Journal of Systems Architecture最新文献

Error correction and erasure codes and steganographic channels use related methods, but are investigated separately. We detail an idea from literature for a steganographic channel in a transmission with error correction code and experimentally investigate it with respect to bandwidth, robustness and detectability. We expand this construction to provide an example of multi-level steganography, i.e., a steganographic channel within a steganographic channel. Furthermore, we investigate the advantages on bandwidth and stealthyness that reversibility of such a steganographic channel brings, together with a new proposal for a covert channel in error-corrected data.

Resistive random access memory (ReRAM) is a promising technology for AI Processing-in-Memory (PIM) hardware because of its compatibility with CMOS, small footprint, and ability to complete matrix–vector multiplication workloads inside the memory device itself. However, redundant computations are brought on by duplicate weights and inputs when an MVM has to be split into smaller-granularity sequential sub-works in the real world. Recent studies have proposed repetition-pruning to address this issue, but the buffer allocation strategy for enhancing buffer device utilization remains understudied. In preliminary experiments observing input patterns of neural layers with different datasets, the similarity of repetition allows us to transfer the buffer allocation strategy obtained from a small dataset to the computation with a large dataset. Hence, this paper proposes a practical compute-reuse mechanism for ReRAM-based PIM, called CRPIM, which replaces repetitive computations with buffering and reading. Moreover, the subsequent buffer allocation problem is resolved at both inter-layer and intra-layer levels. Our experimental results demonstrate that CRPIM significantly reduces ReRAM cells and execution time while maintaining adequate buffer and energy overhead.

With the arrival of 5G technology and the popularization of the Internet of Things (IoT), mobile edge computing (MEC) has great potential in handling delay-sensitive and compute-intensive (DSCI) applications. Meanwhile, the need for reduced latency and improved energy efficiency in terminal devices is becoming urgent increasingly. However, the users are affected by channel conditions and bursty computational demands in dynamic MEC environments, which can lead to longer task correspondence times. Therefore, finding an efficient task offloading method in stochastic systems is crucial for optimizing system energy consumption. Additionally, the delay due to frequent user–MEC interactions cannot be overlooked. In this article, we initially frame the task offloading issue as a dynamic optimization issue. The goal is to minimize the system’s long-term energy consumption while ensuring the task queue’s stability over the long term. Using the Lyapunov optimization technique, the task processing deadline problem is converted into a stability control problem for the virtual queue. Then, a novel Lyapunov-guided deep reinforcement learning (DRL) for delay-aware offloading algorithm (LyD2OA) is designed. LyD2OA can figure out the task offloading scheme online, and adaptively offload the task with better network quality. Meanwhile, it ensures that deadlines are not violated when offloading tasks in poor communication environments. In addition, we perform a rigorous mathematical analysis of the performance of Ly2DOA and prove the existence of upper bounds on the virtual queue. It is theoretically proven that LyD2OA enables the system to realize the trade-off between energy consumption and delay. Finally, extensive simulation experiments verify that LyD2OA has good performance in minimizing energy consumption and keeping latency low.

Video streaming has dominated Internet traffic over the past few years, spurring innovations in transport protocols. The QUIC protocol has advantages over TCP, such as faster connection setup and alleviating head-of-line blocking. Multi-path transport protocols like Multipath QUIC (MPQUIC) have been proposed to aggregate the bandwidth of multiple links and provide reliable transmission in poor network conditions. However, reliable transmission incurs unnecessary retransmission costs for MPQUIC, resulting in deteriorating performance, especially in real-time video streaming. Partially reliable transmission, which supports both reliable and unreliable delivery, may perform better by trading off data reliability and timeliness. In this paper, we introduce MPR-QUIC, a multi-path partially reliable transmission protocol for QUIC. Based on MPQUIC, MPR-QUIC extends unreliable transmission to provide partially reliable transmission over multiple paths. Specific schedulers are designed in MPR-QUIC based on priority and deadline, respectively, for video streaming optimization. Video frames with high priority are transmitted first since frames with low priority cannot be decoded before their arrival. Additionally, to alleviate rebuffering and freezing of the video, as many frames as possible should be delivered before the deadline. We evaluate MPR-QUIC experimentally on a testbed and in emulations. Results show that the rebuffer time of MPR-QUIC is significantly decreased by 60% to 80% when compared to state-of-the-art multi-path transmission solutions. The completion ratio of transmitted data blocks is increased by almost 100%.

An anonymous transit pass system allows passengers to access transport services within fixed time periods, with their privileges automatically deactivating upon time expiration. Although existing transit pass systems are deployable on powerful devices like PCs, their adaptation to more user-friendly devices, such as mobile phones with smart cards, is inefficient due to their reliance on heavy-weight operations like bilinear maps. In this paper, we introduce an innovative anonymous transit pass system, dubbed $AnoPas$, optimized for deployment on mobile phones with smart cards, where the smart card is responsible for crucial lightweight operations and the mobile phone handles key-independent and time-consuming tasks. Group signatures with time-bound keys (GS-TBK) serve as our core component, representing a new variant of standard group signatures for the secure use of time-based digital services, preserving users’ privacy while providing flexible authentication services. We first constructed a practical GS-TBK scheme using the tag-based signatures and then applied it to the design of AnoPas. We achieve the most efficient passing protocol compared to the state-of-the-art AnoPas/GS-TBK schemes. We also present an implementation showing that our passing protocol takes around 38.6 ms on a smart card and around 33.6 ms on a mobile phone.

In cloud computing, the current challenge lies in managing massive data, which is a computationally overburdened environment for data users. Outsourced computation can effectively ease the memory and computation pressure on overburdened data storage. We propose an outsourced unbounded decryption scheme in the standard assumption and standard model for large data settings based on inner product computation. Security analysis shows that it can achieve adaptive security. The scheme involves the data owner transmitting encrypted data to a third-party cloud server, which is responsible for computing a significant amount of data. Then the ripe data is handed over to the data user for decryption computation. In addition, there is no need to give the prior bounds of the length of the plaintext vector in advance. This allows for the encryption algorithm to run without determining the length of the input data before the setup phase, that is, our scheme is on the unbounded setting. Through theoretical analysis, the storage overhead and communication cost of the data users remain independent of the ciphertext size. The experimental results indicate that the efficiency and performance are greatly enhanced, about 0.03S for data users at the expense of increased computing time on the cloud server.

Emerging Persistent Memory (PM) usually has the serious drawback of expensive write activities. Thus, existing PM-oriented B${}^{+}$-trees mainly concentrate on alleviating the write overhead (i.e., reducing PM writes and flush instructions). Unfortunately, due to the improper data organization in the sorted leaf node, existing solutions cause massive data migration when data inserting or node splitting occurs. In this paper, we propose a write-optimized PM-oriented B${}^{+}$-tree with aligned flush and selective migration, called WPB${}^{+}$-tree, to solve the above problems. WPB${}^{+}$-tree first adopts a buffer-assisted mechanism that temporarily stores the newly inserted data to reduce the overhead of entry shifts. Second, WPB${}^{+}$-tree employs a selective migration of node entries scheme to achieve less than half of the data migration when a node is split. Moreover, existing PM-oriented B${}^{+}$-trees usually employ a coarse-grained lock to avoid thread conflicts, which can severely degrade the concurrency efficiency. Thus, we further propose a fine-grained lock technique for WPB${}^{+}$-tree, namely, parallel-efficient WPB${}^{+}$-tree (WOPE), to improve the concurrency efficiency. We implement the proposed WPB${}^{+}$-tree and WOPE on Linux and conduct extensive evaluations with actual persistent memory, where WOPE achieves 23.5%, 30.7%, and 15.3% of performance improvement (insert, read, and scan) over the straightforward solutions (i.e., SSB-Tree, Fast&Fair, and wB${}^{+}$tree), and 10.1% of performance improvement over WPB${}^{+}$-tree, on average.

The automotive Electrical/Electronic (E/E) architecture with Time-Sensitive Networking (TSN) as the backbone network and Controller Area Network (CAN) as the intra-domain network has attracted extensive research attention. In this architecture, the CAN-TSN gateway serves as a vital hub for communication between the CAN and TSN networks. However, with frequent information exchange between domains, multiple real-time applications inevitably compete for the same network resources. The limited availability of schedule table entries and bandwidth allocation pose challenges in scheduling design. To mitigate the transmission conflicts at the CAN-TSN gateway, this paper proposes a CAN-to-TSN scheduler consisting of two primary stages. The first stage introduces the Message Aggregation Optimization (MAO) algorithm to aggregate multiple CAN messages into a single TSN message, ultimately decreasing the communication overhead and the schedule table entries number. The second stage proposes the Exploratory Message Scheduling Optimization (EMSO) algorithm based on MAO. EMSO disaggregates and reassembles the CAN messages with small deadlines within the currently un-scheduled TSN message to improve the acceptance ratio of CAN messages. Experimental results demonstrate that EMSO achieves an average acceptance ratio of CAN messages 4.3% higher in preemptive mode and 8.2% higher in non-preemptive mode in TSN than state-of-the-art algorithms.

We present high performance, multi-threaded implementations of three GEMM-based convolution algorithms for multicore processors with ARM and RISC-V architectures. The codes are integrated into CONVLIB, a library that has the following unique features: (1) scripts to automatically generate a key component of GEMM, known as the micro-kernel, which is typically written in assembly language; (2) a modified analytical model to automatically tune the algorithms to the underlying cache architecture; (3) the ability to select four hyper-parameters: micro-kernel, cache parameters, parallel loop, and GEMM algorithm dynamically between calls to the library, without recompiling it; and (4) a driver to identify the best hyper-parameters. In addition, we provide a detailed performance evaluation of the convolution algorithms, on five ARM and RISC-V processors, and we publicly release the codes.

The advent of cloud computing has made cloud server outsourcing increasingly popular among data owners. However, the storage of sensitive data on cloud servers engenders serious challenges for the security and privacy of data. Public Key Authenticated Encryption with Keyword Search (PAEKS) is an effective method that protects information confidentiality and supports keyword searches. Identity-Based Authenticated Encryption with Keyword Search (IBAEKS) is a PAEKS variant in identity-based settings, designed for solving the intractable certificate management problem. To the best of our knowledge, only two IBAEKS schemes exist in the literature, both presented with weak security models that make them vulnerable against what is known as Fully Chosen Keyword attacks. Moreover, the existing IBAEKS schemes are based on the time-consuming bilinear pairing operation, leading to a significant increase in computational cost. To overcome these issues, in this paper, we first propose an enhanced security model for IBAEKS and compare it with existing models. We then prove that the existing IBAEKS schemes are not secure in our enhanced model. We also propose an efficient pairing-free dIBAEKS scheme and prove that it is secure under the enhanced security model. Finally, we compare our proposed scheme with related constructions to indicate its overall superiority.