| |
Home
About FileTek
Services/Solutions
Products
Press
Jobs
Support
|
|
Case Study: Near-Line
Storage System Makes Aged BW Data More Easily Accessible - page
4
NLS
Testing
Once the installation phase was completed, the pilot group ran the tests
shown in Table 2 to determine whether this system provided the
hoped-for improved performance and new functionality, and whether it addressed
Entergy's data and storage capacity concerns. These tests serve as a good
guide for an NLS-BW integrator to assess any similar implementation.
Table 2
|
Purpose of Test 1
|
|
|
To test all the features and functions delivered
in the new SAP StorHouse API
|
|
|
Method
|
|
|
- Create table
- Create indexes
- Migrate data
- Rebuild Oracle ODS from StorHouse OD
|
|
|
Objects Used
|
|
|
- /ZSTHCOMTROL(API)
- BEx query
|
|
Purpose of Test 2
|
|
|
To demonstrate that partitioned data residing in
Oracle and StorHouse can both be accessed by BW query objects
|
|
|
Method
|
|
|
- One month of data remained in Oracle
- Two older months of data were migrated to StorHous
|
|
|
Objects Used
|
|
|
BEx query run against MultiCube accessing, InfoCube
(Oracle), and remote-cube (StorHouse)
|
|
Purpose of Test 3
|
|
|
To demonstrate that BW access rules are applied
to BW data objects resident in StorHouse
|
|
|
Method
|
|
|
The same BW query object was used to access:
- Data resident in Oracle and
- Data resident in StorHouse
|
|
|
Objects Used
|
|
|
BEx query
|
|
Purpose of Test 4
|
|
|
To demonstrate the time delay associated with accessing
data stored on tape and managed by StorHouse
|
|
|
Method
|
|
|
Two copies of three months of data were migrated
to two different StorHouse managed ODSs
|
|
|
Objects Used
|
|
|
Three InfoSet queries executed against these ODSs
|
|
|
Purpose of Test 5
|
|
|
To demonstrate how the StorHouse system can be used
to manage non-BW data and provide users with SQL level access to
this data
|
|
|
Method
|
|
|
Three months of legacy data were loaded into a
StorHouse relational structure
|
|
|
Objects Used
|
|
|
MS Access was used as the SQL tool and was ODBC-connected
to StorHouse
|
 |
To provide timed benchmarks
for the Entergy pilot would be misleading, because performance is specific
to a given implementation and query complexity. FileTek has seen some
organizations access their data within 10 seconds. For others, access
can take up to 40 seconds. However, retrieving the same information using
a traditional archiving system might take hours or even days depending
on the available personnel (most likely DBAs).
Lessons
Learned
For any organization
with anticipated or current challenges dealing with large amounts of aged
data, Entergy's experience with an NLS system offers guidance. First,
the pilot system experienced only minimal setbacks. (We learned, for example,
that BW 2.1C does not support the selective delete function. This problem
has been corrected in BW 3.0A.) One unexpected benefit that came from
this pilot was the ability to transfer mainframe legacy data (in addition
to BW data) to the StorHouse repository. This showed the flexibility of
an NLS system to adapt to different IT requirements.
The key lesson learned
was the importance of planning an aged-data archiving architecture before
data deluge presents itself. A good strategy and careful implementation
are key factors to a successful execution. Before you deploy an NLS system,
consider implemented and planned ODs and InfoCube designs, the acceptable
response time for accessing aged data, the best time to migrate data to
the NLS system, and the appropriate access methods to the aged data.
Back
|
|
More Information
About the author
To access BWExpertOnline, click:
|
