Wireless Distributed Computing

Description | People | Publications


We are witnessing a rapid growth in the number of connected devices, which is only going to grow further with the Internet of Things (IoT). Most of the these devices are constrained by their batteries and have limited processing capabilities, and thus cannot locally run computationally intensive tasks. These connected devices can, however, share resources with each other and collaborate to create a platform with large computational resources for the execution of the complex applications. This is in contrast to the traditional cloud services that provide high performance and reliable servers. With long RTTs of a wide area network and possibly long setup times, the remote servers might incur a high delay. The connected devices, we consider, are not as powerful as cloud servers, but can be accessed by faster device-to-device (D2D) communication.


In above figure, we illustrate the basic idea of WDC. Given an application consisting of multiple tasks, we wish to assign them to the networked devices with the goal of minimizing the total expected cost. The cost can depend on performance metrics like the energy consumption and the overall application latency. We analyze different D2D transmission models and the computational models, design algorithms to minimize the execution cost and prove performance guarantees for them.

Note that these collaborating devices are connected with wireless links which have time-varying gains. The collaborating wireless devices are mobile and can get disconnected arbitrarily and come back later. The link qualities are, therefore, highly dynamic and variable. The quantity of the resources shared by a device is dependent on its own local processes. Hence, the performance of each device may vary with time.

The models for the transmission link and computational speed considered are as follows:

  • Known Deterministic
  • Known Stochastic
  • Unknown Stochastic
  • Adversarial


Yi-Hsuan Kao
Pranav Sakulkar
Kwame Lante Wright
Bhaskar Krishnamachari


  1. Yi-Hsuan Kao, Bhaskar Krishnamachari, “Optimizing Mobile Computational Offloading with Delay Constraints“, IEEE Global Communications Conference (GLOBECOM), 2014.
  2. Yi-Hsuan Kao, Bhaskar Krishnamachari, Moo-Ryong Ra and Fan Bai, “Hermes: Latency Optimal Task Assignment for Resource-constrained Mobile Computing“, International Conference on Computer Communication (INFOCOM), 2015.
  3. Yi-Hsuan Kao, “Optimizing Task Aassignment for Ccollaborative Computing over Heterogeneous Network Devices“, PhD Thesis, USC, 2016.