An Irregular Graph Based Network Code for Low-Latency Content Distribution.

Area of Science:

• Computer Science
• Information Theory
• Network Engineering

Background:

• Increasing demand for low-latency content distribution requires efficient network coding strategies.
• Existing generation-based network coding often uses constant generation sizes, limiting adaptability.

Purpose of the Study:

• To propose a framework for designing generation-based network codes with non-constant generation sizes.
• To optimize network code design for improved decoding performance and reduced overhead.

Main Methods:

• Characterizing network codes as irregular graphs and evaluating graph properties.
• Extending and-or tree evaluation for decoding performance analysis.
• Formulating optimization problems to design degree distributions for generation sizes.

Main Results:

• Achieved lower decoding cost and transmission overhead compared to constant generation size schemes.
• Demonstrated satisfactory decoding speed.
• Validated the effectiveness of the proposed optimization framework.

Conclusions:

• The proposed generation-based network coding with variable generation sizes offers an efficient solution for low-latency content distribution.
• This approach is particularly suitable for dynamic, unknown, or cyclic network topologies where node resources are constrained.