Towards an internet-scale overlay network for latency-aware decentralized workflows at the edge

Area of Science:

• Computer Science
• Distributed Systems
• Network Engineering

Background:

• Edge computing offers cloud-like services with lower latency than traditional cloud environments.
• Challenges exist in seamless service discovery, secure sharing, and unified interfaces for edge computing at internet scale.
• Lack of trust and interoperability hinders workflow composition from diverse edge services.

Purpose of the Study:

• To design Viseu, a latency-aware blockchain framework for Virtual Internet Services at the Edge.
• To address challenges in edge service discovery, secure sharing, and workflow composition.
• To facilitate scientific collaboration by enabling native service composition across peers.

Main Methods:

• Developed Viseu, a blockchain framework incorporating latency and peer reputation for service selection.
• Enabled secure resource, service, and data sharing among peers in an untrusted environment.
• Facilitated workflow composition using peer services, moving beyond traditional provider-consumer roles.

Main Results:

• Viseu provides a solution for network service discovery at the edge, considering latency and reputation.
• Peers can securely share resources and services without relying on trusted third parties.
• Enabled native scientific collaboration through flexible workflow composition from diverse edge services.

Conclusions:

• Viseu effectively minimizes end-to-end latency and data loss by utilizing services from multiple edge peers.
• The framework promotes scientific collaboration by allowing seamless integration of services from various providers.
• Viseu establishes a scalable and secure platform for edge computing services at the internet scale.