Companies started conducting large-scale analysis with Internet of Things (IoT) data on cloud computing platforms to increase operating efficiency and revenue. IoT platforms based on fog computing and edge computing frameworks are expected to be used for data analysis and to communicate with other Internet services. In this context, a major concern is application virus and IoT platform attacks, thus the application services must be robust enough against those threats.

A next-generation framework for building application services running on future cloud platforms for IoT is the MirageOS Unikernel programming framework. This platform focuses on higher robustness in the whole software stack, including the operating system and application layers and in the server virtualization hypervisor. However, the network throughput of MirageOS on ukvm is not high since currently it is not optimized.

To solve this particular issue, the authors propose a scheme to accelerate the network performance of MirageOS on ukvm. This proposal enables both higher robustness and network performance for IoT service applications on cloud computing platforms, while not requiring additional scheme modifications in existing MirageOS-based applications.

The proposed approach is interesting. The authors characterize the problem; describe and discuss MirageOS and ukvm, their approach and solution, and the evaluation method and evaluation results; and describe and compare a related network approach.

In conclusion, this is a very useful paper. The authors (and I) believe that their paper can contribute to the realization of higher robustness and network performance for IoT cloud service applications, as their proposal’s evaluation results show.