Abstract:
Integrating the smart phone in our daily life where many info's and tools can be
accessed, is a new horizon for achieving a better quality of life. Accessing and
analyzing data at the right time with the right price is the core of human need. We are
targeting a professional environment where some complex applications are being
developed and used on mobile devices, such as Health Care or Educational
Environments. Mobile devices, however, remain short in resources including energy,
computation, and storage. Edge computing emerged as an effective solution to enable
mobile devices processing complex computations through offloading to
geographically close computing servers. While offloading provides access to powerful
servers, frequent transmission of tasks over wireless links leads to draining the battery
of mobile devices. In this work, we propose two contributions to help in the adaptation
of mobile offloading. First, we elaborate a design methodology related to the
development of applications in any professional environment targeting energy savings
whereby we introduce a concept of adding annotations dynamically in the headers of
webpages. Second, we provide a mobile offloading model that makes use of the
multiple wireless interfaces of mobile devices to transfer computation tasks to edge
servers only when energy savings are expected and delay requirements can be met. We
consider two widely available wireless interfaces namely, Wi-Fi and ZigBee, and
formulate our problem as integer linear programming to decide whether to offload on
any of the existing interfaces or compute locally while minimizing the total energy consumption and meeting time limits. The work presents various results under
different system parameters and demonstrate considerable gains in energy
consumption.
