Measuring and Scaling

Engineering has strong convincing and proofing because it is very close to methodologies that can be measured and scaled. By scaling the measured values, you can know where you are at. If you do not use the measured values, you will be caught off guard to the surprises that will take place in your projects.

If you are lucky, have time and your customer has patience, you can try to overcome problems. For this, you need to have a performance model. Your performance model has 3 main components.

• Your goals,
• Your measurements,
• Scales.

After creating your own model, we need to answer the question "What Will We Measuring”. Components such as memory, CPU, errors, times, disk I/O, network I/O are the first to come to mind. Performance improvement is a repetitive process. You must repeat and measure your tests to see the results/benefits of your improvements. This process should take until your results overlap with your performance components or apply for code optimization, design change or an alternative method. Create your key scenarios. Performance scenarios define your priorities and workload of various values. You know where to focus with your well-defined performance components for your key scenarios and when to complete the repetitive process. You should keep your performance improvement discipline; you should give up work that will bring little benefit that are away from the system. Do not hesitate to exclude technical features that will not work clearly with your performance components. Know the costs of your solutions to you and your customer.

•                 Being late is the starting point for negative consequences.
•                 Identify your performance components to measure.

Performance components can be the data such as response time, resource usage, user capacity, data volume, and volume of motion. If the software’s perform differs from your performance component values and your target values by measuring, there is only one thing to say; “Your project has failed”. You don't know how good your performance is without your performance components. You can increase your performance by spending too little or too much effort. You can unnecessarily design very complex designs or easily connect the structure with an extremely simple approach. Or you can refer to interesting security methods that will never be improved and have not been used before. In short, you should avoid using complex design or unnecessary technological approaches. Otherwise, after a point, you may find yourself in a weak state, far from making good engineering decisions.

Performance and scalability are the main features of quality. If you are working with software processes such as ITIL or CMMI, you should balance performance well with other quality features such as security, availability, integrity, manageability, interoperability and sustainability.

As a result, having a software performance model, measuring, scaling and repeating will also bring you to a positive point in your project.