Evaluating DBMS
There are various database management systems available in the market. Each type has its features and can be used for varied purposes. The large number of DBMS makes it difficult to choose the DBMS that should be implemented to solve our problem. To choose the most suitable DBMS, we need to evaluate from various systems. We perform a structured approach to evaluate the database systems.
Evaluation Methodology
This step involves two analyses; both these analyses should be performed to determine the most suited DBMS.
These two analyses are:
- Feature analysis: In this phase, we determine whether the DBMS provides all the features required for the operations that are to be performed on the data and shortlist the DBMS.
- Performance Analysis: In this phase, we analyse only those shortlisted DBMS, evaluate the systems’ efficiency, and choose one with maximum efficiency.
Feature Analysis
This analysis serves two purposes:
- It directly eliminates the DBMS that is not competent to fulfil all the needs of the user.
- We can rank the remaining systems based on the features provided by them.
This is a subjective approach as the database implementation is not necessary and can only be done with the help of reading the documentation of the database management system. This is relatively cheaper as compared to any other approach. This method does not measure the actual performance of the management system. Therefore, it cannot be decisively determined the best database management system.
Performance Analysis
The purpose of this analysis is:
- To analyse the behaviour of the database management system in the user system.
- To gather the data on the performance of the DBMS and rank them accordingly.
Using this method, we can decide the best DBMS that should be implemented to give the maximum result.
Techniques to Analyse the Performance of DBMS
There are various methods to judge the performance of the database management system. Some of the most used methods are Analytical and Stimulation modelling and Benchmarking.
Now, discuss these methods individually.
Analytical Modelling Technique
In this model, we use certain equations that are used to map the performance of the database management system to system parameters. This is quite an accurate method to perform performance analysis. Two models that are used under this technique are Queueing models and Cost models.
Queueing Models
The database system is provided with various resource systems and tasks are provided to the database management system. The system performance is evaluated based on the performance of the system while performing these assigned tasks.
Cost Models
This model effectively calculates the performance of physical databases. They are used to estimate the storage cost. We can determine the average response time to perform a task by the database management system.
Simulating Modelling
In simulation modelling, we construct a real-time model of the system to replicate the behaviour of the actual system. It performs the task and gathers the data, using this gathered data it evaluates the real-time performance of the system. It constitutes an application program, database system, operating system and hardware environment. This is more suitable for evaluating the performance of complex systems. But as the complexity of the system increases the evaluation of the performance of the system becomes more difficult.
Benchmarking
This technique is preferred when there is less number of databases that are to be evaluated. It requires implementing the management system in an almost similar environment so that the data acquired using this method is closest. As the environment for each database management system is created separately, this method is costly and consumes a lot of time in setting up the environment. This method is still performed because this method produces the most accurate result upon which the performance of the management system is evaluated.
In this method, we need to define and implement system configuration, a database and a workload. The key component in this method is the workload that can either be specific to an application or it can be general. All the test is performed using this workload. Various systems operate on these workloads and the result and data from each implementation is used to compare and evaluate the database management system.
Difference Between Analytic and Simulating Modelling and Benchmark
| Based on | Analytic and Simulating Modelling | Benchmarking |
| Usability | It is useful in studying the behaviour of the database | It is more used to compare the functionality of the various system. |
| Preference | These techniques are preferred when dealing with a large number of databases. | This technique is performed when dealing with limited databases. |
| Scope | The scope of testing is limited | The testing is performed in an almost similar environment. |
| Cost | The cost of implementation is less. | The cost of implementation is quite high. |
| Actual Database Evaluation | The evaluation of the DBMS cannot be performed on the actual database. | In this technique, we can evaluate the DBMS on the actual database. |
| Accuracy | These methods are relatively less accurate. | This method is more accurate. |