We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.DC

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Distributed, Parallel, and Cluster Computing

Title: A Case for a Programmable Edge Storage Middleware

Abstract: Edge computing is a fast-growing computing paradigm where data is processed at the local site where it is generated, close to the end-devices. This can benefit a set of disruptive applications like autonomous driving, augmented reality, and collaborative machine learning, which produce incredible amounts of data that need to be shared, processed and stored at the edge to meet low latency requirements. However, edge storage poses new challenges due to the scarcity and heterogeneity of edge infrastructures and the diversity of edge applications. In particular, edge applications may impose conflicting constraints and optimizations that are hard to be reconciled on the limited, hard-to-scale edge resources. In this vision paper we argue that a new middleware for constrained edge resources is needed, providing a unified storage service for diverse edge applications. We identify programmability as a critical feature that should be leveraged to optimize the resource sharing while delivering the specialization needed for edge applications. Following this line, we make a case for eBPF and present the design for Griffin - a flexible, lightweight programmable edge storage middleware powered by eBPF.
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Cite as: arXiv:2111.14720 [cs.DC]
  (or arXiv:2111.14720v1 [cs.DC] for this version)

Submission history

From: Animesh Trivedi [view email]
[v1] Mon, 29 Nov 2021 17:21:09 GMT (692kb,D)

Link back to: arXiv, form interface, contact.