SAP recommends to implement Nearline Storage as the data management strategy for cold data sitting in SAP BW systems. The offloading of cold SAP BW data into NLS is usually done to reduce the footprint of SAP database before or after migration to SAP HANA. But, the standard solution from SAP using SAP NLS interface with SAP IQ has limitations that prevents SAP users from archiving more data into NLS – this is due to BW stragglers that cause DTP failures when accessing the archived NLS area.
Here are the 4 biggest challenges that SAP customers encounter when offloading data with SAP NLS into SAP IQ:
CHALLENGE #1: UNABLE TO ARCHIVE WARM DATA
When archiving with standard NLS interface all loads that access the locked archived area would fail. To process the DTP the 1.) reload of data is necessary, 2.) re-run the DTP and 3.) archive data again. The data in DTP’s that are accessing locked archived area are called BW Stragglers. Look below how NLS WRITER Add-on enables DTP accessing the archived area to process seamlessly.
CHALLENGE #2: SLOW PERFORMANCE ON NLS DATA
The idea to offload data into NLS always brings the question how fast users can access the NLS data. At first please refer to the blog post from Goetz Lessmann “Benchmarking the NLS solutions” which answers the question which storage for NLS compromise the speed and cost of the storage the most. One of the most often reasons of bad NLS response is reporting on Navigational attributes where the basis attribute has more than 100 records. The symptom is usually the significant decrease of query performance or memory overflow dumps. This NLS bottleneck was already solved by NLS add-on from Datavard that allows to replicate Master data into the NLS database where the join over the large navigation attributes and NLS data is executed before data is transferred back to SAP BW.
CHALLENGE #3: IMPLEMENTATION EFFORT
By doing many Nearline Storage implementation I have learnt about the biggest effort drivers. These are especially identification of HOT, WARM and COLD data to find the best candidates for NLS archiving and explaining business that the data is no longer accessed. Secondly there are several technical activities to be done, such as identification and translation or lookups, creating DAP’s, archiving requests and scheduling of initial and ongoing archiving into NLS.
CHALLENGE #4: UNABLE TO EXTEND SAP IQ
The NLS storage should according to the value of offloaded data be balancing the value of data and the cost of storage. To extend the current SAP IQ database might be a problem due to the cost of the storage and hardware. Therefore, the newly supported NLS storage types such as Hadoop play an important role for offloading SAP data per its value and costs. Please read another blog of mine “NLS on Hadoop” to understand the benefits of Hadoop as the most recommended SAP NLS storage option.
SOLUTION FOR BW STRAGGLERS: NLS WRITER ADD-ON TO ENABLE NLS FOR WRITING & UPDATING
The nearline storage bottleneck can be identified by an DTP error message in RSA1: Error: “Data Package X: Errors in processing” -> “Data record X locked by archiving request XYZ”.
The Nearline Storage in read only mode means:
- Loads accessing the archived area are failing
- Unable to archive more with NLS
- SAP IQ is not good for warm data
The problem can be solved by installing an Add-on with NLS WRITER from Datavard. Loads accessing the archived area will seamlessly change data in NLS (insert, update and delete mode is enabled). No reload of data is necessary to process a DTP accessing Nearline Storage.
NLS writer results in further benefits such as:
- Load accessing the archived area will update data in NLS automatically
- Enables to archive more data into NLS (WARM data)
- Further options to store data in Hadoop which is easily extended
NLS writer is part of the OutBoard DataTiering add-on.
WEBINAR: What is going on in your SAP landscape
Date: Monday 3rd April, 2pm SGT | 8 am CET
Speaker: Tim Oehme, Solution Architect at Datavard