Brian Kim

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2020-205

December 17, 2020

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-205.pdf

By expanding the computational power of the public cloud to include resources closer to clients, edge computing serves as a natural extension of cloud computing. The resulting reductions in latency and network bandwidth have continued attracting developers to deploy applications at the edge. To accommodate the evolving ecosystem of edge applications, however, the underlying system must be rich in functionality, providing primitives for handling the geo-distributed topology and dynamic environment, while simultaneously delivering on client needs. Current offerings have not been holistic along these dimensions, as core abstractions such as shared data and location-agnostic communication have largely been overlooked. In this paper, we articulate the requirements of edge applications that must be addressed to encourage application evolvability. We follow up by proposing our Vertex system, a programming model and underlying execution environment that provides a unified approach for programming the edge.

Advisors: Scott Shenker

---

BibTeX citation:

@mastersthesis{Kim:EECS-2020-205,
    Author= {Kim, Brian},
    Title= {Vertex: Programming the Edge},
    School= {EECS Department, University of California, Berkeley},
    Year= {2020},
    Month= {Dec},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-205.html},
    Number= {UCB/EECS-2020-205},
    Abstract= {By expanding the computational power of the public cloud to include resources closer to clients, edge computing serves as a natural extension of cloud computing. The resulting reductions in latency and network bandwidth have continued attracting developers to deploy applications at the edge. To accommodate the evolving ecosystem of edge applications, however, the underlying system must be rich in functionality, providing primitives for handling the geo-distributed topology and dynamic environment, while simultaneously delivering on client needs. Current offerings have not been holistic along these dimensions, as core abstractions such as shared data and location-agnostic communication have largely been overlooked. In this paper, we articulate the requirements of edge applications that must be addressed to encourage application evolvability. We follow up by proposing our Vertex system, a programming model and underlying execution environment that provides a unified approach for programming the edge.},
}

EndNote citation:

%0 Thesis
%A Kim, Brian 
%T Vertex: Programming the Edge
%I EECS Department, University of California, Berkeley
%D 2020
%8 December 17
%@ UCB/EECS-2020-205
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-205.html
%F Kim:EECS-2020-205