Cardiac arrhythmia detection and classification is of outmost importance for early diagnosis to reduce significantly the rates of morbidity and mortality of patients with heart disease. In particular for patients with silent cardiac symptomatology, the advances in wearable sensing technology offer a promising solution for on-line, real-time detection of intermittent tachyarrhythmia events that otherwise may evolve undetected. In this paper the authors examine some of the key issues that outline the path towards integrating various aspects for ECG signal acquisition and analysis with current trends in wearable sensing technology and present recent results towards on-line real-time arrhythmia classification considering the IoMT (Internet of Medical Things) approach.