rainer.schlachter

February 2011 Previous month Next month

SAP BW 730: What’s New in the SAP BW Near-Line Storage Solution

 

 

Introduction

 

SAP NetWeaver BW customers have to handle increasingly large data volumes. This typically leads to a steady increase in hardware requirements and maintenance costs. Keeping the system running efficiently and at low operation costs is a permanent challenge.

Therefore, SAP NetWeaver BW offers a near-line storage (NLS) interface that enables customers to separate frequently accessed data from older, less frequently accessed data.

With this solution, the data volume of the frequently accessed online database can be kept relatively low. Less frequently accessed data is archived in time partitions on near-line storage solutions by certified third party partners. Data in near-line storage resides in a highly compressed state in cost-efficient storage with fewer backups to reduce operational costs.

To ensure consistency between the online and near-line partition, the archived time partitions are write-protected and read only. Data in NLS can be accessed directly using BW queries without reloading. Both near-line storage and online storage consistently reflect the BI persistency of an InfoProvider. Below you can see which InfoProviders and features are supported in the different releases.

 

Evolution of the SAP NetWeaver BW Near-Line Storage Solution:

 

Release 7.0

 

  • Archiving of Basic InfoCubes and Standard Datastore Objects
  • NLS Look-Up Support during ETL Processes ( Available as of SAP NetWeaver BW 7.0 Support Package 10, Performance optimized temporal Look-Up available as of SAP NetWeaver BW 7.0 Support Package 21)

 

Release 7.01               

 

  • Request based Archiving for Write-Optimized DataStore Objects

 

Release 7.30

 

  • Queries on top of MultiProviders transparently read the NLS Partitions of the consisting InfoProvider. You can enable NLS use on the query selection screen, in BexQuery Designer,  on MultiProvider level or on PartProvider level
  • Archiving of uncompressed data from InfoCubes. It is no longer a precondition to compress data before archiving.
  • Archiving of semantically partitioned objects (SPO) based on standard InfoCubes and DataStore objects
  • Archiving of semantically partitioned objects (SPO) based on write-optimized DataStore Objects will be delivered with SP3

 

Miscellaneous enhancements

 

  • Automatic rebuild of BWA index during deletion is possible
  • Query selections on navigational attributes are passed to the near-line storage (performance optimization)
  • NLS data access for InfoProviders as sources within APD
  • Data transfer processes (DTP) on InfoCubes extracts data transparently from the online and near-line partitions
  • Separate transformation rule type ‘DataStore Object Lookup’ implemented to read data transparently from online and near-line partitions

 

Queries on top of MultiProviders

 

With SAP NetWeaver BW 730, you can now set up queries on MultiProvider level to read the NLS partitions of the PartProviders.

Configuration for NLS use is possible on four different levels:       

    

  • In the query selection screen 
  • In the properties of the BEx Query Designer
  • In the properties of the MultiProvider
  • In the properties of the PartProvider

 

The prioritization is top-down,  which means that if NLS use is enabled/disabled on query level, the NLS settings from the underlying objects (MultiProvider and PartProvider) are ignored.

For example, the NLS partitions of the PartProvider are read when NLS use is enabled on query level, even though NLS use is disabled on MultiProvider level or PartProvider level.

The settings for NLS use on the four different levels are described below.

In the BEx Query Designer, you can also define an ‘NLS variable’ to let a specific query run either in ‘Read Near-Line Storage’ mode or ‘Do not Read Near-Line Storage’ mode.

 

Definition of a near-line storage variable for the query selection screen:

 

 

 

 

 

 

 

Properties of BEx Query Designer:

 

 

 

 

Properties of the MultiProvider:

 

 

 

 

Properties of the PartProvider:

 

  

 

NLS use can also be delegated from the query level to the PartProvider level by using the following settings:

 

  •  BEx Query Designer: ‘Use Near-Line Storage According to Provider Settings’
  •  MultiProvider: ‘Near-Line Access Set the Same as for PartProvider’
  •  PartProvider: ‘Near-Line Access Switched On/Off

 

This configuration allows you to switch specific NLS PartProvider partitions on and off. In the Statistics Data for the Query Runtime (transaction RSRT), you can see how many PartProviders, including aggregates and near-line partitions (suffix $N), effectively contribute to the query result.

 

 

 

 

Archiving of non-compressed data for InfoCubes

 

The option to archive non-compressed data was requested by customers for whom data compression was a time consuming operation and did not really lead to significant compression rates (highly granular data).

Data compression was a precondition for archiving InfoCube data since release 700. Before data can be compressed, it must be:

 

  1. distributed to subsequent data marts
  2. rolled up to existing aggregates

 

As of SAP NetWeaver BW 730, it is no longer necessary to compress data before archiving. To ensure data consistency in different data layers, the two preconditions above must be met before data can be archived.

The flag ’Allow Archiving of Non-Compressed Data’ is a central setting in the data archiving process.

The flag can be changed in the DAP even if data has already been archived. If you change this flag (Allow Archiving of Non-Compressed Data) in the DAP definition, all archiving requests that are not completely finished (Status ‘70 Deletion Phase Confirmed and Request Completed’) must be finished in the old ‘Compression Mode’ or must be set to invalid before you  can archive data in the new ‘Compression Mode’.

 

 

Archiving of uncompressed data results in partially  or completely  archived requests. This is shown in the Requests tab in the column ‘Selectively Deleted Records/Archived Requests’.

 

 

Note that partially or completely archived requests cannot be deleted and subsequently reloaded with delta DTP requests. The consequences of deleting these requests are described in Note 1360751. This Note is referenced in the popup that appears when you delete a partially or completely archived request.

 

 

 

 

Archiving of Semantic Partitioned Objects (SPO)

 

If you create a data archiving process for a semantically partitioned object (SPO), the data archiving process  (DAP) will apply to all PartProviders.
You can only create a data archiving process for semantically partitioned standard InfoCubes and standard DataStore objects. This function is not available for write-optimized DataStore objects in 730 SP01. Archiving of semantically partitioned objects based on write-optimized DataStore objects will be possible in release 730 SP3.

 

 

The partitioning criteria of the SPO are listed under Additional Partitioning Characteristics and cannot be changed. These criteria can be used to semantically preselect specific partitions when scheduling an archiving process.

 

 

 

 

When you create an archiving process for the SPO, you can decide whether you want to archive across all partitions or only archive a specific partition.
If you choose a specific partition, you can select a physical partition from a drop-down menu.

 

 

If you choose Cross-Partition, you can restrict the partitions with the partition criteria of the SPO under Further Restrictions.

 

 

 

Note that if the SPO partition criteria is time-dependent (as shown in the example above with the characteristic ‘calendar year’), whole partitions can be archived, which means these partitions are empty. You can either delete or repartition these partitions. If you decide to delete or repartition the empty partitions, you cannot reload the archived data anymore because the relevant partitions do not exist anymore. In this case you have to repartition or recreate the old partitions before reloading.
However, if the SPO partition criteria are time-independent (for example, partitioned by sales organization), archiving time slices does not result in empty partitions.

As of SAP NetWeaver BW 730, archiving requests for semantically partitioned objects cannot be scheduled with transaction SARA. This functionality is only available with TA RSA1 or using process chains (transaction RSPC). Only reload requests have to be scheduled with transaction SARA if an ADK archive is involved in the DAP configuration.

 

 

Automatic rebuild of BWA index during deletion is possible

 

In SAP NetWeaver BW 730, the BWA, online and near-line partitions stay consistent during the archiving process. After data has been copied to the archive, the corresponding BWA index is invalidated and can be rebuilt during the deletion phase of the archiving request. 

 

 

If you use process chains for data archiving, we recommended that you use the process step ‘Initial Activation and Filling of BWA Indexes’ to rebuild the BWA index.  If you have set the ‘Rebuild BWA Index’ flag in the Data Archiving Process Deletion step, the following question appears:

 

 

 

 

 

Query selections on navigational attributes

 

Navigational attributes are master data attributes that can be used to dynamically filter the result of queries. As master data attributes are located outside in the star schema of the InfoCube persistence, they are not part of the archived NLS data.
Therefore, query restrictions for navigational attributes are not transferred to NLS. The attribute values are assigned subsequently during OLAP (application server) and filtered in the result set. This will impact on performance if the filter condition for navigational attributes is highly restrictive.
As of SAP NetWeaver BW 7.30, selections for navigational attributes are converted to a characteristic selection for the attributes with characteristics. The converted selection is then passed to the near-line storage instead of the navigational attribute selection.

 

 

NLS data access for InfoProviders as sources within APD

 

As of SAP NetWeaver BW 7.30, the Analysis Process Designer is enabled to read data transparently from the online partitions and near-line partitions.  This makes it easy to include historically archived data in your analysis process.

 

 

 

DTP on InfoCubes can extract data transparently from online and near-line partitions

 

 

 

 

Separate transformation rule type ’ DataStore Object Lookup’

 

SAP NetWeaver BW 7.30 introduces a separate transformation rule type: ‘DataStore Object lookup’. If the ‘Looked-Up DataStore Object’ contains near-line partitions, data will be read consistently, regardless of whether data is contained in the online partition, the near-line partition or both.

 

 

 

 

 

Links

 

Documentation

RKT Session

 

Partner links:

 

PBS Software GmbH     or SAP EcoHubPBS

SAND Technology Inc   or SAPEcoHubSAND

IBMHowtoGuide           or SAPEcoHubIBMDB2

 

 

Disclaimer

 

This document is not subject to your license agreement or any other service or subscription agreement with SAP. SAP has no obligation to pursue any course of business outlined in this document or any related presentation, or to develop or release any functionality mentioned therein. This document, or any related presentation about SAP's strategy and possible future developments, directions, and functionality of products and/or platforms, are all subject to change and may be changed by SAP at any time for any reason without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or functionality. This document is provided without a warranty of any kind, either express or implied, including but not limited to, the implied warranties of merchantability, fitness for a particular purpose, or non-infringement. This document is for informational purposes and may not be incorporated into a contract. SAP assumes no responsibility for errors or omissions in this document, except if such damages were caused by SAP intentionally or grossly negligent.

All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ materially from expectations. Readers are cautioned not to place undue reliance on these forward-looking statements, which speak only as of their dates, and they should not be relied upon in making purchasing decisions.

 

 

 

 

 
 

               

 

Actions

Filter Blog

By date: