基于节点角色划分的混合ICN缓存协调方案

26 浏览量更新于2024-08-29 收藏 418KB PDF 举报

“A Hybrid ICN Cache Coordination Scheme Based on Role Division between Cache Nodes” 本文是一篇研究论文，探讨了信息中心网络（Information-Centric Networking，ICN）中的混合缓存协调策略，该策略基于缓存节点之间的角色划分。ICN是一种与传统TCP/IP主机为中心的方法截然不同的网络范式，它依赖于自认证内容、广泛分布的网络内缓存和灵活的请求路由，以在架构层面上提供高效且可扩展的内容分发。ICN的缓存机制与传统的Web缓存有所不同，由于其任意的网络拓扑和对线速操作的需求，使得缓存协调成为一项关键任务。大多数现有的ICN缓存策略主要关注如何优化内容存储和检索，以减少延迟和网络拥塞。然而，由于ICN网络中可能存在大量分布式缓存节点，单纯依赖本地决策往往无法实现全局最优。因此，本文提出了一种混合策略，通过将缓存节点划分为不同角色，如权威节点、辅助节点等，以实现更高效的协调。这种角色划分允许节点根据其特定职责来管理内容，从而提高整体性能。具体来说，权威节点可能负责存储热门和长期稳定的内容，而辅助节点则可能处理临时热点或区域性的内容需求。通过这种方式，可以更好地平衡负载，减少冗余复制，并确保内容新鲜度。此外，论文可能还讨论了如何动态调整节点角色以适应网络条件的变化，以及如何利用兴趣包（Interest Packet）的反馈信息来优化内容分发策略。此外，文章可能还涉及到了缓存选择算法、内容替换策略、以及如何处理内容的版本控制和安全问题。这些算法和策略可能采用了机器学习或优化理论，以在满足用户需求的同时，最大化网络资源的利用率。这篇论文深入研究了ICN环境下通过缓存节点角色划分实现的混合缓存协调机制，为ICN的高效运行提供了新的视角和解决方案。这种方法有望解决ICN中的缓存协调难题，提升整个网络的性能和用户体验。

A hybrid ICN cache coordination scheme

based on role division between cache nodes

Guoqiang Zhang

∗†§

, Xiaohui Wang

†

,QianGao

†¶

and Zhen Liu

‡

∗

State Key Laboratory of Information Security, Institute of Information Engineering,

Chinese Academy of Sciences, Beijing, China

†

School of Computer Science and Technology, Nanjing Normal University, Nanjing, China

‡

Beijing Jiaotong University, Beijing, China

Department of Computer Science and Engineering, University of California, Riverside, USA

University of Massachusetts Boston, USA



Corresponding author’s email address:guoqiang@ict.ac.cn

Abstract—Information-Centric Networking(ICN) is a para-

digm shift from traditional TCP/IP’s host-centric approaches,

which relies on self-certiﬁable content, built-in pervasive in-

network caching and ﬂexible request routing to provide native

support for efﬁcient and scalable content distribution from the

architectural layer. Caching in ICN diverts from traditional Web

Caching due to its arbitrary network topology and the require-

ment for line-rate operation, which makes cache coordination a

crucial task. However, most of the cache coordination schemes

proposed to date use uniform caching coordination policies within

the same cache network, which is either hard to achieve good

balance between user QoE and resource utilization or hard to

scale. This paper proposes a hybrid cache coordination scheme

that bases on cache node role separation. The cache network

is divided into a core area and several edge areas. The core

area applies an explicit and off-path cache coordination policy

such as the HASH-based cache coordination to improve cache

diversity within the cache network, whereas the edge areas can

apply any implicit and on-path cache coordination policies such

as LCE and LCD. The separation allows this scheme to be easily

extended to large networks, and simulation results show that the

hybrid schemes can achieve good balance between user QoE and

network resource utilization.

I. BACKGROUND

Recently, the usage pattern of the Internet has shifted

from sender-driven end-to-end communication to receiver-

driven many-to-one content retrieval. The mismatch between

TCP/IP’s unicast communication primitive and the applica-

tion’s usage model is one of the root causes for today’s

trafﬁc explosion. Information Centric Networking (ICN) is a

paradigm shift from the host-centric approaches [1] [2] [3]

[4]. ICN relies on the decoupling between content providers

and consumers, self-certiﬁable content, pervasive in-network

caching and ﬂexible request routing to satisfy the massive,

asynchronous and heavy-tailed user requests, which provides

native support for efﬁcient and scalable many-to-one content

distribution from the architectural layer. In addition, it is

also believed that ICN is a good candidate to reduce energy

consumption in massive content dissemination [5] [6].

Pervasive in-network caching is regarded as a fundamental

building block of ICN. The pervasiveness makes ICN’s caching

mechanism different from traditional Web caching. Unlike

Web caching, the cache network topology in ICN is no

longer hierarchical, but arbitrary [7]. Meanwhile, as a basic

service, ICN’s caching should support line-speed operation,

which limits the cache capacity of each cache node [8] [19].

Hence, cache coordination between different nodes becomes

very crucial to reduce cache redundancy and improve the utility

of cache resources.

Given a cache network, the primary goals of caching are

mainly two-folded: to improve the user’s quality of experi-

ence(QoE) from users’ perspective, and to improve network

resource utilization and reduce cross-domain trafﬁc from the

network’s perspective [12] [10]. Unfortunately, these two ob-

jectives are sometimes contradictive. In order to improve the

user’s QoE, one tendency is to place as many popular content

as possible to those cache nodes near users. This tendency,

just as the on-path caching policy LCE does [18], will make

many duplicated content copies of popular content, and thus

reducing the cache diversity within the network. On the other

hand, to fulﬁll the latter goal, one tends to increase the cache

diversity. One extreme of this approach is the HASH-based

cache coordination [9] [11]. This off-path cache coordination

policy leaves at most one copy of each content in the cache

network regardless of the content popularity. This, however,

may degrade the user’s experience. So, a good cache coordi-

nation scheme needs to balance between these two goals.

Besides the tradeoff between QoE and network resource

utilization, there is a growing concern of how to make cache

coordination between different networks [20] and how to make

a cache coordination scheme scalable so that it can be used

in large ISPs. This requires the ability for different cache

networks or different network portions of the same large

network to apply different cache coordination policies rather

than to adopt a universal cache coordination policy. Conse-

quently, a good cache coordination scheme needs to support

the enforcement of ﬂexible or even autonomous policies.

This paper proposes a hybrid cache coordination scheme

based on role division between cache nodes. It divides a

network into one core area and several edge areas. The scheme

can make use of existing partition of large ISPs to establish

the partition, such as OSPF areas. And, for networks without

pre-existing partition, it applies the k -core decomposition to

partition the cache nodes. Cache nodes in the core area applies

下载后可阅读完整内容，剩余5页未读，立即下载

weixin_38560768

粉丝: 5
资源: 895

基于节点角色划分的混合ICN缓存协调方案

mipi转lvds icn6202参考资料

mipi转rgb icn6211参考资料

An adaptive per-application storage management scheme based on manifold learning in information centric networks

ICN

Geometric Routing on Flat Names for ICN

Link Prediction in Signed Networks based on The Similarity and Structural Balance Theory

Context-Aware Distributed Service Provisioning based on Anycast for Information-Centric Network

icn.rar_ICN

On Allocating Cache Resources to Content Providers

中心网络ICN

最新资源