QoS-Aware Packet Chunking Schemes for M2M Cloud Services

With the emergence of machine-to-machine (M2M) communications, huge numbers of devices have been connected and massive amounts of traffic are exchanged. However, since M2M applications typically generate packets of very small size, the number of incoming packets is becoming larger than the maximum number of packets that a router can handle per second, and network throughput is decreased. That is, network resources cannot be used efficiently. Therefore, the M2M cloud network operator makes chunks of these small packets, but a waiting time for chunking is inevitable, and thus application requirements may not be satisfied. This paper proposes new packet chunking schemes to both meet the QoS application requirement and improve the router´s achievable throughput. The proposed schemes need multiple buffers and then classify the arrival packets to one buffer based on their acceptable waiting time. After that, the chunk packet is made based on an attribute of each packet to meet the application requirements. Through simulation experiments, we showed that the proposed scheme attains excellent performance even in cases of overloaded and imbalanced traffic.