Virtual Connection: Selective Connection System for Energy-Efficient Wearable Consumer Electronics

Gyeonghwan Hong, Dongkun Shin

Abstract: Wearable consumer electronics such as smartwatch and smartglasses generally use peer-to-peer (P2P) communications to connect to their companion devices such as smartphones. The recent wearable consumer electronics are equipped with multiple heterogeneous network interfaces, which have different power consumption and transfer bandwidths. To optimize energy efficiency, a network interface must be selected for communication depending on workload or system state. Such a technique is referred to as selective connection. However, the current selective connection systems cannot support the P2P-based wearable devices. We propose a protocol-independent user-level system, called virtual connection, that decouples applications from network-specific operations for selective connections. The proposed virtual connection system consists of user-level API and user-level middleware components. To establish the usefulness of our virtual connection, we present a case study of the selective P2P connection for wearable consumer electronics equipped with Bluetooth and Wi-Fi Direct. We demonstrate that, compared to existing selective connection systems, the P2P-based selective connection system implemented with the virtual connection can reduce the energy consumption of the wearable consumer electronics by 1.26-2.31 times.