Abstract

In-network caching as a core function in ICN can greatly improve the content distribution efficiency and reduce redundant network traffic. It is a challenge to design caching strategies in ICN to improve the performance of in-network caching. Traditional lightweight strategies are inefficient due to problems such as edge caching pollution and slow diffusion speed. Current caching strategy with popularity prediction does not fully consider the content capacity of ICN network. They usually have huge computation overhead and cannot meet line-speed requirement. This paper focuses on the research of lightweight on-path caching. Our caching strategy is proposed according to the edge-degree of routers along the forwarding path. This strategy can quickly form a hierarchical content placement with little overhead. A large number of experiments are conducted to examine the performance of our strategy in different topologies and workloads. The simulation result reveals that our caching strategy can reduce the average download latency by 10.2% and increase the cache hit ratio by 9.6% at most compared with other lightweight strategies.

Highlights

  • INTRODUCTIONThe system has to have a holistic view of the topology and manage the dynamic topology with very high cost

  • Information-Centric Networking (ICN) [1] is proposed as an effective future Internet architecture to address the problem of content distribution efficiency in current network architecture

  • 2) We propose a Node Edge-degree-based Caching strategy to push popular content to the edge of the network quickly and to help forming a hierarchical content placement in the network

Read more

Summary

INTRODUCTION

The system has to have a holistic view of the topology and manage the dynamic topology with very high cost These strategies can improve the cache hit ratio by putting popular content copies in the center of networks, but they cannot effectively reduce the download latency because requests have to travel multiple routers to hit the node with high centrality. B. PROBLEM FORMULATION AND ANALYSIS Average download latency is an important indicator to measure performance of network caching, which is the time interval between when the content request is sent and the data packet arrives. That we have the request path of each node for the content item f according to the spanning forest Gf , which can be derived from all the connected graph Tfi, we can start to calculate the download latency. With Proposition 1 and 2, we conclude that content should be cached hierarchically in the network, the more popular content should be closer to the edge of the network

DESIGN OF NEC
PROBLEM AND DISCUSSION IN NEC
1: Switch end receives a data packet 2: read the local edge-degree as e 3
SIMULATION SETUP
Findings
CONCLUSION
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call