Resource Allocation for Mobile Edge Computing Platform with Multiple Resource Providers

Main Achievement: 

To tackle the main issue of how to strike a good balance between computing power and communication resources, contributing to total service latency, we first model the latency-driven cooperative task computing as an optimization problem, and an algorithm based on dynamic programming, namely, CTC-DP, is proposed for the cooperative task computing in the special case with a single user. 

Then we design a heuristic algorithm, CTC-All, which combines the CTC-DP approach with “one-for-all” concept to provide an approximate solution for both heterogeneous resource allocation and cooperative task computing in the general case with multiple users. 

Besides, we develop a business/incentive-model whose key tasks are to decide the service fee (how much should application service providers pay for the resource) and resource allocation for application/service among multiple resource providers (which providers supply how much amount of resource to which users)


Quantitative Impact:

The proposed scheme, compared with single server approach, can significantly reduce about 88% service latency of the cooperative task computing operation and properly deal with heterogeneous resource allocation while handling the tradeoff between communication resource allocation and computing task assignment in the time domain. The scheme also can reach a win-win situation for each user. 

Further, via leveraging dynamic-programming design approach, we efficiently reduce total running time with the increasing number of users, which validates the feasibility and scalability of our proposed scheme.

End Goal:

We want to develop a highly efficient and sustainable Fog/edge computing platform. High efficiency is achieved via building a high-performance cooperative task computing framework with latency-driven task assignment and heterogeneous resource allocation across multiple fog/edge nodes. Sustainability can be ensured by developing a business/incentive-model whose goal is to maximize the total benefit of computing/storage resource providers, application service providers, and users under the condition that everyone should have a positive benefit.

 (Updated in Jul, 2017)