Cache Management Research Articles

A Reliable Transport Scheme for Human Opportunistic Networks

The human opportunistic networks (ONs) formed by hand-held smart devices can facilitate peer-to-peer communication when humans are on the move, despite contemporaneous end-to-end paths rarely existing. In some scenarios, where network resources, especially power, are scarce and the traffic is large, the data delivery is prone to poor user experience and unbounded delay, although the best effort mechanism “store-carry-forward” is used. To cope with that, most transport/routing schemes obtain an acceptable latency at the cost of energy resources. In real-life human ONs, however, excessive energy consumption will trigger passive participation of the relays in message forwarding, so as to save their limited energy resource. Thus, the reliability of these schemes may get worse in real-life human ONs. In this paper, a reliable transport scheme is developed by making an optimal trade-off between the file round-trip delay and the energy consumption of relays. We make use of acknowledgements and coding at the source to enable successful file delivery. When setting up the network model, the cache management rule referred as “full-duplex” strategy is formulated, and then a mathematical model is established to analyze the proposed scheme. This model describes the evolution of packet dissemination and allows both the mean file round-trip delay and the energy consumption up to the reception of the last acknowledgement by the source to be expressed. Subsequently, a new function based on these two metrics is proposed to reflect the number of files that can be delivered under time and energy constraints. Through optimization procedure, the configurations that can maximize the function are obtained; thus, the optimization of these two metrics is achieved. Numerous simulations and comparisons are conducted and the results verify the accuracy of the analytical model. Comparison results show that with limited energy and passive relays, the proposed transport scheme can significantly reduce the energy consumption of file delivery, which obviously alleviates the selfish behavior of nodes. Therefore, the reliability and stability of the communication service in human ONs are enhanced.

Live MPEG-DASH video streaming cache management with cognitive mobile edge computing

Video streaming is expected to account for up to 82% of network traffic by 2021 according to the forecast of CISCO's Visual Networking Index. Dynamic Adaptive Streaming over HTTP (DASH) is the de facto protocol for delivering video streaming services over the Internet. As the cellular network entering the 5G era, more and more video streaming will be live video streaming delivered to mobile phones. However, due to a large amount of video traffic, several techniques are required to improve the user Quality of Experience (QoE). In this work, we first propose a network architecture design for delivering live video streaming over the cellular core network with cognitive Mobile Edge Computing (MEC) servers. We then focused on the optimal cache management by considering several issues, include QoE, cache size, backhaul bandwidth, pre-cache mechanism, and user mobility. Finally, we show a prototype of the proposed MEC-assisted live video streaming system. Our simulation results show the performance improvement of the proposed cache management schemes in terms of QoE-based system utility. Our prototype shows the significant latency reduction in receiving video streams with MEC pre-cache mechanism.

UAV-Assisted Content Delivery in Intelligent Transportation Systems-Joint Trajectory Planning and Cache Management

Unmanned Aerial Vehicles (UAVs) are gaining growing interests due to the paramount roles they play, particularly these days, in enabling new services that help modernize our transportation, supply chain, search and rescue, among others. They are capable of positively influencing wireless systems through enabling and fostering emerging technologies such as autonomous driving, vertical industries, virtual reality and so many others. The Internet of Vehicles is a prime sector benefiting from the services offered by future cellular systems in general and UAVs in particular, and this paper considers the problem of content delivery to vehicles on road segments with either overloaded or no available communication infrastructure. Incoming vehicles demand service from a library of contents that is partially cached at the UAV; the content of the library is also assumed to change as new vehicles carrying more popular contents arrive. Each inbound vehicle makes a request and the UAV decides on its best trajectory to provide service while maximizing a certain operational utility. Given the energy limitation at the UAV, we seek an energy efficient solution. Hence, our problem consists of jointly finding caching decisions, UAV trajectory and radio resource allocation which is formulated mathematically as a Mixed Integer Non-Linear Problem (MINLP). However, owing to uncertainties in the environment (e.g., random arrival of vehicles, their requests for contents and their existing contents), it is often hard and impractical to solve using standard optimization techniques. To this end, we formulate our problem as a Markov Decision Process (MDP) and we resort to tools such as Proximal Policy Optimization (PPO), a very promising Reinforcement Learning method, along with a set of crafted algorithms to solve our problem. Finally, we conduct simulation-based experiments to analyze and demonstrate the superiority of our solution approach compared with four counterparts and baseline schemes.

Design of the IBM z15 microprocessor

The latest-generation IBM Z processor provides enhanced performance and compute capacity compared to its IBM z14 predecessor. This article describes some of the major improvements in both process and design including out-of-order load-and-store sequencing, single-instruction multiple-data and floating point enhancements, a new modulo arithmetic engine for accelerating elliptic curve cryptography, a hardware sort accelerator, and a workflow that modernized the development of these features. Outside of the central processing unit (CPU), the cache sizes have increased on all levels, and each processor chip now contains 12 CPUs. System topology changes have been introduced allowing up to five drawers to exist in a fully populated system. The processor cache subsystem includes numerous improvements in the area of fetch, store, and cache management policies aimed at speeding up both traditional data serving workloads and highly virtualized environments alike.

Learning-based dynamic cache management in a cloud

Caches are an important component of modern computing systems given their significant impact on performance. In particular, caches play a key role in the cloud due to the nature of large-scale, data-intensive processing. One of the key challenges for the cloud providers is how to share the caching capacity among tenants, under the circumstance that each often requires a different degree of quality of service (QoS) with respect to data access performance. The invariant is that the individual tenants’ QoS requirements should be satisfied while the cache usage is optimized in a system-wide manner. In this paper, we introduce a learning-based approach for dynamic cache management in a cloud, which is based on the estimation of data access pattern of a tenant and the prediction of cache performance for the access pattern in question. We consider a variety of probability distributions to estimate the data access pattern, and examine a set of learning-based regression techniques to predict the cache hit rate for the access pattern. The predicted cache hit rate is then used to make a decision whether reallocating cache space is needed to meet the QoS requirement for the tenant. Our experimental results with an extensive set of synthetic traces and the YCSB benchmark show that the proposed method consistently optimizes the cache space while satisfying the QoS requirement.

User-Centric Edge Sharing Mechanism in Software-Defined Ultra-Dense Networks

The emerging mobile edge computing (MEC) evolutionarily extends the cloud services to the network edge. In order to efficiently coordinate distributed edge resources, software defined networking (SDN) at the network edge has been explored to realize the integrated management of communication, computation, and cache (3C) resources. However, many research efforts, in software-defined edge networks, are mainly devoted to 1C or 2C resource sharing. Motivated by high service performance and user demands, we propose a user-centric edge resource sharing model for software-defined ultra-dense network (SD-UDN) where multiple MEC servers around small base stations (SBSs) can share their 3C resources through OpenFlow-enabled switches. In particular, the service models of MEC servers and users are formulated to optimize the service process by minimizing the service delay, which is NP-hard. To address this NP-hard issue, a service association model is constructed based on design structure matrix (DSM), and a simulated annealing algorithm is employed to further optimize the service association model for reducing time complexity and offering a nearoptimal solution. Compared with traditional 1C or 2C resource sharing, the proposed edge resource sharing model can guarantee lower service delay for users.

Intelligent Resource Scheduling with Neutrosophic Knowledge and Optimized Cache Management Using Cuckoo Search Method in Cloud Computing

Intelligent Resource Scheduling with Neutrosophic Knowledge and Optimized Cache Management Using Cuckoo Search Method in Cloud Computing

Addressing variability in reuse prediction for last-level caches

Last-Level Cache (LLC) represents the bulk of a modern CPU processor's transistor budget and is essential for application performance as LLC enables fast access to data in contrast to much slower main memory. However, applications with large working set size often exhibit streaming and/or thrashing access patterns at LLC. As a result, a large fraction of the LLC capacity is occupied by dead blocks that will not be referenced again, leading to inefficient utilization of the LLC capacity. To improve cache efficiency, the state-of-the-art cache management techniques employ prediction mechanisms that learn from the past access patterns with an aim to accurately identify as many dead blocks as possible. Once identified, dead blocks are evicted from LLC to make space for potentially high reuse cache blocks. In this thesis, we identify variability in the reuse behavior of cache blocks as the key limiting factor in maximizing cache efficiency for state-of-the-art predictive techniques. Variability in reuse prediction is inevitable due to numerous factors that are outside the control of LLC. The sources of variability include control-flow variation, speculative execution and contention from cores sharing the cache, among others. Variability in reuse prediction challenges existing techniques in reliably identifying the end of a block's useful lifetime, thus causing lower prediction accuracy, coverage, or both. To address this challenge, this thesis aims to design robust cache management mechanisms and policies for LLC in the face of variability in reuse prediction to minimize cache misses, while keeping the cost and complexity of the hardware implementation low. To that end, we propose two cache management techniques, one domain-agnostic and one domain-specialized, to improve cache efficiency by addressing variability in reuse prediction.

BloCkEd: Blockchain-Based Secure Data Processing Framework in Edge Envisioned V2X Environment

There has been an increasing trend of moving computing activities closer to the edge of the network, particularly in smart city applications (e.g., vehicle-to-everything – V2X). Such a paradigm allows the end user’s requests to be handled/processed by nodes at the edge of the network; thus, reducing latency, and preserving privacy of user data/activities. However, there are a number of challenges in such an edge computing ecosystem. Examples include (1) potential inappropriate utilization of resources at the edge nodes, (2) operational challenges in cache management and data integrity due to data migration between edge nodes, particularly when dealing with vehicular mobility in a V2X application, and (3) high energy consumption due to continuous link breakage and subsequent reestablishment of link(s). Therefore in this paper, we design a blockchain-based secure data processing framework for an edge envisioned V2X environment (hereafter referred to as BloCkEd ). Specifically, a multi-layered edge-enabled V2X system model for BloCkEd is presented, which includes the formulation of a multi-objective optimization problem. In addition, BloCkEd comprises an optimal container-based data processing scheme, and a blockchain-based data integrity management scheme, designed to minimize link breakage and reducing latency. Using Chandigarh City, India, as the scenario, we implement and evaluate the proposed approach in terms of its latency, energy consumption, and service level agreement compliance.

Page Reusability-Based Cache Partitioning for Multi-Core Systems

Most modern multi-core processors provide a shared last level cache (LLC) where data from all cores are placed to improve performance. However, this opens a new challenge for cache management, owing to cache pollution. With cache pollution, data with weak temporal locality can evict other data with strong temporal locality when both are mapped into the same cache set. In this article, we propose page reusability-based cache partitioning (PRCP) for multi-core systems to maximize cache utilization by minimizing cache pollution. To achieve this, PRCP divides pages into two groups: (1) highly-reused pages and (2) lowly-reused pages. The reusability of each page is collected online via periodic page table scans. PRCP then dynamically partitions the shared cache into two corresponding areas using page coloring technique. We have implemented PRCP in Linux kernel and evaluated it using SPEC CPU2006 benchmarks. The results show that our scheme can achieve comparable performance to the optimal offline MRC-guided process-based cache partitioning scheme without a priori knowledge of workloads.

A collaborative caching strategy for content-centric enabled wireless sensor networks

The Content-Centric Networking (CCN) is an efficient traffic handling technology by accessing content with its name instead of its physical location and achieving in-network caching. Indeed, CCN caching ensures high content availability, network traffic reduction, and low retrieval latency which reduces congestion and improves end-to-end delay. Moreover, it could greatly improve the efficiency of content delivery in Wireless Sensor Networks (WSNs). In this article, we propose to exploit this feature to enable CCN in WSN environments. The CCN architecture enables the content caching on each sensor-node in WSN and several research studies have been devoted to the caching management issue in such a context. However, caching the content on all the nodes is not a good strategy in terms of resource utilization. It is, therefore, necessary to determine where to cache and how to handle it in order to optimize the resources while realizing a high-interest satisfaction rate. Thus, our objective is to study existing caching strategies and propose a novel one that takes into account the node centrality and its distance from the source of the content. Through extensive simulations, we examine the performance of our scheme under different configurations and demonstrate how it outperforms the traditional caching strategies such as LCE (Leave Copy Everywhere) and LCD (Leave Copy Down).

Selective bypassing and mapping for heterogeneous applications on GPGPUs

Modern GPGPU supports executing multiple tasks with different run time characteristics and resource utilization. Having an efficient execution and resource management policy has been shown to be a critical performance factor when handling the concurrent execution of tasks with different run time behavior. Previous policies either assign equal resources to disparate tasks or allocate resources based on static or standalone behavior profiling. Treating tasks equally cannot efficiently utilize the system resources, while the standalone profiling ignores the correlated impact when running tasks concurrently and could hint incorrect task behavior. This paper addresses the above drawbacks and proposes a heterogeneity aware Selective Bypassing and Mapping (SBM) to manage both computing and cache resources for multiple tasks in a fine-grain manner. The light-weight run time profiling of SBM properly characterizes the disparate behavior of the concurrently executed multiple tasks, and selectively applies suited cache management and workgroup mapping policies to each task. When compared with the previous coarse-grained policies, SBM can achieve an average of 138% and up to 895% performance enhancement. When compared with the state-of-art fine-grained policy, SBM can achieve an average of 58% and up to 378% performance enhancement.

Enhanced query processing over semantic cache for cloud based relational databases

Semantic cache diminishes the expectancies of data retrieval over distributed system like cloud-based systems, by reusing already extracted data. It improves the performance of the retrieval system with limited bandwidth which is a key requirement in cloud, fog, edge and other mobile computing technologies. Cache management and query processing over semantic cache are two key activities which should be handled carefully. Query dispensation performs a significant part in terms of reusability in already retrieved data efficiently. Competent query dispensation algorithms made the semantic cache more efficient. This paper identifies four issues that cause decreasing the efficiency of query processing algorithm. The solution is provided to the identified issues to improve the efficiency of query processing. Improvement in efficiency is discussed along each of the issues. Efficiency analysis for enhanced algorithm is also provided and compared with state-of-the-art algorithms in the field. The proposed approach significantly improves the efficiency in many ways.

A $$3+\varOmega (1)$$ Lower Bound for Page Migration

We address the page migration problem, one of the most classical online problems. In this problem, we are given online requests from nodes in a network for accessing a single page, i.e., a data set stored in a node, and asked to determine a node for the page to be stored in after each request. Serving a request costs the distance between the request and the page at the point of the request, and migrating the page costs the migration distance multiplied by the page size. The objective is to minimize the total sum of the service and migration costs. This problem is motivated by efficient cache management in multiprocessor systems. In this paper, we prove that no deterministic online page migration algorithm is $$(3+o(1))$$ -competitive, where the o-notation is with respect to the page size. Our lower bound first breaks the barrier of 3 by an additive constant for an arbitrarily large page size and disproves Black and Sleator’s conjecture even in the asymptotic sense.

Modified Floyd Warshall Algorithm for Cache Management in Information Centric Network

Modified Floyd Warshall Algorithm for Cache Management in Information Centric Network

Reuse Distance-based Victim Cache for Effective Utilisation of Hybrid Main Memory System

Hybrid main memories comprising DRAM and Non-volatile memories (NVM) are projected as potential replacements of the traditional DRAM-based memories. However, traditional cache management policies designed for improving the hit rate lack awareness of the comparative latency of read-write for NVM blocks where the write latency is more than the read latency. Therefore, developing cache management techniques that reduce costly write-backs of the NVM blocks, yet maintain a fair hit rate in the cache, is of paramount importance. We propose two techniques based on the use of a small victim cache associated with the last-level cache that helps in retaining on the chip critical DRAM and NVM blocks. Victim cache being a scarce resource, we intend to keep only performance-critical blocks in the victim cache by exploiting the idea of reuse distance. The first technique, Victim Cache Replacement Policy, works on the replacement policy of the victim cache by preferential eviction of DRAM blocks over NVM blocks. However, the second technique, Prioritized Partitioning of victim cache, logically partitions the victim cache, giving a smaller share to the DRAM blocks and a relatively larger share to the NVM blocks. Experimental evaluation on full-system simulator shows significant improvement in system performance and reduction in the number of write-backs to the NVM partition of the main memory compared to the baseline and existing technique. Additionally, NVM reads and DRAM miss rate are also improved, leading to further performance enhancement.

Per-Operation Reusability Based Allocation and Migration Policy for Hybrid Cache

Recently, a hybrid cache consisting of SRAM and STT-RAM has attracted much attention as a future memory by complementing each other with different memory characteristics. Prior works focused on developing data allocation and migration techniques considering write-intensity to reduce write energy at STT-RAM. However, these works often neglect the impact of operation-specific reusability of a cache line. In this paper, we propose an energy-efficient per-operation reusability-based allocation and migration policy (ORAM) with a unified LRU replacement policy. First, to select an adequate memory type for allocation, we propose a cost function based on per-operation reusability – gain from an allocated cache line and loss from an evicted cache line for different memory types – which exploits the temporal locality. Besides, we present a migration policy, victim and target cache line selection scheme, to resolve memory type inconsistency between replacement policy and the allocation policy, with further energy reduction. Experiment results show an average energy reduction in the LLC and the main memory by 12.3 and 21.2 percent, and the improvement of latency and execution time by 21.2 and 8.8 percent, respectively, compared with a baseline hybrid cache management. In addition, the Energy-Delay Product (EDP) is improved by 36.9 percent over the baseline.

An adaptive cache management approach in ICN with pre-filter queues

In Information-Centric Networking (ICN), transmission does not depend on the ends of communication, but on the content itself. In-network cache plays an important role in ICN, it empowers nodes in ICN better mobility. It also shortens serving path, lightens load of server, reduces traffic and time delay, its efficiency seriously affects performance of entire network, and thus cache management in ICN draws much attention of researchers recently. Although there are a lot approaches have been proposed, a cache management scheme with better adaptability and less cost remains to be studied. To this end, we propose State-value-and-Cache-rate-Method (SCMethod) to manage cache resources in ICN, which comprises cache deployment policy and cache replacement policy. In cache level, we add pre-filter queues in front of cache queue to filter out popular content to store in cache. In the perspective of node, according to factors that node’s state and relative location on data forwarding path, we define node state and cache rate to select nodes with maximum of state value or minimum cache rate to cache content, and effectively mitigate data redundancy. In order to increase dynamics and adaptive of system for mobility of nodes, we employ cache hit ratio as feedback to dynamically adjust the number of pre-filter queues in every node. With pre-filter queues, we improve Least Recently Used (LRU) cache replacement method. We conduct extensive experiments in simulator Icarus (Saino et al., 2014) with both tree topology and realistic internet topologies, define four metrics to quantitatively evaluate performance of SCMethod and verify its efficacy of reducing latency and load. Simulation results demonstrate that SCMethod we proposed outperform several classic cache schemes of ICN.

An NDN Cache Management for MEC

To enhance network performance, the named data networking architecture (NDN) caches data-packets in the network nodes on a downstream network path. Then it uses such cached requested data-packets to respond to new request-packets. Hence, a cache management scheme (CMS) is the essential point of NDN. CMS generally considers two main factors. One is a short response time and the other is storage efficiency. To rapidly respond to requests, CMS generally tries to cache data-packets near users as much as possible. To efficiently manage storage, it uses the popularity of the data. That is, proportionally to the popularity of the data, it increases the number of nodes caching data-packets and manages the lifetime of caches. However, few data objects are as popular as many users globally enjoy in the real world. Hence, if the assumptions about content- usage are practically changed, CMS can waste cache storage and not significantly improve network efficiency. We show that many caches have expired and are not used at all. To improve such inefficiency of CMS, this paper propose to simultaneously apply two cache decision factors, the expected frequency of a cache hit and the popularity of data. That is, it proposes to gradually cache transmitted data in nodes in which their expected cache-usage frequency is relatively high. To show the effectiveness of our proposal, we implement LdC (a limited domain cache policy) and evaluate the performance of LdC. The evaluation result shows that it can enhance the cache-storage efficiency by up to 65% compared with existing CMS without degrading the network efficiency.

Miss Penalty Aware Cache Replacement for Hybrid Memory Systems

Current DRAM-based memory systems face the scalability challenges in terms of memory density, energy consumption, and monetary cost. Hybrid memory architectures composed of emerging nonvolatile memory (NVM) and DRAM is a promising approach to large-capacity and energy-efficient main memory. However, hybrid memory systems pose a new challenge to on-chip cache management due to the asymmetrical penalty of memory access to DRAM and NVM in case of cache misses. Cache hit rate is no longer an effective metric for evaluating memory access performance in hybrid memory systems. Current cache replacement policies that aim to improve the cache hit rate are not efficient either. In this article, we take into account the asymmetry of the cache miss penalty on DRAM and NVM, and advocate a more general metric, average memory access time (AMAT), to evaluate the performance of hybrid memories. We propose a miss penalty aware LRU-based cache replacement policy (MALRU) for hybrid memory systems. MALRU is aware of the source (DRAM or NVM) of missing blocks and preserves high-latency NVM blocks as well as low-latency DRAM blocks with good temporal locality in the last level cache. The experimental results show that MALRU can improve system performance by up to 22.8% and 13.1%, compared to LRU and the state-of-the-art hybrid memory aware cache partitioning technique policy, respectively.