begin your end-to-end Analytic View design phase.
� Be cautious making changes to existing Analytic Views (Post Development). Upstream content
will stop working until it is re-activated. In some cases, removing columns, will require re-work
on up-stream content.
Tips and Tricks for SAP HANA standalone
13
� Universe Design Tips� Using Analytic Views as the source for your Universe is great when that Universe is used
primarily for Dashboards, Analytics and other reporting needs that contain Group By and
Aggregate functions. All queries executed against an Analytic View must contain a measure.
When you use an Analytic View within a Universe very little Universe design is required as
most of the modeling and metadata were already defined in SAP HANA. It is also a good idea
to leverage Analytic Views as current versions of SAP BusinessObjects Explorer and future
versions of Webi will bind directly to SAP HANA Analytic Views.
� Using SAP HANA Columnar Tables within a Universe (Traditional Universe Design) will provide
more flexibility for a variety of reporting requirements. Best if used for Operational Reporting,
Multi-Fact Data Synchronization and situations where measures are not required.
� Multi-Fact Data Synchronization can be pushed down to SAP HANA by setting the
JOIN_BY_SQL parameter to true within the Universe.
� The SAP HANA ODBC / JDBC drivers are only supported in the Information Design Universe
(IDT.UNX)
� ODBC is easier to setup and maintain on Windows based BOE environments
� JDBC is easier to setup and maintain on UNIX / LINUX environments.
Tips and Tricks for SAP HANA standalone
14
� BW powered by HANA� Frequent SAP HANA patching was required during the initial POC. However, the current patch
level (Patch 33) has proven to be more stable and we expect fewer and less frequent patching
as the product matures. SAP has been very diligent in resolving issues but I would recommend
adding time to any project estimate to account for unforeseen issues.
� Hardware vendor selection is important. Watch for vendors recommending supplemental
hardware components. (Network Switches Etc..) Make sure that all supplemental hardware is
compliant with your current environment.
� Frequent SAP HANA database settings were changed during the POC to find the right mix to
support the customers in-house developed BW processes. Loading and processing of complex
BW flows may require increasing parallel treads and memory settings on SAP HANA.
� Some BW modeling optimizations were required to fully support BW on HANA. In most cases
the modeling code was poorly developed (not following best practices) in other cases SAP BW
and or HANA required patches. Note: These issues had little or no impact on the project.
Observations BW powered by SAP HANA
15
Performance Improvements
Baseline Faster Report Speed Faster Report Speed BW Powered by HANAHANA Standalone
Database
Report Response Times: • Reports: • Reports: • Reports: • Reports:
•Avg 90 sec •Avg 40 sec •Avg 32 sec •Avg 6.4 sec •Avg < 1 sec
•CRM 400 sec •CRM 161 sec •CRM 23 sec •CRM 4.6 sec •CRM < 1 sec
•Data Loads 15 hrs •Data Loads 15 hrs •Data Loads 15 hrs •Data Loads 4.6 hrs •Full Loads 3.2 hrs
•Reports are faster then
BWA
•Delta Data Loads are
much faster (3.2x)
BW HANA With BOBJ
4.0 SP4
HANA Stand Alone
With BOBJ 4.0 SP4
•Reports are faster
then BWA and BW
Powered by HANA
•Data Loads are very
fast
BW with Legacy DB2
database. No BOBJ
Legacy DB-BWA With
BOBJ 3.1 SP2
Legacy DB-BWA With
BOBJ 3.1 SP4
•No BWA or BOBJ
Reports
•BWA 2.5x faster
•Early version of BOBJ not
stable
•No Improvement on Data
Loads
•BOBJ 3.1 SP4 very Stable
•Reports are faster due to
Query Striping
Customer’s History & Performance Gains
* Directly Observed during recent POC with customer
16
Performance Improvements
0
50
100
150
200
250
300
350
400
Base Line BWA BOBJ 3.1 SP2 BWA BOBJ 3.1 SP4 BW HANA HANA Standalone
AVG (Sec)
CRM (Sec)
Query Response Times
* Directly Observed during recent POC with customer
17
Performance Improvements
0
2
4
6
8
10
12
14
16
Base Line BWA BOBJ 3.1 SP2 BWA BOBJ 3.1 SP4 BW HANA HANA Standalone
Load Times(Hrs)
Data Load Times
* Directly Observed during recent POC with customer
18
� Reduce Expensive SAN storage by moving data into memory with
compression.
� Better integration of BW with the SAP HANA database appliance
means more operations are managed by the in-memory database
without the need to duplicate data multiple times.
� Multi-Providers on DSOs, in many cases, will work as well (if not
better) then cubes on BWA. Further reducing the data footprint.
� Near Line storage, directly integrated into BW 7.3, can be utilized for
infrequently accessed data. Allows for BW to maintain multiple
storage tiers at different costs levels.
� User will experience faster response times making them more willing
to deep-dive into the data. This will keep them focused on the
Analysis task while increasing their knowledge and productivity.
ROI for Solutions Based on SAP HANA
19
ROI for Solutions Based on SAP HANA
Object Type BW on DB2 Step 1. Cleaning up to move to HANA
Reduction in System Tables
Decommission Obselete Objects
Reduce Layers with HANA
Near line
PSA 5,842 100 100 100 100 100
Legacy DB Overhead 6,141 - - - - -
Change Log 4,174 - - - - -
DSO 1,487 1,487 1,487 1,312 1,201 1,201
System Tables 1,167 1,167 300 300 300 300
Cube 708 708 708 648 591 141
Master Data 408 408 408 408 408 408
Temp 7 7 7 7 7 7
Total (GB) 19,934 3,877 3,010 2,775 2,607 2,157
6 x Comp 3322 646 502 462 434 359
Reduced Data Footprint with BW on SAP HANA
* Directly Observed during recent POC with customer. Your results might very based on the type and makeup of the data.
20
ROI for Solutions Based on SAP HANA
ITEM SAVINGS
SAN (Fast Storage) Eliminated (Very Expensive)
DB2 (Database) Eliminated
Total # of Servers Reduced by 30 -50%
SAP Licenses and Hardware Increased Hardware and Software
Staff Reallocated (DB2 and Storage)
BOBJ No Change
Total Estimated Savings (5 yr) 30-40% reduction in TCO
Overall Cost Savings Estimates
* Directly Observed during recent POC with customer
21
ROI for Solutions Based on SAP HANA
Happy and Productive Users
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
User Acceptance
User Acceptance
22
� SAP HANA is a fusion of both software and hardware. It was built from the
ground-up to support in-memory data storage. Legacy vendors are still trying to
develop in-memory solutions based on legacy DB architecture.
� There are three main solution categories for SAP HANA. SAP HANA standalone,
BW powered by HANA and a wide variety of Rapid software solutions.
� For most organizations, SAP HANA solutions will yield significantly faster query
response times.
� There will be an opportunity for most organizations to reduce their TCO for BW
when powered by SAP HANA.
� User Acceptance is hard to measure in dollars, but when users are able to use
BI in seconds and not minutes, they will be more productive and willing to
leverage these tools.
Key Learnings
23
� Data Modeling vs Analytic Modeling in SAP HANA – “All Things
BOBJ BI” http://wp.me/p2868w-4W
� Creating Analytic Models in SAP HANA Studio - Question and
Answer Responses – “All Things BOBJ BI”
http://wp.me/p2868w-54
� SAP BusinessObjects Explorer 4.0, powered by SAP HANA – “All
Things BOBJ BI” http://wp.me/p2868w-2P
� SAP HANA FAQ – “All Things BOBJ BI” http://wp.me/p2868w-N
� https://www.experiencesaphana.com – SAP Hosted Site with
great information on SAP HANA.
� SAP HANA product page - http://www.sap.com/solutions/technology/in-memory-computing-
platform/index.epx
Links and References
Thank you for participating.
Please provide feedback on this session by completing a short survey via the event
mobile application.
SESSION CODE: 1310
Learn more year-round at www.asug.com