Dependability Analysis of Data Storage Systems in Presence of Soft Errors

Kishani, Mostafa; Tahoori, Mehdi B.; Asadi, Hossein

doi:10.1109/tr.2018.2888515

Cited by 19 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These errors can happen in the SSD controller and the flash memory packages. There are research works that investigate the effect of the soft errors on the SSD controller and the availability and reliability of the storage systems [23]. On the other hand, for addressing the soft errors on the flash memory packages, the error correction unit is utilized to correct the data errors.…”

Section: Hardware Architecturementioning

confidence: 99%

Computational storage: an efficient and scalable platform for big data and HPC applications

et al. 2019

View full text Add to dashboard Cite

In the era of big data applications, the demand for more sophisticated data centers and high-performance data processing mechanisms is increasing drastically. Data are originally stored in storage systems. To process data, application servers need to fetch them from storage devices, which imposes the cost of moving data to the system. This cost has a direct relation with the distance of processing engines from the data. This is the key motivation for the emergence of distributed processing platforms such as Hadoop, which move process closer to data. Computational storage devices (CSDs) push the "move process to data" paradigm to its ultimate boundaries by deploying embedded processing engines inside storage devices to process data. In this paper, we introduce Catalina, an efficient and flexible computational storage platform, that provides a seamless environment to process data in-place. Catalina is the first CSD equipped with a dedicated application processor running a full-fledged operating system that provides filesystem-level data access for the applications. Thus, a vast spectrum of applications can be ported for running on Catalina CSDs. Due to these unique features, to the best of our knowledge, Catalina CSD is the only in-storage processing platform that can be seamlessly deployed in clusters to run distributed applications such as Hadoop MapReduce and HPC applications in-place without any modifications on the underlying distributed processing framework. For the proof of concept, we build a fully functional Catalina prototype and a CSD-equipped platform using 16 Catalina CSDs to run Intel HiBench Hadoop and HPC benchmarks to investigate the benefits of deploying Catalina CSDs in the distributed processing environments. The experimental results show up to 2.2× improvement in performance and 4.3× reduction in energy consumption, respectively, for running Hadoop MapReduce benchmarks. Additionally, thanks to the Neon SIMD engines, the performance and energy efficiency of DFT algorithms are improved up to 5.4× and 8.9×, respectively.

show abstract

Section: Hardware Architecturementioning

confidence: 99%

Computational storage: an efficient and scalable platform for big data and HPC applications

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Consequently, users are reluctant to store critical data using cloud storage. In contrast, on-premises storage can help to prevent data leakage because the user can set the security policy of the local server to control how the data are stored and who has access [12,13]. However, local storage demands a high cost to install and maintain the hardware, with the risk that all data can become inaccessible due to a local server failure [14].…”

Section: Introductionmentioning

confidence: 99%

Design and Implementation of Virtual Private Storage Framework Using Internet of Things Local Networks

2020

View full text Add to dashboard Cite

This paper presents a virtual private storage framework (VPSF) using Internet of Things (IoT) local networks. The VPSF uses the extra storage space of sensor devices in an IoT local network to store users’ private data, while guaranteeing expected network lifetime, by partitioning the storage space of a sensor device into data and system volumes and, if necessary, logically integrating the extra data volumes of the multiple sensor devices to virtually build a single storage space. When user data need to be stored, the VPSF gateway divides the original data into several blocks and selects the sensor devices in which the blocks will be stored based on their residual energy. The blocks are transmitted to the selected devices using the modified speedy block-wise transfer (BlockS) option of the constrained application protocol (CoAP), which reduces communication overhead by retransmitting lost blocks without a retransmission request message. To verify the feasibility of the VPSF, an experimental implementation was conducted using the open-source software libcoap. The results demonstrate that the VPSF is an energy-efficient solution for virtual private storage because it averages the residual energy amounts for sensor devices within an IoT local network and reduces their communication overhead.

show abstract

“…These external disturbances create extra charges into the struck hardware component, causing a bit-flip. In turn, such a bit-flip can corrupt data, cause a jump through the executing program or even wrongly control an actuator [12,15].…”

Section: Introductionmentioning

confidence: 99%

Automatic Implementation of Control Flow Error Detection Techniques

Vankeirsbilck

Hallez

Boydens

2019

Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control

View full text Add to dashboard Cite

Modern embedded systems are prone to erroneous bit-flips introduced in its hardware by external disturbances such as alpha particles, electromagnetic interference or intentional external attackers. In order to protect embedded systems against these disturbances, a wide variety of software-implemented detection techniques have been proposed, a.o. by the authors of this paper. Implementing those techniques, however, can be arduous and error-prone since they have to be implemented in low-level code, e.g. assembly. To overcome this problem we propose a compiler extension, in the form of a plugin, that can automatically add any supported technique to the low-level code of the target program. We discuss the internal working of our compiler extension and conclude with a demonstration using an example program and validate the effectiveness of the introduced countermeasures by running a fault injection campaign.

show abstract

Dependability Analysis of Data Storage Systems in Presence of Soft Errors

Cited by 19 publications

References 34 publications

Computational storage: an efficient and scalable platform for big data and HPC applications

Computational storage: an efficient and scalable platform for big data and HPC applications

Design and Implementation of Virtual Private Storage Framework Using Internet of Things Local Networks

Automatic Implementation of Control Flow Error Detection Techniques

Contact Info

Product

Resources

About