Very often, I am asked a basic question: “Do we really need BW on HANA if we have SAP HANA?” Indeed, migration of SAP NetWeaver BW 7.3 to SAP HANA can create a landscape that benefits from the strengths of both products. Professionals leveraging such landscapes can integrate data and provide reports that users can access throughout the organization. Users can also take advantage of HANA’s ability to process data in real time. More than one migration method exists, though, and users need to understand the particular contexts and advantages of each type.

To help readers gain understanding, I explore different ways to realize SAP NetWeaver BW on HANA, some considerations for HANA-specific BW data models, and some vital BW maintenance techniques.

There are three possible ways to implement SAP NetWeaver BW on HANA, depending on your current landscape:

1. Fresh installation
2. Full operating system database in-place migration
3. Post-copy automation (PCA) migration

In a single environment, SAP NetWeaver BW and SAP HANA models can complement each other. I have seen companies use SAP NetWeaver BW on HANA as an enterprise data warehouse to integrate heterogeneous data and provide highly analytical reports with cross-organizational views. SAP HANA models, meanwhile, are often used to process large volumes of data in real time.

Users should consider an implementation of SAP NetWeaver BW on HANA for several reasons. Prebuilt contents in both systems not only enable a foundation for report customization but can also serve as templates and best-practice reference models for in-house development. SAP NetWeaver BW offers standard content that is not currently available in SAP HANA Agile DataMart solutions, such as Inventory Movements cubes, Finance cubes, and Logistics cubes. SAP HANA, however, has rapidly growing amounts of HANA Live content, which can be compared to BW’s standard contents.

SAP NetWeaver BW on HANA offers new features, such as BW Workspaces, which enable greater flexibility and faster development and can reduce overall Total cost of Ownership (TCO).  SAP NetWeaver BW on HANA-specific Layered Scalable Architecture (LSA)++ architecture and SAP HANA can reduce redundant data footprints in SAP NetWeaver BW and enable virtualization of business logic. This is due mainly to SAP HANA’s ability to perform data manipulation during report run time.

SAP NetWeaver BW architecture stores pre-calculated data in multiple layers. With SAP HANA, though, this multi-layer storage is no longer required, shortening the duration of data load storage to HANA. Furthermore, in the event of catastrophic data loss or business logic changes, the enterprise corporate memory and the Open DataStore (ODP) concepts allow reload of data within SAP NetWeaver BW without re-extracting data from the source system. This enables SAP NetWeaver BW on HANA to be self-sufficient. This corporate memory function is in addition to the standard SAP HANA database backup and recovery process.

Users adapting SAP NetWeaver BW on HANA can harness the speed of HANA and still take advantage of their familiarity with SAP NetWeaver BW 7.3.

Three Ways to Migrate SAP HANA BW on HANA

There are three methods for migrating SAP NetWeaver BW 7.3 to Hana: New installation, in-place migration, and PCA migration (Figure 1). I discuss the pros and cons of each method.

Figure 1

Three ways to migrate SAP NetWeaver BW onto HANA

Fresh Installation

A fresh installation is a new install of SAP NetWeaver BW on HANA without existing reports, BW models, and data from the legacy BW system. A fresh installation creates a dual landscape in which the existing legacy SAP NetWeaver BW system is not impacted during the development of SAP NetWeaver BW on HANA. Users are able to continue using the existing the legacy BW application without interruptions until the validation of SAP NetWeaver BW on HANA content is complete.

A fresh installation would be considered when there is no existing SAP NetWeaver BW deployment in place at the organization. Such an installation might also be appropriate when users are planning to build new SAP NetWeaver BW data models or they are planning a reload of data or a redesign of existing data models. SAP NetWeaver BW on HANA offers the opportunity to improve data models with newer BW artifacts and simplified development architectures such as LSA++ for SAP NetWeaver BW on HANA. With these new artifacts and LSA++ concepts available to companies, users may desire to take on a fresh installation rather than migrate their existing system.

With a fresh installation, the business should consider the cost and resource impact of additional hardware, support, and landscape maintenance, because they will have an existing BW system running as users manage the migration. Though a fresh installation of SAP NetWeaver BW on HANA may require additional support and hardware, it does avoid the complexity of a migration of an existing BW system.

Full OS/Database/In-Place Migration

A full operating system and database migration implementation achieves direct migration of the existing landscape, including existing reports, models, and data from SAP NetWeaver BW into SAP NetWeaver BW on HANA. Companies that have a SAP NetWeaver BW system that can afford to be offline for a number of days would have the most to gain from this low-cost approach. This migration option enables faster turnaround compared to a fresh install, and does not require the user to maintain a dual landscape during migration.

This migration approach, though, does require extended downtime of the existing landscape, as users are migrating the existing live system into SAP NetWeaver BW on HANA instead of using copied or parallel systems. In addition, because the existing system is being migrated to SAP NetWeaver BW on HANA, it lacks a reference point for reconciliation of data once the migration is completed. With the fresh install and the PCA approach (explained in the next section), there is a dual landscape involving the existing system and the migrated system. Therefore, in the fresh and PCA methods, the migrated system can be compared against the existing system for data quality.

This full OS/database migration option also carries a high risk as it completely replaces the existing landscape. In case of migration failure, such as faulty hardware or failed data transfer due to a faulty network, reversal of the migration may take significant effort. The reversal effort depends on the number of affected systems, the size of the database, and the number delta records needed to be loaded in once the reversal is completed.

PCA or Parallel System Approach

PCA (e.g., a parallel system approach) is a less risky migration approach than a full OS/database migration. The PCA approach enables shorter downtime as the migration occurs in parallel compared to in-place migration and, in case of failed migration, the existing system does not lose data. PCA enables the system to copy the existing BW landscape into SAP NetWeaver BW on HANA while enabling delta cloning and loading of the two parallel systems (the existing BW and the new SAP NetWeaver BW on HANA). The delta cloning enables the source system to keep track of delta records for different BW systems at the same time. Once the SAP NetWeaver BW on HANA migration is completed, the cloned delta can be repointed into the new system and the delta loads can continue. This cloning approach eliminates the need for reloading of data into SAP NetWeaver BW on HANA.

Because this option enables the migration of the existing BW system without modifying the BW data models, the user should be satisfied that the technical migration of the existing BW landscape into a SAP NetWeaver BW on HANA system is sufficient to obtain adequate HANA performance without optimizing the customized routines or BW models.

This migration of BW may take one to three days depending on the database size, hardware size, any added migration steps such as upgrades, and Unicode migration. However, the timing of delta and delta queue management has to be given full consideration prior to beginning the migration process. Based on my experience, even though full shutdown is not required, migration and delta cloning should be scheduled during a period of minimal business activity.

SAP NetWeaver BW on HANA Modeling

When planning one of these migration scenarios, organizations must be aware of SAP NetWeaver BW on HANA-specific modeling considerations to ensure that the migration and resulting BW and HANA landscape are as efficient as possible.

3.x Extraction, Transformation, and Loading (ETL) objects can be used on SAP BW on HANA. Based on lab tests and user experience, however, a migration to 7.x versions yields better performance and provides greater flexibility with delta management. The 7.x ETL can support more than one delta from a single Persistent Staging Area (PSA). SAP has provided built-in wizard-based programs for easy and mass migration of ETLs to 7.x versions.

If you are migrating a BW system to SAP HANA, you may need to optimize existing transformations in order to increase HANA performance. These transformation codes were written originally for non-HANA use, and the codes may not perform well with the HANA database. Therefore, users should consider optimizing the codes for increased performance. One example would be to optimize SELECT statements (see SAP Notes 1662726 and 1092539: Note 1662726 - Optimizing select with FOR ALL ENTRIES in SAP HANA database and Note 1092539 - Master Data Lookup with prefetch).

For a SAP NetWeaver BW on HANA migration of existing installations—in either a full OS or a PCA approach—users must migrate certain BW InfoProviders, such as Cubes, to achieve optimal performance.

Prior to BW 7.3 SP8, the Data Stores Objects (DSOs) needed to be migrated to gain optimal performance. However, with the updated SP08, there is no need to migrate DSOs to HANA.

InfoCubes require migration to gain optimal performance. The migration of an InfoCube can take up to 30 minutes depending on the size of the cube. Reporting may not be possible during this time.

SAP NetWeaver BW on HANA also offers enhanced BW artifacts such as CompositeProviders and BW Workspaces. A CompositeProvider enables the data modeler to join InfoProviders. The CompositeProvider operations are transferred to the HANA database level to gain optimal performance. BW Workspace is designed to enhance the self-service model. The BW Workspace enables users to use existing BW objects such as BW cubes, combine them with ad-hoc objects such as Excel data, and create ad-hoc reporting.

Maintaining the BW system is crucial to system stability and high performance. Whether users are planning on migrating existing BW systems to HANA or planning fresh installations, they must be prepared to provide continuous system maintenance.

BW Migration Cockpit for BW on HANA

The Migration Cockpit provides a one-stop solution which enables administrators and migration specialists to manage any of the three migration processes from beginning to post-migration. The migration cockpit not only provides the ability to execute migration and maintenance steps in order but also provides quick links to relevant documentation of each of these functionalities.

The migration cockpit consists of multiple tabs, each designed to focus on specific tasks. The Checking tab (Figure 2) automates the monitoring of the best-practice guidelines and pre-requisites for migration of an existing SAP NetWeaver BW deployment to the SAP HANA platform. It is important to use this Checking tab before and after migration activities. Consistency checks and planning-function checks can also be run whenever there is a failure with activation, run reporting, or data loads.  

Figure 2

Checking tab of the SAP NetWeaver BW on HANA migration cockpit

For additional details on each of the migration activities and steps, refer to SAP Note 1729988 –SAP NetWeaver BW powered by SAP HANA - Checklist Tool. Each tool also offers documentation that explains the functions of each tool. You can access this information by clicking the Documentation button. I recommend that the migration team read these documents prior to executing these steps.

The Sizing tab (Figure 3) is crucial to migration projects, as the sizing tool estimates required hardware size for migration and also provides projections for future growth based on these percentage estimates. In addition, the sizing program provides table names, the actual size of each table, and the relevant HANA configuration (scaling out the required nodes). Refer to SAP Note – 1736976 - Sizing Report for BW on HANA and SAP Note 1855041 - Sizing Recommendation for Master Node in BW-on-HANA.

Figure 3

Sizing tab

The Sizing tab also provides the ability to configure and manage near-line storage. The Sybase IQ-based near-line storage features tight integration points to manage data and provides the ability to report data loads seamlessly. The ability to store old data in near-line form is crucial for successful data warehouse design for various reasons, such as cost management, improved performance, and low maintenance. In addition, near-line storage can be configured as corporate memory in the LSA architecture.

Data efficiency is critical to a successful data warehouse. One of the ways to achieve this is to closely monitor system tables to keep them from being flooded with system information. With the Housekeeping tab (Figure 4), these functions can be automated to keep the system healthy. For further details underlying system tables and the methods to keep them clean, refer to SAP Note 1829728 - BW Housekeeping Task List and to the SAP_BW_HOUSEKEEPING program.

Figure 4

Housekeeping tab

The Migration tab (Figure 5) displays various pre- and post- migration activities. This tab provides a one-stop solution for the entire set of migration activities and keeps track of migration activities as they progress.

Figure 5

Migration tab

The Conversion wizard (Figure 6) is applicable to systems in which 3.x ETLs are available. This tool enables automated and orderly conversion steps. The Conversion tab also provides a list of ETLs that can be improved for performance with the reduction of data persistency.  Refer to SAP Note 1908367 – SAP NetWeaver BW Transformation Finder for further details.

Figure 6

Conversion tab

The Optimization tab (Figure 7) provides analysis for fine-tuning existing ABAP codes. This provides a number of optimization analyzer tools for custom ABAP programs including the start, end, and transfer routines. The migration team can run these analysis tools to identify the codes that require updates to perform optimally in HANA. Refer to SAP Note 1847431 – SAP NetWeaver BW ABAP Routine Analyzer for further details.

Figure 7

Optimization tab

Haran Vinayagalingam

Haran Vinayagalingam is a Practice Lead with  SAP’s HANA Services Center of Excellence team. He is a certified HANA architect with experience implementing SAP HANA and SAP BW on HANA for large-scale enterprises. Along with numerous SAP HANA implementation experiences, he is the North American solution-delivery owner for SAP HANA Live.

You may contact the author at haran.vinayagalingam@sap.com.

If you have comments about this article or publication, or would like to submit an article idea, please contact the editor.