You are designing a partitioning strategy for a SQL Server Analysis Services (SSAS) cube.
New data is loaded in real-time into the data warehouse that feeds the cube. Between 10
million and 15 million rows of data are loaded into the main fact table each day from a
Microsoft Azure SQL Database.
You have the following requirements:
Maximize cube query performance during business hours.
Ensure that data is available in the cube as soon as possible after it is loaded into the data
warehouse.
You need to design a partitioning strategy that meets the requirements.
What should you do? (More than one answer choice may achieve the goal. Select the BEST
answer.)
A.
Partition the cube by day for history, using hybrid OLAP (HOLAP) storage mode. Create a
daily partition that uses multidimensional OLAP (MOLAP) storage mode during the day.
Reprocess the partition incrementally during the day.
B.
Partition the cube by week for history. Create a daily partition that uses multidimensional
OLAP (MOLAP) storage mode. Process the partition periodically to add new data.
C.
Partition the cube by day for history, using multidimensional OLAP (MOLAP) storage
mode. Create a daily partition that uses proactive caching during the day. Reprocess the
partition in full MOLAP storage mode at night
D.
Partition the cube by day for history, using multidimensional OLAP (MOLAP) storage
mode. Create a daily partition that uses relational OLAP (ROLAP) storage mode during the
day. Reprocess the partition in full MOLAP storage mode at night.
I think it should be C.
There is no min storage requirement. Only performance for the cube and “data is available in the cube as soon as possible” for the day… I think it’s D.
See 5.4.2 from this paper:
https://msdn.microsoft.com/en-us/library/dn749781.aspx
I tend to C.In Rolap you get the data as soon as posible but the performance can suffer.
If the daily load was 1M (not 15M), D is the best option but it
depents on a DWH hardware configaration also.
With proactive caching you can get access to new data a few minutes later
but the performance is much better.
my first guess also C
I think C, can use low latency MOLAP