There was a tweet from Jon Reed about a question at SAP TechED Madrid, asking if SAP Hana and Sybase IQ are like Woody Allen’s movie “Hannah and Her Sisters”. It reminded me of this great webcast from late August of last year, which I will summarize below.
This was an ASUG webcast that SAP’s Balaji Krishna and Yuvaraj Raghuvir provided to ASUG last year. The information may have changed since this webcast and of course the usual disclaimer applies, the information is subject to change.
On the agenda:
- Sybase IQ positioning
- Why SAP BusinessObjects & IQ Together
- Integration points for IQ and BusinessObjects BI
We have an information explosion; an explosion of information and speed with which business needs to react. In a perfect world the information would be current and complete so business decisions are most accurate and looking for an immediate answer. It has been expensive to improve performance – the information access, complex systems, duplicating information, lengthy validation approval process
Sybase IQ is open, ANSI SQL compliant. It has data compression, is columnar, it is easy to compress and can get real time feeds with Sybase Replication server
Figure 1, Source: SAP
As Figure 1 shows the lowest layer is the Data Warehouse with volumes of data. EIM turns data into quality information you can trust. BI “turns data into insight”.
Figure 2: Source, SAP
Figure 2 shows several options with SAP Data Warehousing and analytical databases. SAP says “complement” each other. All offerings have been integrated with BusinessObjects and EIM for analytical reporting initiatives. The right side of Figure 2 shows the offerings.
Figure 3, Source: SAP
Figure 3 shows a “paradigm shift”, where we are moving from a rear-view to a windshield approach to look forward to make better decisions. Analytics is everywhere.
The right side of Figure 3 shows the scenarios such as calculating profitability of millions of customers everyday.
Figure 4: Source: SAP
Traditional warehouses cannot get to the new paradigm such as senior management not getting the information they need and manual analysis in Excel and Access.
Figure 5, Source: SAP
Sybase IQ has more than 1900 customers as shown in Figure 5. It is mature, offering data compression and “moves data closer to the user”
Examples are shown on the right of Figure 5 including analyzing all federal tax returns in the US.
Figure 6, Source: SAP
Figure 6 shows “what the analysts say”. What it means – distributed query processing in 15.3 and 15.4 – federated across storages
Federated analytics – Sybase IQ is able to do a distributed query across multiple node storages.
Figure 7, Source: SAP
What is under the hood of Sybase IQ. It is designed for analytics “from within”.
Each column is stored separately, with a unique column based. It is ANSI SQL compliant
Queries are flexible and ad hoc. Optimized storage with terabytes of information, queries with only columns used means save on # of CPU’s and storage. It supports any data schema. It is designed “internally to suit analytical performance”.
Figure 8, Source: SAP
How does this fit into overall business analytics with SAP?
In the area of event processing and data replication there are Sybase products (Sybase Replication Server), coming together with SAP BusinessObjects Data Services.
Figure 9, Source: SAP
Figure 9 shows some leading use cases for Sybase IQ such as organizations with overburdened transaction systems. It involves a proof-of-concept to show how IQ can add value using customer’s data. It is a complementary solution.
Advanced analytics requires sophisticated processing and see an advantage of Sybase IQ. Web based shops want to provide analytics to end user (scenario 4 in Figure 9).
Figure 10, Source: SAP
Figure 10 shows accelerated reporting systems with a variety of data sources (on the left). There are different options and then offload reporting data to Sybase IQ.
Figure 11: Source, SAP
Figure 11 shows moving from the source system to the data warehouse, VLDB scaling and deep analysis, the columnar store allows for speed for complex queries. It includes tools to support decision-making and visualization, using IQ as a data mart . It is a centralized data warehouse, using federated data marts, consuming other sources of information.
Figure 12, Source: SAP
Figure 12 shows an example of data compression using 1 TB of data in a database, need to build indices which take up space. You need to build summary tables, which adds more space.
With Sybase IQ, few indices need to be built. Using the same data, Sybase IQ will store it in less than a TB
Figure 13: Source, SAP
Figure 13 shows how you can use the text index and third party plug-ins
Text Search, Text Analytics and Text mining are some of the add-ons, to combine text analytics with in database.
Figure 14, Source: SAP
On the left you have Data Quality, Data Services, Data Profiling, use to load from multiple data sources on to Sybase IQ
Complex Event Processing (CEP) engine that comes with Sybase IQ used for Operational BI reporting on transactional data
On the right you see the SAP BusinessObjects BI which work with Sybase IQ, all from a single vendor.
Figure 15, Source: SAP
Figure 15 shows BusinessObjects and Sybase IQ integration and key integration points to ensure that IQ works best with BusinessObjects
Data Services can load from multiple data sources, load to IQ, and use BusinessObjects reporting.
Integration surrounds BusinessObjects Data Services and BI platform to optimize integration points
Data Services has the kind of “push downs” to IQ faster and better
BI on the front end: what can be done from a semantic layer with Sybase IQ.
Figure 16, Source: SAP
Key value proposition with Data Services is ways to move data and improve the quality of data, ensure that data is in place. Text Data Processing is extraction based on unstructured data.
Figure 17, Source: SAP
Figure 17 shows that SAP wants Data Services to be the best ETL for Sybase IQ, so IQ can be loaded fast
It calls all the high performance API’s to load data at a faster rate, using Data Quality algorithms from Sybase IQ and integrating Data Quality Libraries in Sybase IQ inside the Sybase IQ kernel.
Figure 18, Source: SAP
Reporting with BI platform provides for visualization; Sybase IQ ODBC JDBC drivers to act “seamlessly”
Figure 19, Source: SAP
BI platform has access directly to Sybase IQ. Discovery and analysis is fast, and response time is fast.
Modules have direct access to IQ with no intermediate layers.
Semantic Layer/Universe Layer ensures the various questions and exploration queries – IQ must understand how Semantic Layer works .
Schemas – semantic layer to understand IQ as a special repository – schemas and security, access controls
Figure 20, Source: SAP
Figure 20 shows they will certify and support different versions and “optimize the BI products”
Question & Answer
Q: In SAP implementation (which has ECC and BW), how would you compare ‘SAP BW on HANA’ with ‘Sybase IQ’. Earlier without HANA in picture, moving data from BW to SybaseIQ made sense but now it’s little confusing. Please clarify.
A: We have a use case where you can use HANA in conjunction with IQ (for aging data). BW along with IQ offers a smart storage solution that builds on top of the BW’s Near Line Storage option. Large volumes of data in BW can now be tiered into IQ. IQ can simultaneously provide analytics directly on the data loaded into it. The use cases are (1) BW with IQ via the NLS and (2) HANA as high value operational data mart along with IQ for aged data analytics.
Q: How expensive they are in comparison to other databases like DB2, Oracle, SQL server etc and above all HANA a SAP product.
A: IQ is cheaper than any of the others mentioned and is used only for OLAP…its not a transaction DB
Q: What are the performance stats between Sybase IQ and HANA?
A: HANA is much faster (refer to the metrics published under sap.com/hana) than IQ plus HANA provides real time analytics.
Q: How does BI connect to IQ: relational model or MDX model?
Q: From an ETL and modeling perspective, BW is time-intensive. How does Sybase IQ/Data Services compare?
A: BW – benefits of complex queries in OLAP. Sybase IQ is more relational. Data loading is faster, doing the transformations in Data Services. Cannot say which is better.
Our thanks to Balaji and Yuva for supporting ASUG with this webcast. I hope we learn more at ASUG Annual Conference in May.