Intrusion Detection Systems (IDS) are crucial in monitoring and identifying potential threats within network systems. Researchers are continuously exploring the integration of machine learning and deep learning methodologies to extract features from network traffic. Reliable benchmark datasets are essential to test and assess the effectiveness of IDS. However, outdated datasets can lead to performance decline and class imbalance, affecting detection rates and accuracy. This paper emphasizes the importance of adapting IDS methodologies to the latest datasets for enhanced security and privacy in the rapidly evolving technological landscape. A framework is proposed for a new dataset and a comparison of available datasets based on additional requirements is presented to generate an effective and enhanced IDS dataset. The findings from a radar chart visualization reveal varying degrees of feature presence in different datasets, guiding the development of an effective and enhanced IDS dataset. The proposed features aim to rectify information deficiencies in previous datasets and improve the overall efficiency of IDS models, ensuring their relevance in addressing contemporary network security challenges.