What Is The Purpose Of Data Warehousing Tuning?
A data warehousing keeps changing and it is unforeseen what question the customer is going to write later on. Therefore it becomes more challenging to track a information factory program. In this section, we will talk about how to track the different factors of a information factory such as efficiency, information fill, concerns, etc.
Difficulties in Data Warehouse Tuning
Tuning a information factory is a challenging process due to following reasons:
Data factory is dynamic; it never continues to be continuous.
It is hard to estimate what question the customer is going to write later on.
Business requirements change eventually.
Customers and their information keep changing.
The customer can change from one group to another.
The information put on the factory also changes eventually.
Note: It is very essential to have an extensive knowledge of information factory.
Here is a list of purpose actions of performance:
Average question reaction time
Check out rates
Time used per day query
Memory utilization per process
I/O throughput rates
Following are the indicates keep in mind.
It is necessary to specify the actions in service level contract (SLA).
It is of no use trying to track reaction time, if they are already better than those required.
You must have genuine objectives while making efficiency evaluation.
It is also essential that users have possible objectives.
To cover the complexness of it from the customer, aggregations and opinions should be used.
It is also possible that the customer can write a question you had not updated for.
Data Load Tuning
Data fill is a critical part of over night handling. Nothing else can run until information fill is finish. This is the access point into it.
Note: If there is a wait in shifting the information, or in appearance of information then the entire product is affected poorly. Therefore it is very essential to track the information fill first.
There are various techniques of adjusting information fill that are mentioned below:
The very common strategy is to place information using the SQL Part. In the process, normal assessments and restrictions need to be conducted. When the information is placed into the desk, the rule will run to check for enough area to place the information. If enough room is not available, then extra area may have to be assigned to these platforms. These assessments take the a chance to perform and are costly to CPU.
The second strategy is to avoid all these assessments and restrictions and place the information straight into the preformated prevents. These prevents are later written to the data source. It is quicker than the first strategy, but it can work only with whole prevents of information. This can lead to some area waste.
The third strategy is that while running the information into the desk that already contains the desk, we can maintain indices.
The 4th strategy says that to fill the information in platforms that already contain information, drop the indices & reproduce them when the information fill is finish. The choice between the third and the 4th strategy relies on how much information is already packed and how many indices need to be renewed. oracle course is always there for you to provide quality based training in Pune.