This paper introduces a distributed service-oriented system, which is developed to provide ECG (electrocardiogram) monitoring, analysis and storage services. The Service-Oriented Architecture system design is introduced for ECG signal transmission and processing. The implementation of cloud-based web-services and overall system architecture is described. The presented system includes a T-shirt with five electrodes intended for the acquisition of the signal. The ECG data for the experiment were recorded while the participant was moving. The signal replicates real conditions and the ECG data contain different high and low frequency noise. Therefore, this paper includes analysis of data filtering methods, model selection and ECG parameter calculation algorithms. The DWT algorithm was selected for the high frequency noise reduction and the BEADS method was used for trend removal. It was experimentally identified that these algorithms are effective and can be used in the system under development. The tests covering overall system were performed on an Amazon cloud computing infrastructure. The results are presented together with a discussion of various constraints of service-oriented performance.