Is there a SAND/DNA Nearline solution available for SAP BW 3.1/3.5 as well as for SAP NetWeaver BI 7.0?
Yes, there is an ABAP-certified SAND/DNA Nearline solution available for SAP BW 3.1 and 3.5. Certification (for SAP-certified Integration, Powered by NetWeaver status) took place in December 2005.
Features of SAND/DNA for SAP BI:
- Efficient storage of PSA data, ODS Objects and InfoCubes from one or more SAP BW instances, with at least 85% compression
- Virtual Providers enable direct access to stored ODS Objects via BEx or any other certified Business Intelligence tool
- MultiProviders enable access to combined data from SAP BW and SAND/DNA
- Ability to feed stored InfoProviders back into the SAP BW process when they are required to fulfill new analytic requirements
- Ability to use stored PSA data and InfoProviders as DataSources for Transfer and Update Rules that derive new InfoProviders or Data Layers
- Integration of all operations into batch processes via SAP BW process chains
- Can be used in APD processes to derive new InfoProviders or to feed Data Mining processes
- Reduced administration requirements:
- Automatic storage of designated tables
- No need to create indexes
- Any table/column can be queried
- Automatic creation of SAND/DNA structures
- Automatic adaptation to structural changes in SAP BW
- Continued satisfaction of Service Level Agreements (SLAs) with optimized resource usage
- Increased SAP BW system responsiveness for data loads and queries
- Lower Total Cost of Ownership due to reduced resource consumption
- Greater information scope with no reduction in system availability, permitting in-depth Business Performance Management (BPM)
- Rapid access to the data for reporting or for rebuilding KPIs to satisfy continuously changing analytical demands.
“With the SAND/DNA nearline solution, we have found a way to control the explosion of data inside SAP NetWeaver BI. The SAND solution optimizes data warehouse resource usage and allows us to fulfill our Service Level Agreements no matter how much data is involved” - Bardo Gutberlet, Head of SAP System Development / Management, RI-Solution GmbH
SAND Technology’s experience with existing customers for SAP BI shows that the need for a NLS solution is independent of data volume. Customer data volumes range from 0.5 TB to > 6 TB data.
It is the expected data growth and the desire to limit further investments in hardware and administration that typically drives the decision to implement SAND/DNA for SAP BW 3.1/3.5 (even though SAP BW 3.1/3.5 is capable of running multiple terabytes without a nearline component).
Yes, there are some differences. For SAP BW 3.1/3.5, SAND Technology offers a Nearline interface for InfoCubes, ODS Objects and PSA data based on its own SAP-certified ABAP code (that is, SAND has ownership of the Nearline processes).
For SAP NetWeaver BI 7.0, SAP offers their own Nearline Interface as described in SP 7, and this is part of their officially delivered package (SAP has ownership of the nearline processes).
The same version of SAND/DNA is used for all SAP BI versions, including SAP NetWeaver BI 7.0.
Based on the experience of several SAND/DNA for SAP NetWeaver BI (SAP BW 3.1/3.5 and SAP NetWeaver BI 7.0) customers, compression rates between 85% and 95% can realistically be expected, depending on Info Object type, data volume and data diversity.
RI-Solution selected SAND/DNA over the other candidate solutions because of its ability to compress selected data effectively and still keep it accessible for analysis or for building new ODS Objects and InfoCubes.
“…The original size of the data that we’ve moved to nearline storage so far is around 650 GB, and the size of our nearline storage is only about 25 GB - meaning we’ve achieved a compression rate of 96%” - Bardo Gutberlet, Head of SAP System Development / Management, RI-Solution GmbH (Nov 2006)
Based on the experience of several SAND/DNA for SAP NetWeaver BI customers (SAP BW 3.1/3.5 and SAP NetWeaver BI 7.0), SAND/DNA storing data “nearline” is slower by a factor of 1.x overall than SAP NetWeaver BI keeping the same data “online” in a relational database.
The installation of SAND/DNA for SAP BW 3.1/3.5 takes from 2 to 4 hours. Additionally, the SAP DBA must import some transport orders into SAP BW 3.1/3.5. After this, the system is up and running and ready to store data coming from SAP BW 3.1/3.5 via SAND’s Nearline Interface.
How long does it take to get first results?
Directly after the installation, SAND consultants will conduct a 1.5 day workshop. The workshop gives a detailed overview of the NLS Interface of for SAP BW 3.1/3.5 and its interaction with SAND/DNA. In this workshop, selected InfoObjects are stored in SAND/DNA, deleted from for SAP BW 3.1/3.5 and accessed via BEx. As there is no administration (i.e. indexing) required, customers can expect the process to be straightforward. The workshop also covers recommendations for “best practices”.
No, it is not possible to store masterdata in SAND/DNA for SAP BI 3.1/3.5.
“Transactional” InfoCubes enable multiple parallel write access to Cubes, which is mainly used in BPS.
The SAND/DNA solution works in both modes of the cube, i.e. “can be loaded” and “can be planned”.
“Transactional” ODS Objects are not accessible for analysis, because they are not InfoProviders and cannot be accessed by BEx. For this reason, SAND/DNA for SAP BI 3.1/3.5 does not support nearline storage of “transactional” ODS objects.
Because “transactional” InfoCubes or ODS’s contain frequently updated and accessed data, SAND does not recommend storing “transactional” InfoCubes or ODS’s nearline.
Backup and recovery of compressed files is implemented using standard software and file backup/restore methodologies. The files do not have to be uncompressed in order to be processed. Backup of the repository is implemented using a SAND utility.
SAND recommends that SAP BW 3.1/3.5 customers start with SAND/DNA for SAP BW 3.1/3.5 when the date for “going live” with SAP NetWeaver BI 7.0 is further than 3-6 months away, and when there are some “pain points” that need to be addressed immediately.
No, in fact the contrary is the case. In situations where migration from SAP BW 3.1/3.5 to SAP Netweaver BI 7.0 has to be considered at the same time as the introduction of an ILM concept using nearline storage, the order of the various steps must be determined. It is advisable to take advantage of nearline storage in SAP BW 3.1/3.5 to address at least the most important pain points, such as the largest ODS objects. Especially in the case of objects in the data provisioning layer or data acquisition layer, it is easy to reap benefits quickly by introducing nearline first, while leaving smaller objects as they are and applying nearline concepts to them only after the migration to SAP Netweaver BI 7.0. This ensures early benefits from nearline storage while reducing the amount of effort required to migrate metadata for nearline data storage objects as part of the SAP Netweaver BI 7.0 migration project. In this way SAP BW 3.1/3.5 customers are able to:
- Decrease TCO sooner rather than later
- Relieve the pressure of SLA requirements
- Familiarize themselves with the new SAP ILM approach and organize the migration to SAP NetWeaver BI 7.0 accordingly from the outset
- Start with a “skinny” relational database in SAP NetWeaver BI 7.0, and thus benefit from the SAP ILM architecture from day one.
Does SAND provide tools to migrate SAND/DNA for SAP BW 3.1/3.5 to SAND/DNA for SAP NetWeaver BI 7.0?
Yes, because SAND has several customers for SAND/DNA for SAP BW 3.1/3.5, SAND will provide dedicated transformation tools.
To be specific, there are minor metadata differences between the version of SAND/DNA for SAP BW 3.1/3.5 and SAND/DNA for SAP NetWeaver BI 7.0.
Therefore the following migration steps are required:
- Define the new DAPs (Data Archiving Processes)
- Migrate SAP BW metadata (covered by a tool from SAND)
- Migrate SAND metadata (covered by a tool from SAND)
SAND uses the same version of SAND/DNA for both SAP BW 3.1/3.5 and SAP NetWeaver BI 7.0, so all nearline data from SAP BW 3.1/3.5 is still valid in SAP NetWeaver BI 7.0. No reloading is required.
SAP Customers can submit their SAND Nearline-related OSS messages directly to XX-PART-SAND in SAP’S OSS system, and SAND will respond.
SAND/DNA can be deployed on different hardware architectures or configurations to meet customer needs in terms of performance, scalability, SLAs and redundancy. To enable high scalability, parallelism is achieved using multiple instances of the SAND/DNA executor components. The number of concurrent instances running on the hardware configuration controls the level of parallelism. SAND/DNA can take advantage of any number of processors running in a single SMP server or on a distributed architecture.
A typical SAND/DNA configuration will be based on a small Blade Server equipped with a minimum of 2 processors (64-bit). The memory recommendation is 2 GB per processor, which means that a minimal configuration will be 2 processors with 4 GB of RAM.
Finally, a local disk is required to process and store temporary results. The recommendation is to use a minimum size of 73 GB with a speed of 15,000 RPM.
SAND/DNA can be installed on the same hardware as SAP BI, but sharing is not recommended if the SAP BI hardware is being used at more than 50%. From a capacity planning point of view, it is recommended to install SAND/DNA on its own hardware.
SAND/DNA is available for servers running:
- OSF1 HP Compaq Alpha server 5.1A & up,
- AIX version 5.3 & up
- HP-UX-11.00 & HP-UX-11.11 with clibs from 11.00
- Solaris 8 & 9
- Windows NT 2000 SP4 & up
- Linux_X86 (RedHat 4 & SuSe 9 enterprise)
- Itanium 2 & Solaris 10 (x86) available upon request
