Caching‐enabled networks and its applications into the emerging networking paradigms and technologies
Jianhui Lv, Xingwei Wang, Qing Li, Tian Gong Pan · Internet Technology Letters · 2021
Recently, the characteristic of in-network caching has been accepted and canonized by some prevalent networking paradigms, such as Peer-to-Peer Networking (P2PN), Content Delivery Networks (CDN), Information-Centric Networking (ICN), Hybrid ICN (HICN) and IPv6 Content Networking (6CN). Among them, (a) P2PN, CDN and ICN are proposed by the academic communities, while HICN and 6CN are proposed by the industrial community, i.e., Cisco; (b) P2PN, CDN and 6CN are overlay solutions, while ICN and HICN are the built network-layer solutions. No matter which networking paradigms, they all belong to the Caching-Enabled Networks (CEN).1 With the rapid development of mobile devices and applications diversification, the CEN has wide application value thanks to its in-network caching characteristic. For example, applying ECN into some networking paradigms and technology fields can facilitate their performance improvement. This Editorial includes recent publications that investigate CEN and the CEN-related applications into the emerging networking paradigms and technologies. This special issue provides a valuable reference to the field of CEN. In particular, this special issue selects the high-quality papers based on their contributions. We cannot guarantee that these selected papers cover all the topics of this special issue. However, we are pleased to present this special issue, which covers some key problems in CEN and its related applications. The first paper "Q-learning-based congestion control strategy for information-centric networking"2 exploits the Q-learning-based reinforcement learning method to further address the ICN congestion control issue. Meanwhile, the Q-learning is used to learn the network state, at the same time, a value which depends on the transmission delay is generated. The second paper "6CN-driven routing and content delivery scheme for building the smart physical education"3 proposes a novel system framework based on 6CN to facilitate the building of smart physical education, including content request stage based on segment routing and content delivery stage based on identity forwarding. The third paper "In-network caching based NBA-traffic offloading mechanism: A D2D communication mode"4 studies the NBA-traffic offload mechanism based on in-network caching and D2D communication mode, where the router is enabled with the caching ability and the information exchange is based on the direct communication between two D2D groups. The fourth paper "In-network caching and SDN-aided transmission synchronization for cross-domain English translation"5 leverages in-network caching to reduce the cross-domain latency and SDN6 to synchronize cache transmission, which adopts Lyapunov optimization to stabilize streaming data. The fifth paper "Ant colony optimization based Information-Centric networking delivery strategy via flow analysis and scheduling"7 proposes the content delivery strategy of ICN with the load balance consideration. It uses ACO8 to schedule the optimal path based on the dynamically analyzed network status. The sixth paper "A hierarchical task scheduling strategy in mobile edge computing"9 proposes a hierarchical task scheduling strategy. The system framework includes the edge computing layer and the frog computing layer, in which the former is responsible for processing the simple tasks (outputting the sketchy result) and the latter is used to addressing the complex tasks (outputting the precise result). The seventh paper "Distribution of teaching surveillance video via edge computing"10 firstly proposes a framework for teaching surveillance video content distribution through the network, storage, and computing capabilities of edge computing; then provides an edge caching architecture and a cache update strategy by using an LSTM network. The eighth paper "Mobile edge computing based video surveillance model for improving the performance of extended training"11 takes full advantages of mobile edge computing to design the video surveillance model, improving the performance of extended training. The proposed framework consists of cloud center, wireless base station, edge computing devices and mobile wearables. The ninth paper "In-network caching based live broadcasting for short track speed skating"12 proposes an SDN-based ICN-assisted edge network live broadcast architecture to realize the allocation of transcoding tasks and transmission of massive streaming data. The tenth paper "Adaptive genetic algorithm inspired energy-efficient cache deployment strategy for caching-enabled networks"13 proposes an energy-efficient cache deployment strategy, including two stages. The first stage is the mathematical modeling for cache hit ratio and energy consumption, in which a bi-objective optimization problem is built. The second stage is the model solving, where the adaptive genetic algorithm is used. The eleventh paper "Software-defined information-centric networking based exercise intensity evaluation of volleyball player: An efficient convolutional neural network method"14 conducts the exercise intensity evaluation of volleyball player based on SDN and ICN. To be specific, SDN is used to monitor the exercise intensity data via the functions of centralized control and global network view; ICN router is responsible for storing the important and frequently used exercise intensity data via the ability of in-network caching. In addition, CNN model is exploited to train the exercise intensity data and further to conduct the behaviors of volleyball player. The twelfth paper "ABR optimization for live sports events with ICN-enabled network framework"15 uses the Q-learning method based on neuronal population coding to optimize ABR. Besides, it also introduces ICN paradigm to improve the traditional distribution framework in order to store the hot sports streaming at the edge caches. The thirteenth paper "A GAN-based data augmentation method for human activity recognition via the caching ability"16 expands the origin dataset via a GAN-based augmentation neural network. The experimental results suggest that the generated data has a certain substitution and complementary effect in terms of the real data based on the caching deployment ability. The fourteenth paper "Light in-network caching enabled IP network: A large-scale content delivery in MAN"17 presents a light content delivery scheme depending on identity, with the chain-based fine-grained caching support. Given the NP-hardness of delivery scheme in mesh topology, this paper makes a case study in MAN to address the large-scale content delivery scenario. The fifteenth paper "Stochastic neural network based data analysis-related talent recruitment optimization via CDN server"18 uses SNN to optimize the data analysis-related talent recruitment. Furthermore, in order to guarantee the on-line resume screening, this paper does resume screening under CDN scenario, where CDN has two functions, that is, data storage and hot resumes recommendation. The sixteenth paper "ICN routing strategy based on reinforcement learning and neural network"19 presents a RL and NN based ICN routing strategy to improve the distribution efficiency and the stability. Meanwhile, RL is used to obtain the stable routing, while NN is used to predict network delay and enhance the routing efficiency. The seventeenth paper "A caching-enabled light control scheme for aerobics video"20 proposes a transmission control scheme for aerobics video. Meanwhile, video compression and congestion control are performed in the simultaneous way, and they are completed by using the explored in-network cache. The eighteenth paper "Edge computing based ice-snow data analysis with NDN paradigm support"21 uses the inherent name space of NDN to differentiate different application types. Furthermore, regarding edge computing, this paper adopts PSO to perform the tasks offloading. The nineteenth paper "ICN-driven group psychology visualization analysis mechanism using reinforcement learning"22 uses RL to achieve the group psychology visualization analysis instead of the abstract data presentation. In order to accelerate the process of data visualization under the large-scale environment, this paper also introduces ICN paradigm to support the name-based community detection by separating IP addresses. The twentieth paper "Tennis-resource allocation mechanism in software-defined data center network"23 studies the multi-tenant tennis-resource allocation scenario. At first, the max-min fairness is used to determine the allocated number of tennis-resources for each round. Then, the allocation state is adjusted according to the current situation until the final allocation decision is determined. The twenty-first paper "CNN-based politics public opinion analysis of undergraduates: A case study with CDN deployment"24 uses CNN to realize the politics public opinion analysis of undergraduates, which has two functions. On one hand, greatly help the government departments eliminate the crisis timely; On the other hand, correctly guide the political education of undergraduates. Besides, this paper also presents a case study based on CDN deployment. The twenty-second paper "Smart opinion formation maximization to improve political education of undergraduates in Social Information-Centric Networking"25 considers the scenario regarding the political education of undergraduates and proposes a heuristic opinion formation maximization algorithm, including: the opinion formation model is built based on individual intimacy and social influence; in order to make sure the opinion formation maximization, the collaborative evolution model is built. The twenty-third paper "Natural language processing-based lexical meaning analysis: An application of in-network caching-oriented translation system"26 addresses the communication bottleneck of distributed parallel training. At first, an in-network caching-oriented training system architecture is proposed, which utilizes the in-network caches to reduce the parameter transmission and reduce the communication overhead. Then, an improved attention model based on the variation algorithm is proposed to further reduce the model size and improve the lexical meaning analysis capability. The twenty-fourth paper "Online preschool education optimization based on edge computing in the era of COVID-19"27 adopts the edge computing to optimize online preschool education, where a task unloading algorithm based on genetic algorithm is designed to minimize the computing delay of terminal tasks. The twenty-fifth paper "0-1 Knapsack problem driven resource scheduling in caching-enabled network: A case study on sports video"28 converts the cache resource scheduling into a KP01 with a given total cache budget. For the proposed strategy in CEN, this paper makes a detailed case study by considering a hot application, that is, sports video transmission. The Guest Editors would like to thank the Authors and the Reviewers for their effort in the production of these selected papers. In particular, we would like to thank K Li and X Fu for the effort of CFP, even through this special issue did not receive some submissions from USA and Germany.