Latency as a key performance indicator (KPI) of 5G communication has attracted a significant amount of efforts from both industry and academia to provide the enhanced technology solution to meet the requirement of low latency from the diverse 5G use cases. With the evolution of technologies and eco-system in 5G, it becomes not only more important but also more challenging to manage a large number of elements which have been connected and orchestrated to provide an end-to-end communication service across different segments of the network. This paper introduces a framework approach to define and evaluate latency in 5G based on the proposed 5G systems from four research projects. The use case and architecture are well studied to identify and analyse the contribution of latency. The proposed reference framework of latency is validated by mapping it to the testbed of these projects.