Search

Dynamic Adaptive Streaming over HTTP (DASH) has prevailed as the dominant way of video transmission over the Internet. This technology is based on receiving small sequential video segments from a server. However, one challenge that has not been adequately examined is the obtainment of video segments in a way that serves both the needs of the network and the improvement in the Quality of Experience (QoE) of the users. One effective way to achieve this is to implement and study caching and DASH technologies together. This paper investigates this issue by simulating a network with multiple video servers and a video client. It then implements both the peer-to-many communications in the context of adaptive video streaming and the video server caching algorithm based on proposed criteria that improve the status of the network and/or the user. Specifically, we investigate the scenario of delivering DASH-based content with the help of an intermediate server, apart from a main server, to demonstrate possible caching benefits for different sizes of intermediate storage servers. Extensive experimentation using emulation reveals the interplay and delicate balance between caching and DASH, guiding such network design decisions. A general tendency found is that, as the available buffer size increases, the video playback quality increases to some extent. However, at the same time, this improvement is linked to the random cache selection algorithm.

With outstanding features, machine learning (ML) has become the backbone of numerous applications in wireless networks. However, the conventional ML approaches face many challenges in practical implementation, such as the lack of labeled data, the constantly changing wireless environments, the long training process, and the limited capacity of wireless devices. These challenges, if not addressed, can impede the effectiveness and applicability of ML in wireless networks. To address these problems, transfer learning (TL) has recently emerged to be a promising solution. The core idea of TL is to leverage and synthesize distilled knowledge from similar tasks and valuable experiences accumulated from the past to facilitate the learning of new problems. By doing so, TL techniques can reduce the dependence on labeled data, improve the learning speed, and enhance the ML methods' robustness to different wireless environments. This article aims to provide a comprehensive survey on the applications of TL in wireless networks. Particularly, we first provide an overview of TL, including formal definitions, classification, and various types of TL techniques. We then discuss diverse TL approaches proposed to address emerging issues in wireless networks. The issues include spectrum management, signal recognition, security, caching, localization, and human activity recognition, which are all important to next-generation networks, such as 5G and beyond. Finally, we highlight important challenges, open issues, and future research directions of TL in future wireless networks. Ministry of Education (MOE) National Research Foundation (NRF) This work was supported in part by the Australian Government through the Australian Research Council's Discovery Projects Funding Scheme under Project DE210100651 (by Dr. Hoang Dinh); in part by the Joint Technology and Innovation Research Centre-a partnership between the University of Technology Sydney and the VNU Ho Chi Minh City University of Technology (VNU HCMUT); in part by the Programme DesCartes-the National Research Foundation, Prime Minister's Office, Singapore, through its Campus for Research Excellence and Technological Enterprise (CREATE) Programme and its Emerging Areas Research Projects (EARP) Funding Initiative; in part by the Alibaba Group through the Alibaba Innovative Research (AIR) Program and the Alibaba-NTU Singapore Joint Research Institute (JRI); in part by the National Research Foundation, Singapore, through the AI Singapore Programme (AISG) under Grant AISG2-RP-2020-019; and in part by the Singapore Ministry of Education (MOE) Tier 1 under Grant RG16/20. The work of Cong T. Nguyen was supported in part by Vingroup JSC and in part by the Master, PhD Scholarship Programme of the Vingroup Innovation Foundation (VINIF), Institute of Big Data, Code 2021.TS.006. The work of Quoc-Viet Pham and Won-Joo Hwang was supported in part by the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (MSIT) under Grant NRF-2019R1C1C1006143 and Grant NRF-2019R1I1A3A01060518, in part by the Institute of Information and Communications Technology Planning and Evaluation (IITP) Grant funded by the Korea Government (MSIT) under Grant 2020-0-01450 [Artificial Intelligence Convergence Research Center (Pusan National University)], and in part by the BK21 Four, Korean Southeast Center for the 4th Industrial Revolution Leader Education.

In today's data landscape, data streams are well represented. This is mainly due to the rise of data-intensive domains such as the Internet of Things (IoT), Smart Industries, Pervasive Health, and Social Media. To extract meaningful insights from these streams, they should be processed in real time, while solving an integration problem as these streams need to be combined with more static data and their domain knowledge. Ontologies are ideal for modeling this domain knowledge and facilitate the integration of heterogeneous data within data-intensive domains such as the IoT. Expressive reasoning techniques, such as OWL2 DL reasoning, are needed to completely interpret the domain knowledge and for the extraction of meaningful decisions. Expressive reasoning techniques have mainly focused on static data environments, as it tends to become slow with growing datasets. There is thus a mismatch between expressive reasoning and the real-time requirements of data-intensive domains. In this paper, we take a first step towards bridging the gap between expressivity and efficiency while reasoning over high-velocity IoT data streams for the task of event enrichment. We present a structural caching technique that eliminates reoccurring reasoning steps by exploiting the characteristics of most IoT streams, i.e., streams typically produce events that are similar in structure and size. Our caching technique speeds up reasoning time up to thousands of times for fully fledged OWL2 DL reasoners and even tenths and hundreds of times for less expressive OWL2 RL and OWL2 EL reasoners.

This paper considers LEO satellite- and cache-assisted UAV communications for content delivery in terrestrial networks, which shows great potential for next-generation systems to provide ubiquitous connectivity and high capacity. Specifically, caching is provided by the UAV to reduce backhaul congestion, and the LEO satellite supports the UAV's backhaul link. In this context, we aim to maximize the minimum achievable throughput per ground user (GU) by jointly optimizing cache placement, the UAV's resource allocation, and trajectory while cache capacity and flight time are limited. The formulated problem is challenging to solve directly due to its non-convexity and combinatorial nature. To find a solution, the problem is decomposed into three sub-problems: (1) cache placement optimization with fixed UAV resources and trajectory, followed by (2) the UAV resources optimization with fixed cache placement vector and trajectory, and finally, (3) we optimize the UAV trajectory with fixed cache placement and UAV resources. Based on the solutions of sub-problems, an efficient alternating algorithm is proposed utilizing the block coordinate descent (BCD) and successive convex approximation (SCA) methods. Simulation results show that the max-min throughput and total achievable throughput enhancement can be achieved by applying our proposed algorithm instead of other benchmark schemes., 14 pages, 6 figures

Cache-enabled Fog radio access network (F-RAN) is a promising technology to alleviate the traffic congestion and boost the contents delivery success rate. The efficiency of disseminating the cached contents in F-RAN can be boosted by enabling multicast service at the fog access points (F-APs). This paper proposes a joint random caching and multicasting optimization scheme for wireless backhauled F-RAN. Using tools from stochastic geometry, the expression of the successful transmission probability (STP) is derived by carefully analyzing the different types of serving F-APs and interferers. Then, a closed-form expression of the asymptotic STP in the high SNR region is derived to reduce the complexity. The joint caching and multicasting optimization problem is formulated to maximize the STP. The optimization problem is complex and non-convex in general. A novel projected cuckoo search algorithm (PCSA) is proposed to obtain the optimal content placement that maximizes the STP. The numerical simulation results show that PCSA outperforms the original cuckoo search algorithm (CSA) and the proposed asymptotically joint caching and multicasting scheme outperforms the benchmark caching schemes by up to 15% higher STPs.

High-speed internet has boosted clients’ traffic needs. Content caching on mobile edge computing (MEC) servers reduces traffic and latency. Caching with MEC faces difficulties such as user mobility, limited storage, varying user preferences, and rising video streaming needs. The current content caching techniques consider user mobility and content popularity to improve the experience. However, no present solution addresses user preferences and mobility, affecting caching decisions. We propose mobility- and user-preferences-aware caching for MEC. Using time series, the proposed system finds mobility patterns and groups nearby trajectories. Using cosine similarity and CF, we predict and cache user-requested content. CF predicts the popularity of grouped-based content to improve the cache hit ratio and reduce delay compared to baseline techniques.

Distributed systems offer a reliable and scalable solution for tackling massive and complex tasks that cannot be handled by a single computer. However, standard consensus protocols used in such systems often replicate data without considering the workload, leading to unnecessary retransmissions. This thesis proposes using machine learning (ML) to optimize consensus protocols and make them adaptable to recurring workloads. It introduces a cache that encodes frequently-transmitted data between nodes to reduce network traffic. To implement this, the thesis builds a caching layer at all nodes using the decided logs, which represent a consistent view of the application history. The cache can encode and decode incoming log entries to reduce the average message size and improve throughput under limited network bandwidth. The thesis selects an ML-based model that combines various caching policies and adapts to changing access patterns in the workload. Experimental results show that this approach can improve throughput up to 250%, assuming negligible preprocessing overhead. Distribuerade system erbjuder en pålitlig och skalbar lösning för att hantera massiva och komplexa uppgifter som inte kan hanteras av en enskild dator. Konventionella konsensusprotokoll som används i dessa system replikerar emellertid ofta data utan att ta hänsyn till arbetsbelastningen, vilket leder till överflödig dataöverföring. Denna avhandling föreslår att använda maskinin lärning (ML) för att optimera konsensusprotokoll och göra dem anpassade till återkommande mönster i arbetsbelastningen. Den introducerar en cache som kodar och komprimerar data som ofta överförs mellan noder för att minska nätverkstrafiken. För att implementera detta byggs ett cache baserat på den bestämda loggen på alla noder, som representerar en konsekvent syn på programhistoriken. Cachen kan koda inkommande data för att minska genomsnittlig meddelandestorlek och förbättra genomströmning under begränsad nätverksbandbredd. En ML-baserad modell som kombinerar olika cachningpolicyer och anpassar sig till ändrade åtkomstmönster i arbetsbelastningen används. Experimentella resultat visar att denna metod kan förbättra genomströmningen med 250%, under förutsättning att förbearbetningsöverhuvudet är försumbart.

Caching is a fundamental technique widely used in the field of computing to reduce network traffic, server load, and latency. Storing frequently accessed data in a high-speed cache layer can make future requests process faster by involving fewer system components when generating and serving the response. Kalix is a software product that demands a caching solution since it faces latency and is frequently processing partially repeated queries. However, a cache does not guarantee improved performance, which is why the main problems of caching are: determining what content to cache, when to insert or remove cache content, implementing the caching logic, and deciding where to store the cache efficiently. Therefore, this paper theoretically investigates where a cache solution should be implemented within the Kalix system architecture to decrease latency and server load, and evaluates the subsequent cache implementation experimentally. As a result, a client-side cache is implemented which decreases the latency of Kalix by up to 74%, while reducing the I/O load and memory utilization on the server by 98%. The reason for the decrease is that the cache in the client can directly serve the majority of the content, allowing the servers of Kalix to do substantially fewer computations. The evaluation acts as a recommendation for the company behindKalix, Polystar, as to if a cache is beneficial and where the cache can efficiently be deployed, and this paper gives valuable insights into the decision-making of cache placement. Concludingly, implementing the cache positively impacts the Kalix user experience.

Online caching is among the most fundamental and well-studied problems in the area of online algorithms. Innovative algorithmic ideas and analysis - including potential functions and primal-dual techniques - give insight into this still-growing area. Here, we introduce a new analysis technique that first uses a potential function to upper bound the cost of an online algorithm and then pairs that with a new dual-fitting strategy to lower bound the cost of an offline optimal algorithm. We apply these techniques to the Caching with Reserves problem recently introduced by Ibrahimpur et al. [Ibrahimpur et al., 2022] and give an O(log k)-competitive fractional online algorithm via a marking strategy, where k denotes the size of the cache. We also design a new online rounding algorithm that runs in polynomial time to obtain an O(log k)-competitive randomized integral algorithm. Additionally, we provide a new, simple proof for randomized marking for the classical unweighted paging problem., LIPIcs, Vol. 261, 50th International Colloquium on Automata, Languages, and Programming (ICALP 2023), pages 80:1-80:20

Inom nätverkstekniken och datakommunikationen råder idag en tro på tekniken nätverkscache som kan spara data för att senare kunna hämta hem det snabbare. Tekniken har genom åren visat att den effektivt kan skicka den önskade data till sina klienter. Det finns flera cache-tjänster som använder tekniken för Windows-uppdateringar. Bland dessa finns Windows Server Update Services (WSUS) och LanCache. På uppdrag från företaget TNS Gaming AB jämförs dessa tjänster med varandra under examensarbetet. Nätverkscache är ett intressant forskningsområde för framtida kommunikationssystem och nätverk tack vare sina fördelar. Likaså är uppgiften om att jämföra cache-tjänsterna WSUS och LanCache intressant i och med det öppnar upp insikt om vilken tjänst är bättre för företaget eller andra intressenter. Både forskningsområdet och uppgiften är viktiga och intressanta när användare vill effektivisera användningen av internetanslutningen och bespara nätverksresurser. Därmed kan tekniken minska nedladdningstiden. Till det här arbetet besvaras frågor om vilken nätverksprestanda, resursanvändning och administrationstid respektive cache-tjänst har, och vilken cache-tjänst som lämpar sig bättre för företagets behov. I arbetet genomförs experiment, som omfattar tre huvudmättningar, och följs av en enfallstudie. Syftet med arbetet är att med hjälp av experimentets mätningar få en jämförelse mellan WSUS och LanCache. Resultatet av arbetet utgör sedan ett underlag för det framtida lösningsvalet. Resultaten består av två delar. Den första visar att båda cache-tjänsterna bidrar till kortare nedladdningstider. Den andra är att LanCache är bättre än WSUS när det gäller nätverksprestanda och resursanvändning, samt mindre administrationstid jämfört med WSUS. Givet resultat dras slutsatsen att LanCache är cache-tjänsten som är mest lämpad i det här fallet. In the field of network technology and data communication, there is a current belief in the technology of network caching, which can store data to later retrieve it more quickly. Over the years, this technology has proven its ability to efficiently deliver the desired data to its clients. There are several caching services that utilize this technology for Windows updates, among them are Windows Server Update Services (WSUS) and LanCache. On behalf of the company TNS Gaming AB, these services are compared to each other in this thesis. Network caching is an interesting area of research for future communication systems and networks due to its benefits. Likewise, the task of comparing the cache services WSUS and LanCache is interesting as it provides insights into which service is better suited for the company or other stakeholders. Both the research area and the task are important and intriguing when users seek to streamline the use of their internet connection and conserve network resources. Thus, the technology can reduce download times. For this work, questions about the network performance, resource usage, and administration time of each cache service are answered, as well as which cache service that is better suited to the company's needs. The work involves conducting experiments, including three main measurements, followed by a single case study. The purpose of the work is to compare WSUS and LanCache using the measurements from the experiment. The outcome of the work then forms a basis for future solution choice. The results consist of two parts. The first shows that both cache services contribute to shorter download times. The second is that LanCache outperforms WSUS in terms of network performance and resource usage, and also requires less administration time than WSUS. Given the results, the conclusion is drawn that LanCache is the most suitable caching service in this case.

[EN] Data analysis workflows in High Energy Physics (HEP) read data written in the ROOT columnar format. Such data has traditionally been stored in files that are often read via the network from remote storage facilities, which represents a performance penalty especially for data processing workflows that are I/O bound. To address that issue, this paper presents a new caching mechanism, implemented in the I/O subsystem of ROOT, which is independent of the storage backend used to write the dataset. Notably, it can be used to leverage the speed of high-bandwidth, low-latency object stores. The performance of this caching approach is evaluated by running a real physics analysis on an Intel DAOS cluster, both on a single node and distributed on multiple nodes., This work benefited from the support of the CERN Strategic R&D Programme on Technologies for Future Experiments [1] and from grant PID2020-113656RB-C22 funded by Ministerio de Ciencia e Innovacion MCIN/AEI/10.13039/501100011033. The hardware used to perform the experimental evaluation involving DAOS (HPE Delphi cluster described in Sect. 5.2) was made available thanks to a collaboration agreement with Hewlett-Packard Enterprise (HPE) and Intel. User access to the Virgo cluster at the GSI institute was given for the purpose of running the benchmarks using the Lustre filesystem.

Today, billions of smart devices are interconnected via wireless networks, leading to large volumes of video contents circulating through the bandwidth-limited backhaul. This causes network performance to deteriorate. As a mitigation mechanism, caching of highly popular contents to network edges is deployed. We propose a cooperative and demand-aware caching strategy, which is modelled using the Separable Assignment Problem, to maximize the cache hit ratio. This problem is solved with a recursive enumeration method, where dynamic programming is used to fill each edge. The extensive application-level evaluations show that the proposed strategy outperforms existing caching policies. The work of Luis Alonso was supported in part by the Spanish Government under Project PID2021-125799OA-I00.

With increasing demand for video streaming applications, exploiting cached content at the network edge becomes paramount to prevent congestion in the link between the wireless access network and the content providers. However, it is often challenging to exploit the caches in current client-driven video streaming solutions due to two key reasons. First, even those clients interested in the same content might request different quality levels as a video content is encoded into multiple qualities to match a wide range of network conditions and device capabilities. Second, the clients, who select the quality of the next chunk to request, are unaware of the cached content at the network edge. Hence, it becomes imperative to develop network-side solutions to exploit caching, in particular for the scenarios in which multiple video clients compete for some bottleneck capacity. In this article, we propose EdgeDASH which is a network-side control logic running at a WiFi AP to facilitate the use of cached video content. In particular, an AP can assign a client station to a video quality different than its request, in case the alternative quality provides a better utility. This includes, for example, a function of bits delivered from the cache, video bit rate, and the buffer stalls. We formulate the quality assignment problem as an optimization problem and develop several heuristics with polynomial complexity. Our simulations show that EdgeDASH facilitates significant cache hits and decreases the buffer stalls only by changing the client's request by one quality level, however with some increase in session instability. From our analysis, we conclude that EdgeDASH benefits are more visible especially when the clients with identical interests compete for a bottleneck link's capacity, over the baseline where the clients determine the quality adaptation.

This study is a follow-up to a previous study (a replication of Experiment 1 from Shaw & Clayton, 2013) using a larger sample size, refined methodology, and homogeneous coloured 'noisy' and 'quiet' substrates to reduce the possibility of colour matching.

In this paper, we discuss the security and reliability performance of a communication system, where the base station (BS) transmits signals to multiple users and an unmanned aerial vehicle (UAV) jammer sends friendly jamming signals to protect against a full-duplex active eavesdropper with the aid of caching. We derive the closed-form expressions of outage probability (OP) and intercept probability (IP) for the legitimate users and active eavesdropper, respectively. Aimed at minimizing the sum of the IP and the maximum OP among all users, we jointly optimize the BS transmit power, UAV jammer location and jamming power. An efficient algorithm based on alternating optimization (AO) and successive convex approximation (SCA) methods is designed to solve the optimization problem. Numerical results show that compared to other benchmark strategies, the proposed solution improves the reliability and security performance with the aid of caching and UAV friendly jamming.

Data caching has emerged as a promising technique to handle growing data traffic and backhaul congestion of wireless networks. However, there is a concern regarding how and where to place contents to optimize data access by the users. Data caching can be exploited close to users by deploying cache entities at Small Base Stations (SBSs). In this approach, SBSs cache contents through the core network during off-peak traffic hours. Then, SBSs provide cached contents to content-demanding users during peak traffic hours with low latency. In this paper, we exploit the potential of data caching at the SBS level to minimize data access delay. We propose an intelligence-based data caching mechanism inspired by an artificial intelligence approach known as Reinforcement Learning (RL). Our proposed RL-based data caching mechanism is adaptive to dynamic learning and tracks network states to capture users’ diverse and varying data demands. Our proposed approach optimizes data caching at the SBS level by observing users’ data demands and locations to efficiently utilize the limited cache resources of SBS. Extensive simulations are performed to evaluate the performance of proposed caching mechanism based on various factors such as caching capacity, data library size, etc. The obtained results demonstrate that our proposed caching mechanism achieves 4% performance gain in terms of delay vs. contents, 3.5% performance gain in terms of delay vs. users, 2.6% performance gain in terms of delay vs. cache capacity, 18% performance gain in terms of percentage traffic offloading vs. popularity skewness (γ), and 6% performance gain in terms of backhaul saving vs. cache capacity.

The speedy advancement in wireless communication technologies provides considerable development to enable smart cities with applications such as Intelligent Transport Systems (ITS) and the Internet of Medical Things (IoMT). Blockchain is an emerging technology that provides a secure and distributed data storage mechanism useful for smart city applications. The full nodes in the Blockchain contain a record of all the transactions and data blocks of the Blockchain users. As the number of full nodes is less and the number of Blockchain users is high, there is a huge load on the full nodes for accessing and verifying the data by the Blockchain users. Efficient hardware caching techniques are needed to decrease the data access delay. In this paper, we implement different caching techniques on the Field-Programmable Gate Array (FPGA) Network Interface Card (NIC) and analyze their performance for the key-value store caching in the Blockchain. We design the 2-way and 4-way caching techniques on Block Random-Access Memory (BRAM) and compare them with the conventional direct-mapped caching technique in terms of cache hits and cache misses. The improvements in the hit ratio of the 2-way set-associative cache technique with respect to the direct-mapped cache technique for 10 K, 25 K, and 50 K addresses are 0.8%, 0.77%, and 1.67%, respectively. On the other hand, for the same sets of addresses, the hit rate improvement of the 4-way set-associative cache technique with respect to the direct-mapped cache technique is 0.92%, 2.01%, and 2.4%, respectively. The improvements in hit rate for large data sets show that 2-way and 4-way set-associative cache techniques perform better than the direct-mapped cache technique for caching systems.

The paper is motivated by the urgent need, imposed by the COVID-19 pandemic, for trustworthy access to secure communication systems with the highest achievable availability and minimum latency. In this regard, we focus on an ultra-dense wireless network consisting of Femto Access Points (FAPs) and Unmanned Aerial Vehicles (UAVs), known as caching nodes, where there are more than one possible caching node to handle user's request. To efficiently cope with the dynamic topology of wireless networks and time-varying behavior of ground users, our focus is to develop an efficient connection scheduling framework, where ground users are autonomously trained to determine the optimal caching node, i.e., UAV or FAP. Our aim is to minimize users' access delay by maintaining a trade-off between the energy consumption of UAVs and the occurrence of handovers. To achieve these objectives, we formulate a multi-objective optimization problem and propose the Convolutional Neural Network (CNN) and Q-Network-based Connection Scheduling (CQN-CS) framework. More specifically, to solve the constructed multi-objective connection scheduling problem, a deep Q-Network model is developed as an efficient Reinforcement Learning (RL) approach to train ground users to handle their requests in an optimal and trustworthy fashion within the coupled UAV-based femtocaching network. The effectiveness of the proposed CQN-CS framework is evaluated in terms of the cache-hit ratio, user's access delay, energy consumption of UAVs, handover, lifetime of the network, and cumulative rewards. Based on the simulation results, the proposed CQN-CS framework illustrates significant performance improvements in companion to Q-learning and Deep Q-Network (DQN) schemes across all the aforementioned aspects.

Information centric networking (ICN) shifts the focus of existing internet architecture from host-oriented to content-oriented model by enabling in-network caching and content-based forwarding. These ICN features help to increase network performance by decreasing content discovery delay, content server load, and network congestion. To route a content interest inside network such that content can be fetched with minimal time is a challenging task in ICN. The performance of the ICN routing protocol can be significantly improved if the decisions related to content chunk placement and request forwarding are taken in a cooperative fashion. This paper describes a novel strategy for co-operative caching joint request forwarding in ICN, focusing on decreasing content retrieval latency. To do so, the caching strategy leverages the concept of connected dominating set (CDS) for creating a virtual backbone network to eliminate caching redundancy and reduce content discovery delay. It considers content chunk placement and request forwarding tasks as strongly co-related procedures. It exploits the caching information so that the request is forwarded to a content router (CR) with the maximum likelihood of carrying needed data using the betweenness centrality (BC) of CR. The proposed approach also uses the Markov chain-based model to estimate CS hit likelihood and use it as decisional parameter while forwarding the interest packet. This mechanism helps to fetch the content within the shortest possible time duration. The CCJRF-ICN adopts the Dijkstra’s shortest path routing and works in collaboration with the CDS-driven caching joint forwarding mechanism. The simulation study of CCJRF-ICN is done inside ns-3 based ndnSIM-2.0 simulator with performance measures like content store (CS) hit ratio, content discovery delay, mean hop distance, network load, and network overhead. The simulation outcomes demonstrate that CCJRF-ICN outperforms the state-of-the-art strategies for realistic topologies (GEANT, US-26, Euro-28) and shows improvement up to 5-35% against stated performance measures.

The fifth generation cellular networks are required to provide very fast and reliable communications while dealing with the increase of users traffic. In heterogeneous networks (HetNets) assisted with device-to-device (D2D) communications, traffic can be offloaded to small base stations or to users to make content delivery faster and alleviate the traffic burden from the core network. In this paper, we aim to maximize the probability of successfully delivering files to users, referred to as the successful delivery probability, by jointly optimizing the caching placement and channel allocation in cache-enabled D2D-assisted HetNets. First, an analytical expression of the average content delivery delay is derived. The latter is subsequently used to formulate the joint optimization problem of cache placement and channel allocation. Due to the problem’s non-convexity, a linearization transformation is proposed, which allows to find the optimal solution. However, given the high complexity of the problem, we propose a low-complex heuristic approach for channel allocation and caching. Numerical results illustrate the efficacy of the proposed solutions and compare them to the conventional HetNet. Finally, the impact of several key parameters, e.g., transmit power, caching capacity, and QoS requirements, is investigated, which provides design guidelines for D2D-assisted HetNets.

In this paper, we analyze the impact of caching on the performance of a cache enabled system with heterogeneous traffic where one of the users need to be served with confidential data. In this setup, a wireless helper system always serves a dedicated user and it can also serve a user requesting cached content. A cellular network access point is also available to serve the latter user if it cannot retrieve the requested data from the helper’s cache. The impact of caching and secrecy on throughput and delay performance for each user is then examined when the access point can deploy superposition coding to serve both users simultaneously. Two decoding schemes are considered in this work. The first decoding scheme treats interference from parallel transmissions as noise while the second one utilizes the parallel transmission to apply successive decoding for the intended data. Furthermore, network and cache related factors are identified and their impact on the overall performance of the system are analyzed. In order to find the optimal transmission power allocations, two distinct optimization problems are set in this context comparing the two decoding schemes. This will assist to identify the benefits of the considered decoding schemes for each user satisfying the secrecy requirements of the dedicated user and reducing its impact on the overall performance of the system. Funding: Swedish Foundation for Strategic Research (SSF)Swedish Foundation for Strategic Research; Swedish Research Council (VR)Swedish Research Council; Center for Industrial Information Technology (CENIIT); Excellence Center at Linkoping-Lund in Information Technology (ELLIIT); Indo-Sweden Project, Department of Science and Technology, India

Offloading data traffic from Infrastructure-to-Device (I2D) to Device-to-Device (D2D) communications is a powerful tool for reducing congestion, energy consumption, and spectrum usage of mobile cellular networks. Prior network-level studies on D2D data offloading focus on high level performance metrics as the offloading efficiency, and take into account the radio propagation aspects by using simplistic wireless channel models. In this work, we consider a D2D data offloading protocol tailored to highly dynamic scenarios as vehicular environments, and evaluate its performance focusing on physical layer aspects, like energy consumption and spectral efficiency. In doing this, we take into account more realistic models of the wireless channel, with respect to the simplistic ones generally used in the previous studies. Our objective is twofold: first, to quantify the performance gain of the considered D2D offloading protocol with respect to a classic cellular network, based on I2D communications, in terms of energy consumption and spectral efficiency. Second, to show that using simplistic channel models may prevent to accurately evaluate the performance gain. Additionally, the use of more elaborated models allows to obtain insightful information on relevant system-level parameters settings, which would not be possible to obtain by using simple models. The considered channel models range from widely used models based on deterministic path loss formulas, to more accurate ones, which include effects like multipath fading and the associated frequency selectivity of wideband channels. These models have been proposed and validated, in the recent years, through large-scale measurements campaigns. Our results show that the considered protocol is able to achieve a reduction in the energy consumption of up to 35%, and an increase in the system spectral efficiency of 50%, with respect to the benchmark cellular system. The use of different channel models in evaluating these metrics may result, in the worst case, in a sixfold underestimation of the achieved improvement.

In the internet of vehicles (IoVs), vehicle users should provide location information continuously when they want to acquire continuous location-based services (LBS), which may disclose the vehicle trajectory privacy. To solve the vehicle trajectory privacy leakage problem in the continuous LBS, we propose a vehicle trajectory privacy preservation method based on caching and dummy locations, abbreviated as TPPCD, in IoVs. In the proposed method, when a vehicle user wants to acquire a continuous LBS, the dummy locations-based location privacy preservation method under road constraint is used. Moreover, the cache is deployed at the roadside unit (RSU) to reduce the information interaction between vehicle users covered by the RSU and the LBS server. Two cache update mechanisms, the active cache update mechanism based on data popularity and the passive cache update mechanism based on dummy locations, are designed to protect location privacy and improve the cache hit rate. The performance analysis and simulation results show that the proposed vehicle trajectory privacy preservation method can resist the long-term statistical attack (LSA) and location correlation attack (LCA) from inferring the vehicle trajectory at the LBS server and protect vehicle trajectory privacy effectively. In addition, the proposed cache update mechanisms achieve a high cache hit rate.

By enabling name-based routing and ubiquitous in-network caching, Named Data Networking (NDN) is a promising network architecture for sixth generation (6G) edge network infrastructures. However, the performance of content retrieval largely depends on the selected caching strategy, which is implemented in a distributed fashion by each NDN node. Previous research showed the effectiveness of caching decisions based on content popularity and network topology information. This paper presents a new distributed caching strategy for NDN edge networks based on a metric called popularity-aware closeness (PaC), which measures the proximity of the potential cacher to the majority of requesters of a certain content. After identifying the most popular contents, the strategy caches them in the available edge nodes that guarantee the higher PaC. Achieved simulation results show that the proposed strategy outperforms other benchmark schemes, in terms of reduced content retrieval delay and exchanged data traffic.

Understanding of avian incubation is primarily gained from species nesting during spring and summer, under conditions markedly different from those prevailing in late winter. Using nest temperature loggers, we explored how ambient temperature and rainfall influenced off-bout frequency and off-bout duration and how incubation metrics influenced nest success of female Canada jays, a resident of boreal and subalpine forests that incubates eggs under sub-zero temperatures. Females spent 95.3 % ± 0.1 of their time incubating, and variation in attentiveness was driven primarily by variation in off-bout frequency than duration. Females took longer, less frequent off-bouts during warmer days and females that experienced high rainfall took shorter off-bouts. Nest success was negatively related to off-bout frequency but not duration. Our results suggest that attentiveness of females is driven by, and adjusted to, adverse conditions of late winter whereas low off-bout frequencies are driven by the need to avoid attracting nest predators.

The most common problem on routing networks is to compute the shortest paths from a source vertex to a set of target vertices. A variation of it, with applications for recommender systems, asks to merely rank the target vertices with respect to the shortest distances from the source. A classic solution is Dijkstra’s algorithm; however, it is too slow for large but meaningful applications. A setting where the target vertices are fixed but the source vertex is only known at query time allows for preprocessing. Following the line of research on preprocessing the routing network to speed up the computation of shortest paths, we study in this article a novel approach tackling the problem of ranking the static set of target vertices on the routing network with regard to the distance from a given source vertex to these target vertices by leveraging preprocessing. Our approach allows us to generate a partial solution by pre-computing the distances between all targets so that a shortest-path algorithm does not have to determine the shortest path from the source to every target in general. Our proposal can be adopted for both static and time-dependent networks, and it can be used in conjunction with a general shortest-path algorithm. We can experimentally observe significant speed-ups when using our proposed techniques.

The paucity of observations of wild polar bears (Ursus maritimus) caching of food (including hoarding, i.e., burying and remaining with a kill for up to a few days) has led to the conclusion that such behavior does not occur or is negligible in this species. We document 19 observations of short-term hoarding by polar bears between 1973 and 2018 in Svalbard, Greenland, and Canada. Short-term hoarding appears to be influenced by size of the kill and its remaining energetic value after the first meal has been consumed. Fat and meat from smaller seals, such as pup or yearling ringed seals (Pusa hispida), are largely devoured immediately, leaving little to hoard. Carcasses of adult ringed seals, harp seals (Pagophilus groenlandicus), and bearded seals (Erignathus barbatus) may be covered with snow to reduce the chance of kleptoparasitism by another bear or other scavengers visually detecting a dark spot on the ice, while the hoarding bear lies nearby. Hoarding of other species, such as beluga (Delphinapterus leucas) (calves or parts) or other polar bears, appears opportunistic. We review differences in caching, including short-term hoarding behavior between polar bears and brown bears (U. arctos), and hypothesize about factors that may have influenced their evolution.

Sharding is a method for allocating data items to nodes of a distributed caching or storage system based on the result of a hash function computed on the item’s identifier. It is ubiquitously used in key-value stores, CDNs and many other applications. Despite considerable work that has focused on the design and implementation of such systems, there is limited understanding of their performance in realistic operational conditions from a theoretical standpoint. In this paper we fill this gap by providing a thorough modeling of sharded caching systems, focusing particularly on load balancing and caching performance aspects. Our analysis provides important insights that can be applied to optimize the design and configuration of sharded caching systems.

With the development of emerging applications such as augmented reality, more and more computing tasks are sensitive to delay. Caching popular task computation results on the mobile edge computing (MEC) server is an effective solution to meet the latency requirements. When multiple users request the same task, if the computation result is cached on the MEC server, it will return the computation result directly to the user to reduce the delay for repeated computation. In this paper, we use the caching to assist the calculation. Non-orthogonal multiple access (NOMA) is used to further reduce the delay for computation offloading. The optimization problem is formulated as how to make caching and offloading decision to minimize the delay of whole system. In the case of unknown popularity, we use Gated Recurrent Unit (GRU) algorithm to predict the task popularity in time-varying system, and place the computing results of tasks with high popularity on the corresponding server. Based on the predicted popularity, a multi-agent Deep-Q-network (MADQN) algorithm is used to solve the caching and offloading problem. The simulation results show that the prediction error of GRU algorithm can be reduced by increasing the learning rate. Meanwhile, the proposed MADQN can effectively reduce the delay compared with other methods.

Fog radio access networks (F-RANs) can effectively alleviate fronthaul loads and reduce content transmission delay by migrating cloud services to the network edge. This paper addresses a cooperative caching scenario in F-RAN, where each mobile user can acquire the requested contents from any one of its associated fog-computing-based access points (F-APs). However, caching disparate contents in different F-APs will lead to different content delivery delays, since mobile users suffer from diverse channel fadings and interferences when they download contents from different F-APs. Considering limited caching storage in each F-AP, diverse user preferences, unpredictable user mobility and time-varying channel states, an average transmission delay minimization problem is formulated. With the aid of dueling deep-Q-network framework, a delay-aware cache update policy is proposed for mobile users in F-RAN. The proposed cache update policy will decide to replace the stored contents in F-APs with the proper contents at each time slot. Compared with first in first out, least recently used and least frequently used caching policies, simulation experiments are performed to evaluate the performance of the proposed algorithm. Simulation results illustrate that the proposed caching policy yields better average hit ratio and lower average transmission delay than other traditional caching policies.

Various advanced and mission-critical applications are enabled by the emerging technologies in fifth-generation (5G) mobile communication systems. To ensure improved quality of experience (QoE) of users, 5G and beyond networks require ultra-reliable low-latency communications (URLLC). The successful realization of the URLLC entails the advent of new technological concepts. Therefore, this article presents an overview of the enabling techniques for the URLLC. Classification of the enabling techniques is done and an extensive review of the literature is presented to identify the state-of-the-art techniques, limitations, and the potential approaches for alleviating the limitations. It is observed that artificial intelligence (AI)-enabled edge computing and caching solutions are widely explored as promising techniques to effectively guarantee low latency and reliable content acquisition while reducing redundant network traffic and improving the QoE. Therefore, we present a classification of the AI-enabled edge caching solutions and discuss various mechanisms of the caching agents. In particular, we investigate the use of deep learning (DL), deep reinforcement learning (DRL), and federated learning (FL) algorithms. Subsequently, we analyze the performance of the state-of-the-art edge caching schemes and demonstrate the performance gains of FL frameworks over conventional centralized and decentralized DL and DRL frameworks. We confirm that FL edge caching is a viable mechanism in 5G and beyond networks. On the other hand, it is shown that the IEEE 802.1 time sensitive networking and the emerging IETF deterministic networking standards present effective mechanisms when deterministic networks with bounded ultra-low latency are considered. Finally, we present the open issues and opportunities for further research.

Data communication in the present Internet paradigm is dependent on fixed locations that disseminate similar data several times. As a result, the number of problems has been generated in which location dependency is the most crucial for communication. Therefore, Named Data Networking (NDN) is a new network architecture that revolutionized the handling gigantic amount of data generated from diverse locations. The NDN offers in-network cache which is the most beneficial feature to reduce the difficulties of location-based Internet paradigms. Moreover, it mitigates network congestion and provides a short stretch path in the data downloading procedure. The current study explores a new comparative analysis of popularity-based cache management strategies for NDN to find the optimal caching scheme to enhance the overall network performance. Therefore, the content popularity-based caching strategies are comparatively and extensively studied in an NDN-based simulation environment in terms of most significant metrics such as hit ratio, content diversity ratio, content redundancy, and stretch ratio. In this analysis, the Compound Popular Content Caching Strategy (CPCCS) has performed better in terms to enhance the overall NDN-based caching performance. Therefore, it is suggested that the CPCCS will perform better to achieve enhanced performance in emerging environments such as, Internet of Things (IoT), Fog computing, Edge computing, 5G, and Software Defined Network (SDN).

Smart cities are driven by huge amount of data collected from sensors deployed across the city. Sensors typically form a multi-hop network with a base station (BS ) in order to send their data to the command and control center. However, sparse deployment of sensors can leave subsets of the network partitioned from the rest of the network. In such a case, isolated partitions cannot forward their data to the BS . Consequently, network coverage and data fidelity decline. A possible solution to link partitions and provide connectivity is employing mobile data collectors (MDCs). A smart vehicle supporting wireless communication can act as an MDC and carry data between sensors and the BS . Using a single MDC extends the average tour length. To minimize the maximum tour length, multiple MDCs can be employed. To identify sensors to be visited by each MDC, this paper clusters partitions as many as the number of MDCs and assigns an MDC for each cluster. Then two different cooperative data collection schemes are considered based on the availability of inter-MDC data exchange. If MDCs collaborate in data delivery, they meet at certain meeting points for data exchange. Such a cooperation avoids the requirement of visiting the BS for some MDCs and reduces tour lengths. On the other hand, MDCs closer to the BS can experience data loss due to buffer overflow given the higher volume of the accumulated data. Presented approaches are evaluated in terms of maximum tour length, data latency, and data loss. The smart city application is simulated with deployment of sensors on certain amenity types. Geographic data is obtained from a volunteered geographic information system and MDC mobility is restricted with the road network. Obtained results indicate that MDC cooperation decreases maximum tour length at the expense of increased rate of data loss and data latency.

In recent years, the ever-increasing demand for networking resources and energy, fueled by the unprecedented upsurge in Internet traffic, has been a cause for concern for many service providers. Content caching, which serves user requests locally, is deemed to be an enabling technology in addressing the challenges offered by the phenomenal growth in Internet traffic. Conventionally, content caching is considered as a viable solution to alleviate the backhaul pressure. However, recently, many studies have reported energy cost reductions contributed by content caching in cache-equipped networks. The hypothesis is that caching shortens content delivery distance and eventually achieves significant reduction in transmission energy consumption. This has motivated us to conduct this study and in this article, a comprehensive survey of the state-of-the-art green caching techniques is provided. This review paper extensively discusses contributions of the existing studies on green caching. In addition, the study explores different cache-equipped network types, solution methods, and application scenarios. We categorically present that the optimal selection of the caching nodes, smart resource management, popular content selection, and renewable energy integration can substantially improve energy efficiency of the cache-equipped systems. In addition, based on the comprehensive analysis, we also highlight some potential research ideas relevant to green content caching.

By caching the most popular content into mobile devices, users can retrieve content directly from nearby devices through Device to Device (D2D) communications, which can significantly reduce backhaul traffic and improve network performance. Most existing D2D cache placement strategies are proactive approaches, which cannot deal with the problem of timely cache updating. In this paper, we propose a network coding-based socially-aware D2D caching strategy, which takes geographical proximity and the social relationships of users into consideration. First, a physical D2D network with high communication reliability is built according to the geographical proximity, composed of devices with high probability of communicating to each other through stable D2D communications. According to the social relationship between users within the physical D2D network, we partition the devices into communities and rank the devices within same community by their influence degree. Within a community, each caching decision is made independently according to the user-content contribution degree. To calculate the contribution degree, the impact between devices on requesting same content is modeled using an Indian Buffet Process. Devices cache coded blocks instead of the whole content to improve caching efficiency. Simulation results show that the proposed strategy achieves higher cache hit and sum rates compared to other schemes.

Autoencoders implemented by artificial neural networks (ANNs) are utilized to learn the latent space representation of data in an unsupervised manner, and they have been widely used in recommender systems. For instance, several collaborative denoising autoencoder (CDAE) models have shown that their performance gains outperform that of the collaborative filtering based (CF-based) models. In this work, a near-optimal Top- $K$ forecasting solution is proposed for our advanced autoencoder recommender systems. We propose a method which utilizes CDAE model in predicting the Top- $K$ popular videos in an upcoming time period. In order to improve the prediction accuracy, we also propose an autoencoder based recommendation algorithm with the help of $K$ -means clustering that upgrades the performance of the original autoencoder model. The experimental results show that our method increases significantly the Average Precision (AP) and Recall values by nearly 30%. We then further utilize our proposed autoencoder model with clustering in predicting Top- $K$ popular videos. The applications of predicting Top- $K$ popular videos can be used in the video delivery for the Mobile Edge Computing (MEC) environment to avoid bottleneck in the constricted capacity of backhaul link. Namely, the performance gain will be upgraded if our proposed method precisely predicts and caches the Top- $K$ popular videos in advance with the help of a better forecasting model.

The ever-increasing number of diverse and computation-intensive Internet of things (IoT) applications is bringing phenomenal growth in global Internet traffic. Mobile devices with limited resource capacity (i.e., computation and storage resources) and battery lifetime are experiencing technical challenges to satisfy the task requirements. Mobile edge computing (MEC) integrated with IoT applications offloads computation-intensive tasks to the MEC servers at the network edge. This technique shows remarkable potential in reducing energy consumption and delay. Furthermore, caching popular task input data at the edge servers reduces duplicate content transmission, which eventually saves associated energy and time. However, the offloaded tasks are exposed to multiple users and vulnerable to malicious attacks and eavesdropping. Therefore, the assignment of security services to the offloaded tasks is a major requirement to ensure confidentiality and privacy. In this article, we propose a green and secure MEC technique combining caching, cooperative task offloading, and security service assignment for IoT networks. The study not only investigates the synergy between energy and security issues, but also offloads IoT tasks to the edge servers without violating delay requirements. A resource-constrained optimization model is formulated, which minimizes the overall cost combining energy consumption and probable security-breach cost. We also develop a two-stage heuristic algorithm and find an acceptable solution in polynomial time. Simulation results prove that the proposed technique achieves notable improvement over other existing strategies.

The IoT devices produce a massive amount of content compared to the host-centric (IP-based) network. The IP-based network has no strengthened capabilities to handle the growing traffic of the IoT network. The existing research is focusing on proposing Information Centric Network (ICN) framework with potential replacement of IP-based network. Using the caching technique in ICN based IoT significantly strengthen the resource-constrained IoT devices, by reducing the access time with optimized energy consumption. Thus, this paper proposes a caching scheme for IoT contents, known as Central Control Caching (CCC) Scheme. This scheme positions a central entity between different autonomous systems that manage caching and updates the content, by keeping information in a table regarding different contents present on different autonomous systems. This information is utilized when cached contents are exchanged among different autonomous systems. Furthermore, when contents becomes stale, the same information is used to identify the autonomous systems with a copy of the outdated content. The simulation results show that the proposed scheme has outperformed in terms of reduced energy consumption, response latency, and cache-hit compared to state-of-the-art works.

Avoiding problems such as packet loss in the transport network is crucial for mobile network providers to offer high-quality services reliably and without interruptions. In this paper, we propose and compare three different transmission strategies, namely Caching, Network Coding (NC) and Repetition enabled transmission in the User Plane (UP) of mobile backhaul for network operators to prevent such performance degradation. In the proposed NC-enabled transmission method, NC provides robustness to transport network failures such that no further retransmission is required by the User Equipment (UE) compared to conventional approaches where UE applications perform retransmissions. The proposed scheme requires only a minor modification to the packet structure of the UP protocol, which requires a small development effort and no new extensions to the current UE standard features. We also discuss their placement in the O-RAN protocol stack and in the core network, and propose a new architecture that can utilize caching, repetition and NC features in the mobile network architecture. Our simulation results show that an exact 1% packet loss ratio in the backhaul link results in an additional total transmission time of 59.44% compared to the normal GPRS Tunneling Protocol-User Plane (GTP-U) transmission. Applying NC at a rate of 1% and 2% reduces this value to 52.99% and 56.26%, respectively, which is also better than the total transmission time of some previously studied dynamic replication schemes while keeping the bandwidth utilization at low rates. On the cache side, a reduction in latency of about 20% can be achieved with a cache size of 100MB. At the end of the paper, we summarize some of the benefits and limitations of using these three strategies in UP of mobile backhaul networks.

Triangle count and local clustering coefficient are two core metrics for graph analysis. They find broad application in analyses such as community detection and link recommendation. Current state-of-the-art solutions suffer from synchronization overheads or expensive pre-computations needed to distribute the graph, achieving limited scaling capabilities. We propose a fully asynchronous implementation for triangle counting and local clustering coefficient based on 1D partitioning, using remote memory accesses for transferring data and avoid synchronization. Additionally, we show how these algorithms present data reuse on remote memory accesses and how the overall communication time can be improved by caching these accesses. Finally, we extend CLaMPI, a software-layer caching system for MPI RMA, to include application-specific scores for cached entries and influence the eviction procedure to improve caching efficiency. Our results show improvements on shared memory, and we achieve 14x speedup from 4 to 64 nodes for the LiveJournal1 graph on distributed memory. Moreover, we demonstrate how caching remote accesses reduces total running time by up to 73% with respect to a non-cached version. Finally, we compare our implementation to TriC, the 2020 graph champion paper, and achieve up to 100x faster results for scale-free graphs., Comment: 11 pages, 10 figures, to be published at IPDPS'22

A file caching setup to access ATLAS data was deployed for the NEMO HPC cluster in Freiburg, using an XRootD proxy server running in forwarding mode. This setup allows running HEP workflows without the need to operate a large local data storage. Several performance tests were carried out to measure any potential overhead caused by the caching setup with respect to direct, non-cached data access., publishedVersion

Caching is among the most well-studied topics in algorithm design, in part because it is such a fundamental component of many computer systems. Much of traditional caching research studies cache management for a single-user or single-processor environment. In this paper, we propose two related generalizations of the classical caching problem that capture issues that arise in a multi-user or multi-processor environment. In the caching with reserves problem, a caching algorithm is required to maintain at least k_i pages belonging to user i in the cache at any time, for some given reserve capacities k_i. In the public-private caching problem, the cache of total size k is partitioned into subcaches, a private cache of size k_i for each user i and a shared public cache usable by any user. In both of these models, as in the classical caching framework, the objective of the algorithm is to dynamically maintain the cache so as to minimize the total number of cache misses. We show that caching with reserves and public-private caching models are equivalent up to constant factors, and thus focus on the former. Unlike classical caching, both of these models turn out to be NP-hard even in the offline setting, where the page sequence is known in advance. For the offline setting, we design a 2-approximation algorithm, whose analysis carefully keeps track of a potential function to bound the cost. In the online setting, we first design an O(ln k)-competitive fractional algorithm using the primal-dual framework, and then show how to convert it online to a randomized integral algorithm with the same guarantee., LIPIcs, Vol. 245, Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques (APPROX/RANDOM 2022), pages 52:1-52:16

Developing applications on the edge is a difficulttask, as issues arise when multiple nodes have to operate inunison. Common challenges comprise correctness, performance,and forensics analysis after a system crash. The challenges areaggravated when both edge and fog devices get in the picture,since the limited resources of some of them exacerbate anypossible issue, and create the need for a robust mechanism tocollect log data to make sense of the system’s behavior. Theenvisioned monitoring layer should collect log data regardingperformance and errors, protect the data until they are delivered to an external data processing facility, and be easy tointegrate into the deployed system. This paper proposes C-Mon,a resilience-focused monitoring framework for service-orientedapplications that tests timing constraints against live log data. Themonitoring framework is integrated with OpenAPI Generator,the mainstream ReST code generator we used to generate clientand server interfaces and monitoring mechanisms. Log data arecached locally on the disk of the fog/edge devices and transferredto the monitoring server only when enough CPU cycles andnetwork bandwidth are available, at the same time enablingforensics analysis.

Small to medium-scale data science experiments often rely on research software developed ad-hoc by individual scientists or small teams. Often there is no time to make the research software fast, reusable, and open access. The consequence is twofold. First, subsequent researchers must spend significant work hours building upon the proposed hypotheses or experimental framework. In the worst case, others cannot reproduce the experiment and reuse the findings for subsequent research. Second, suppose the ad-hoc research software fails during often long-running computationally expensive experiments. In that case, the overall effort to iteratively improve the software and rerun the experiments creates significant time pressure on the researchers. We suggest making caching an integral part of the research software development process, even before the first line of code is written. This article outlines caching recommendations for developing research software in data science projects. Our recommendations provide a perspective to circumvent common problems such as propriety dependence, speed, etc. At the same time, caching contributes to the reproducibility of experiments in the open science workflow. Concerning the four guiding principles, i.e., Findability, Accessibility, Interoperability, and Reusability (FAIR), we foresee that including the proposed recommendation in a research software development will make the data related to that software FAIRer for both machines and humans. We exhibit the usefulness of some of the proposed recommendations on our recently completed research software project in mathematical information retrieval., Comment: 8 pages, 1 table

International audience; Similarity caching allows requests for an item i to be served by a similar item i ′. Applications include recommendation systems, multimedia retrieval, and machine learning. Recently, many similarity caching policies have been proposed, but still we do not know how to compute the hit rate even for simple policies, like SIM-LRU and RND-LRU that are straightforward modifications of classic caching algorithms. This paper proposes the first algorithm to compute the hit rate of similarity caching policies under the independent reference model for the request process. In particular, we show how to extend the popular timeto-live approximation in classic caching to similarity caching. The algorithm is evaluated on both synthetic and real world traces.

The hippocampal formation (HF) processes spatial memories for cache locations in food-hoarding birds. Hoarding is a seasonal behavior, and seasonal changes in the HF have been described in some studies, but not in others. One potential reason is that birds may have been sampled during the seasonal hoarding peak in some studies, but not in others. In this study, we investigate the seasonal changes in hoarding and HF in willow tits (Poecile montanus). We compare this to seasonal changes in HF in a closely related non-hoarding bird, the great tit (Parus major). Willow tits near Oulu, Finland, show a seasonal hoarding peak in September and both HF volume and neuron number show a similar peak. HF neuronal density also increases in September, but then remains the same throughout winter. Unexpectedly, the great tit HF also changes seasonally, although in a different pattern: the great tit telencephalon increases in volume from July to August and decreases again in November. Great tit HF volume follows suit, but with a delay. Great tit HF neuron number and density also increase from August to September and stay high throughout winter. We hypothesize that seasonal changes in hoarding birds’ HF are driven by food-hoarding experience (e.g., the formation of thousands of memories). The seasonal changes in great tit brains may also be due to experience-dependent plasticity, responding to changes in the social and spatial environment. Large-scale experience-dependent neural plasticity is therefore probably not an adaptation of food-hoarding birds, but a general property of the avian HF and telencephalon.

International audience; Machine learning is revolutionizing the way multimedia information is processed and transmitted to users. After intensive and powerful training, some impressive efficiency/accuracy improvements have been made all over the transmission pipeline. For example, the high model capacity of the learning-based architectures enables us to accurately model the image and video behavior such that tremendous compression gains can be achieved. Similarly, error concealment, streaming strategy or even user perception modeling have widely benefited from the recent learning- oriented developments. However, learning-based algorithms often imply drastic changes to the way data are represented or consumed, meaning that the overall pipeline can be affected even though a subpart of it is optimized. In this paper, we review the recent major advances that have been proposed all across the transmission chain, and we discuss their potential impact and the research challenges that they raise.

International audience; k-Nearest Neighbors aims at efficiently finding items close to a query in a large collection of objects, and it is used in different applications, from image retrieval to recommendation. These applications achieve high throughput combining two different elements: 1) approximate nearest neighbours searches that reduce the complexity at the cost of providing inexact answers and 2) caches that store the most popular items. In this paper we propose to combine the approximate index for the whole catalog with a more precise index for the items stored in the cache. Our experiments on realistic traces show that this approach is doubly advantageous as it 1) improves the quality of the final answer provided to a query, 2) additionally reduces the service latency.

International audience; This paper focuses on similarity caching systems, in which a user request for an object o that is not in the cache can be (partially) satisfied by a similar stored object o , at the cost of a loss of user utility. Similarity caching systems can be effectively employed in several application areas, like multimedia retrieval, recommender systems, genome study, and machine learning training/serving. However, despite their relevance, the behavior of such systems is far from being well understood. In this paper, we provide a first comprehensive analysis of similarity caching in the offline, adversarial, and stochastic settings. We show that similarity caching raises significant new challenges, for which we propose the first dynamic policies with some optimality guarantees. We evaluate the performance of our schemes under both synthetic and real request traces.

Enabling global Internet access is challenging for cellular-based Internet of Things (IoT) due to the limited range of terrestrial network services. One viable solution is to deploy IoT over satellite systems for coverage extension. However, operating a hybrid satellite-terrestrial network (STN) might incur high satellite bandwidth consumption and excessive service latency. Aiming to reduce the content delivery latency from the Internet-connected gateway to the users, this work proposes a wireless two-tier cache-enabled model with full-duplex transmissions where content caches are deployed at the satellite and ground station. A closed-form solution for the successful delivery probability (SDP) of the files is derived considering the requested content distributions and channel statistics. Then, the SDP performance under common caching policies can be conveniently evaluated. The results are also used to optimize cache placement under caching capacity constraints. Numerical results demonstrate the performance improvements of the proposed system over those of single-tier cache-aided and half-duplex transmission systems.