Enhancing the Isolation and Performance of Control Planes for Fog Computing

Area of Science:

• Computer Science
• Networking
• Distributed Systems

Background:

• Fog computing enables low-latency data processing for IoT applications by situating resources near devices.
• Software-Defined Networking (SDN) allows programmable network control logic on a decoupled control plane.
• Existing SDN control planes lack the isolation and high performance needed for multi-tenant, scalable fog computing.

Purpose of the Study:

• To present optimization techniques for Linux-based SDN control planes.
• To enhance isolation and high performance crucial for fog computing environments.
• To address limitations in current control plane architectures for multi-tenancy and scalability.

Main Methods:

• Developed Separate Execution Environment (SE2) to isolate multiple control plane instances.
• Introduced Separate Packet Processing (SP2) to simplify the Linux network stack.
• Evaluated techniques on commodity hardware.

Main Results:

• Achieved a four-fold increase in maximum control plane performance compared to native Linux.
• Demonstrated strong isolation between control plane instances.
• Validated effectiveness on standard hardware.

Conclusions:

• The proposed SE2 and SP2 techniques significantly improve SDN control plane performance and isolation in fog computing.
• These optimizations are vital for supporting scalable and multi-tenant IoT applications.
• The techniques offer a practical solution for enhancing fog computing infrastructure.