The requirements on software and applications are getting harder. If they are perceived as slow or power hungry, the customers will be looking for other solutions instead. The development of open source frameworks is rapid. Frameworks are being built and updated continuously, with different performance, functionality and complexity. This bachelor’s thesis studies and compares WebSocket frameworks in JavaScript, with focus on performance and scalability. A pre-study was made to find out which parameters are of interest when testing and measuring the performance of web servers. A test bench was then built and plain WebSocket, Socket.IO and SockJS were benchmarked. The study shows that there exist significant differences in performance and trends that indicate that some frameworks are superior to others in high concurrency applications. Plain WebSocket can be up to 3.7 times as fast to receive a message from the server compared to Socket.IO and 1.7 times as fast compared to SockJS. Plain WebSocket scales well in terms of response time and memory requirement with higher concurrency levels. Socket.IO requires a lot of memory which is a potential problem. SockJS does not scale well at all and does not manage the same level of concurrency as the other two.