Take one for the team: on the time efficiency of application-level buffer-aided relaying in edge cloud communication

Abstract

Abstract Background Adding buffers to networks is part of the fundamental advance in data communication. Since edge cloud computing is based on the heterogeneous collaboration network model in a federated environment, it is natural to consider buffer-aided data communication for edge cloud applications. However, the existing studies generally pursue the beneficial features of buffering at a cost of time, not to mention that many investigations are focused on lower-layer data packets rather than application-level communication transactions. Aims Driven by our argument against the claim that buffers “can introduce additional delay to the communication between the source and destination”, this research aims to investigate whether or not (and if yes, to what extent) the application-level buffering mechanism can improve the time efficiency in edge-cloud data transmissions. Method To collect empirical evidence for the theoretical discussion, we built up a testbed to simulate a remote health monitoring system, and conducted both experimental and modeling investigations into the first-in-first-served (FIFS) and buffer-aided data transmissions at a relay node in the system. Results An empirical inequality system is established for revealing the time efficiency of buffer-aided edge cloud communication. For example, given the reference of transmitting the 11th data entity in the FIFS manner, the inequality system suggests buffering up to 50 data entities into one transmission transaction on our testbed. Conclusions Despite the trade-off benefits (e.g., energy efficiency and fault tolerance) of buffering data, our investigation argues that the buffering mechanism can also speed up data transmission under certain circumstances, and thus it would be worth taking data buffering into account when designing and developing edge cloud applications even in the time-critical context.

Keywords

• Atomic data entity
• Buffering
• Communication performance
• Edge cloud computing
• Time efficiency