The successful implementation of multi-agent control strategies (MAS) within smart grid applications is heavily reliant on communication frameworks that enable robust inter-agent messaging. In this regard, a promising communication framework is Data Distribution Service (DDS), which is fully distributed, enables low-latency messaging, and provides configurable Quality of Service (QoS) profiles that customize messaging parameters. As such, this paper proposes a DDS-based, fully distributed communication framework (DCF) for smart grid applications. The proposed DCF is configured with QoS profiles that discard outdated messages when agents experience intermittent connection issues, thereby preventing the misaligned execution of distributed control strategies. The performance of the proposed DCF is validated by developing a Software-In-The-Loop platform, where agents utilize the DCF to execute the Asynchronous Weak Commitment (AWC) control strategy within the application of distributed voltage regulation. Experimental results show that the proposed DCF delivers messages at latencies between 4-36 milliseconds, which is sufficient for smart grid applications. The proposed DCF also enhances the efficacy of the AWC strategy when performing voltage regulation by discarding outdated messages and preventing unnecessary control actions in resolving voltage violations. The proposed DCF is validated at a real-world Canadian microgrid, where overvoltage violations are resolved in a timely manner.