COMFORT‐APP

BASIC INFORMATION

Short name: COMFORT-APP
Long name: Computation Offloading for IoT-enabled Applications

Company: WINGS ICT Solutions Ltd. | wings-ict-solutions.eu
Country: Greece

Call: F4Fp-02-Stage1 and F4Fp-02-Stage2 (see call details)
Proposal number: F4Fp-02-Stage1-07 | F4Fp-02-Stage2-07

SUMMARY REMARKS & TESTBEDS

STAGE 1

Computation offloading is an open and interesting problem in the era of Mobile Edge Computing (MEC). COMFORT‐APP aims to provide an offloading mechanism for IoT‐enabled applications. IoT devices have limited resources in terms of energy, computation and networking, thus the execution of computation intensive applications is still restrictive for them. MEC offers the processing power of cloud computing at the proximity of mobile users. Computation offloading can alleviate energy consumption of IoT devices, and overcome their computational limitations, while guaranteeing their stringent execution time constraints.The existing offloading methods use static modeling for both mobile devices and edge servers. This cannot guarantee high Quality of Service (QoS) since overloaded servers can deteriorate the overall performance of mobile application. COMFORT‐APP uses dynamic state‐space modelling to describe the dynamic operation of IoT devices and edge servers and designs novel controllers for the edge server that guarantee their stable operation and the satisfaction of the users time requirements under varying workload. Fed4FIRE+ testbeds and software tools are suitable for experimentation on computation offloading. The results will guide us for conducting more complex experiments on load balancing between mobile edge computing environments and concurrently give valuable feedback to Fed4FIRE+ community on Mobile Cloud Computing.

STAGE 2

The role of computational offloading in current and next‐generation networks is becoming increasingly important due to the proliferation of Internet of Things (IoT) and massive Machine Type Communications (mMTC) technologies, which introduce a vast number of low‐capability, low‐energy devices to the networking ecosystem. Such devices, regularly perform computationally intensive and/or energy‐hungry tasks which lead to inefficient operation, but the latest advancements in Mobile Edge Computing (MEC) offer a much more attractive alternative which can provide low‐latency and high energy efficiency operation, maximizing system performance and extending the devices’ life‐time. COMFORT‐APP aims to provide such an intelligent, dynamic MEC offloading scheme for IoT and MTC devices which will allow the devices’ computational limitation to be alleviated and energy consumption to be minimized while guaranteeing a certain level of Quality of Service (QoS). COMFORT‐APP utilises dynamic state‐space modelling to describe the dynamic operation of IoT devices and edge servers and consequently designs novel controllers for the edge servers that guarantee their stable operation and the satisfaction of the users’ time requirements under varying workload. The stage‐1 experimentation provided proof that even under a basic configuration, significant gains can be observed by using a MEC environment to offload computationally intensive tasks from IoT devices. Using the insights from stage‐1, the COMFORT‐APP will be upgraded with new decision metrics making it more flexible and dynamic, to better adjust to the overall load distribution, while new KPIs and extended experimentation scenarios will be utilized in stage‐2. The results obtained from the Fed4FIRE+ testbeds will guide us towards the optimization of the load balancing process between MEC environments and concurrently will allow us to provide valuable feedback to Fed4FIRE+ community on Mobile Cloud Computing (MCC).Remark: The information in this section may be used in public documents and reports by the Fed4FIRE+ consortium.

MATERIALS