Abstract: Scalable nonvolatile memory DIMMs will finally be commercially available with the release of the Intel Optane DC Persistent Memory Module (or just "Optane DC PMM"). This new nonvolatile DIMM supports byte-granularity accesses with access times on the order of DRAM, while also providing data storage that survives power outages. This work comprises the first in-depth, scholarly, performance review of Intel's Optane DC PMM, exploring its capabilities as a main memory device, and as persistent, byte-addressable memory exposed to user-space applications. This report details the technologies performance under a number of modes and scenarios, and across a wide variety of macro-scale benchmarks. Optane DC PMMs can be used as large memory devices with a DRAM cache to hide their lower bandwidth and higher latency. When used in this Memory (or cached) mode, Optane DC memory has little impact on applications with small memory footprints. Applications with larger memory footprints may experience some slow-down relative to DRAM, but are now able to keep much more data in memory. When used under a file system, Optane DC PMMs can result in significant performance gains, especially when the file system is optimized to use the load/store interface of the Optane DC PMM and the application uses many small, persistent writes. For instance, using the NOVA-relaxed NVMM file system, we can improve the performance of Kyoto Cabinet by almost 2x. Optane DC PMMs can also enable user-space persistence where the application explicitly controls its writes into persistent Optane DC media. In our experiments, modified applications that used user-space Optane DC persistence generally outperformed their file system counterparts. For instance, the persistent version of RocksDB performed almost 2x faster than the equivalent program utilizing an NVMM-aware file system.

Abstract: Collocated data processing and storage are the norm in biological computing systems such as the mammalian brain. As our ability to create better hardware improves, new computational paradigms are being explored beyond von Neumann architectures. Integrated photonic circuits are an attractive solution for on-chip computing which can leverage the increased speed and bandwidth potential of the optical domain, and importantly, remove the need for electro-optical conversions. Here we show that we can combine integrated optics with collocated data storage and processing to enable all-photonic in-memory computations. By employing nonvolatile photonic elements based on the phase-change material, Ge2Sb2Te5, we achieve direct scalar and matrix-vector multiplication, featuring a novel single-shot Write/Erase and a drift-free process. The output pulse, carrying the information of the light-matter interaction, is the result of the computation. Our all-optical approach is novel, easy to fabricate and operate, and sets the stage for development of entirely photonic computers.

Abstract: Artificial intelligence and other data-intensive applications have escalated the demand for data storage and processing. New computing devices, such as phase-change random access memory (PCRAM)-based neuro-inspired devices, are promising options for breaking the von Neumann barrier by unifying storage with computing in memory cells. However, current PCRAM devices have considerable noise and drift in electrical resistance that erodes the precision and consistency of these devices. We designed a phase-change heterostructure (PCH) that consists of alternately stacked phase-change and confinement nanolayers to suppress the noise and drift, allowing reliable iterative RESET and cumulative SET operations for high-performance neuro-inspired computing. Our PCH architecture is amenable to industrial production as an intrinsic materials solution, without complex manufacturing procedure or much increased fabrication cost.

Abstract: The explosive increase in digital communications in the Big Data and internet of Things era spurs the development of universal memory that can run at high speed with high-density and nonvolatile storage capabilities, as well as demonstrating superior mechanical flexibility for wearable applications. Among various candidates for the next-generation information storage technology, resistive switching memories distinguish themselves with low power consumption, excellent downscaling potential, easy 3D stacking, and high CMOS compatibility, fulfilling key requirements for high-performance data storage. Employing organic and hybrid switching media in addition allows light weight and flexible integration of molecules with tunable device performance via molecular design-cum-synthesis strategy. In this review, we present a timely and comprehensive review of the recent advances in organic and hybrid resistive switching materials and devices, with particular attention on their design principles for electronic property tuning and flexible device performance. The current challenges posed with development of organic and hybrid resistive switching materials and flexible memory devices, together with their future perspectives, are also discussed.

Abstract: Laser pulses are the shortest stimulus known to control the magnetization of materials and to switch magnetic devices on the picosecond to femtosecond timescales. Femtosecond laser pulses have been able to trigger the fastest changes in the magnetic state of matter, and thus these pulses may lead to technologies with increased speed and energy efficiency of magnetic data storage and memory. In the past decade, materials enabling optical control of magnetism and concepts of devices employing such opto-magnetic phenomena have been shown. In this Review, we explore ultrafast all-optical switching (AOS) of magnetization as the least-dissipative and fastest method for magnetic writing. We outline the physical processes responsible for mechanisms of AOS, define the materials suitable for optical control of magnetism and test these mechanisms and materials against three important criteria of recording: speed, accompanying dissipations and scalability. In particular, we emphasize that switching magnetization with the help of light outperforms other methods in terms of the speed of the write–read magnetic recording event (less than 20 ps) and the unprecedentedly low heat load (<6 J cm−3). Finally, we outline the integration of AOS in spintronic devices and the perspective of large-scale integration towards magnetic random access memory and other memory applications with low-energy dissipations. Laser pulses can trigger fast changes in magnetic state, facilitating new magnetic data storage and memory devices. This Review outlines the mechanisms of all-optical switching and the materials suitable for the optical control of magnetism and tests these mechanisms and materials in terms of speed, accompanying dissipations and scalability. Finally, the large-scale integration of devices in memory applications with low-energy dissipations is discussed.

Abstract: Optically storing and addressing data on photonic chips is of particular interest as such capability would eliminate optoelectronic conversion losses in data centers. It would also enable on-chip non-von Neumann photonic computing by allowing multinary data storage with high fidelity. Here, we demonstrate such an optically addressed, multilevel memory capable of storing up to 34 nonvolatile reliable and repeatable levels (over 5 bits) using the phase change material Ge2Sb2Te5 integrated on a photonic waveguide. Crucially, we demonstrate for the first time, to the best of our knowledge, a technique that allows us to program the device with a single pulse regardless of the previous state of the material, providing an order of magnitude improvement over previous demonstrations in terms of both time and energy consumption. We also investigate the influence of write-and-erase pulse parameters on the single-pulse recrystallization, amorphization, and readout error in our multilevel memory, thus tailoring pulse properties for optimum performance. Our work represents a significant step in the development of photonic memories and their potential for novel integrated photonic applications.

Abstract: The need for continuous size downscaling of silicon transistors is driving the industrial development of strategies to enable further footprint reduction1,2. The atomic thickness of two-dimensional materials allows the potential realization of high-area-efficiency transistor architectures. However, until now, the design of devices composed of two-dimensional materials has mimicked the basic architecture of silicon circuits3-6. Here, we report a transistor based on a two-dimensional material that can realize photoswitching logic (OR, AND) computing in a single cell. Unlike the conventional transistor working mechanism, the two-dimensional material logic transistor has two surface channels. Furthermore, the material thickness can change the logic behaviour-the architecture can be flexibly expanded to achieve in situ memory such as logic computing and data storage convergence in the same device. These devices are potentially promising candidates for the construction of new chips that can perform computing and storage with high area-efficiency and unique functions.

Abstract: Edge computing is an important tool for smart computing, which brings convenience to data processing as well as security problems. In particular, the security of data storage under edge computing has become an obstacle to its widespread use. To solve the problem, the mechanism combing blockchain with regeneration coding is proposed to improve the security and reliability of stored data under edge computing. Our contribution is as follows. 1) According to the three-tier edge computing architecture and data security storage requirements, we proposed hybrid storage architecture and model specifically adapted to edge computing. 2) Making full use of the data storage advantages of edge network devices and cloud storage servers, we build a global blockchain in the cloud service layer and local blockchain is built on the terminals of the Internet of things. Moreover, the regeneration coding is utilized to further improve the reliability of data storage in blockchains. 3) Our scheme provides a mechanism for periodically validating hash values of data to ensure the integrity of data stored in global blockchain.

Abstract: DNA promises to be a high density data storage medium, but physical storage poses a challenge. To store large amounts of data, pools must be physically isolated so they can share the same addressing scheme. We propose the storage of dehydrated DNA spots on glass as an approach for scalable DNA data storage. The dried spots can then be retrieved by a water droplet using a digital microfluidic device. Here we show that this storage schema works with varying spot organization, spotted masses of DNA, and droplet retrieval dwell times. In all cases, the majority of the DNA was retrieved and successfully sequenced. We demonstrate that the spots can be densely arranged on a microfluidic device without significant contamination of the retrieval. We also demonstrate that 1 TB of data could be stored in a single spot of DNA and successfully retrieved using this method. DNA as a high density storage medium is receiving increasing attention, but long term physical storage is an unsolved problem. Here the authors show that up to 1 TB of data stored as dehydrated DNA spots on a glass cartridge can be retrieved in a spot of water using digital microfluidics with minimal data loss and contamination.

Abstract: Memristors with enormous storage capacity and superior processing efficiency are of critical importance to overcome the Moore’s Law limitation and von Neumann bottleneck problems in the big data and artificial intelligence era. In particular, the integration of multifunctionalities into a single memristor promises an essential strategy of obtaining a high-performance electronic device that satisfies the nowadays increasing demands of data storage and processing. In this contribution, we report a proof-of-concept polymer memristive processing-memory unit that demonstrates programmable information storage and processing capabilities. By introducing redox active moieties of triphenylamine and ferrocene onto the pendants of fluorene skeletons, the conjugated polymer exhibits triple oxidation behavior and interesting memristive switching characteristics. Associated with the unique electrochemical and electrical behavior, the polymer device is capable of executing multilevel memory, decimal arithmetic operations of addition, subtraction, multiplication and division, as well as simple Boolean logic operations. Though designing conductive polymers for memory devices is attractive for future low-cost flexible electronics, a proof-of-concept device has yet to be realized. Here, the authors report a redox-gated polymer memristive processing unit with programmable multilevel storage and logic functionalities.

Abstract: To lower the cost of electricity produced, advanced high-efficiency power cycles operating at temperatures above 600 °C (such as the supercritical CO2 Brayton cycle) are presently being developed for use in both nuclear and concentrating solar power (CSP) plants. Incorporating thermal energy storage into CSP plants allows renewable energy to be generated while simultaneously providing reliability and stability to the grid. Sensible energy storage using molten nitrate salts is used in the majority of CSP plants. However, nitrate salts decompose at around 600 °C, hence an alternative storage medium is required to support the development of next generation high-efficiency CSP plants. Because of practical experience with molten salt storage in the two-tank configuration in industry, continuing to use fluid media is an attractive option, although thermal storage is also possible with other types of storage media (e.g. solids and phase change materials). This paper critically reviews options for energy storage in fluids that are stable over 600 °C. The focus is on three alternative molten salts — fluorides, chlorides and carbonates — which are assessed based on their thermophysical properties and cost. A brief review of liquid metal and molten glass storage options is included for completeness. Corrosion of containment materials is an important consideration in the choice of storage media, because if exotic materials are required, the cost of the storage tanks can dominate the overall storage cost. Therefore, this paper includes a summary of the main corrosion issues relating to containment of the more promising storage fluids considered herein, identifying possible tank materials and corrosion mitigation options.

Abstract: In view of optimizing the configuration of each unit’s capacity for energy storage in the microgrid system, in order to ensure that the planned energy storage capacity can meet the reasonable operation of the microgrid’s control strategy, the power fluctuations during the grid-connected operation of the microgrid are considered in the planning and The economic benefit of hybrid energy storage is quantified. A multi-objective function aiming at minimizing the power fluctuation on the DC bus in the microgrid and optimizing the capacity ratio of each energy storage system in the hybrid energy storage system (HESS) is established. The improved particle swarm algorithm (PSO) is used to solve the objective function, and the solution is applied to the microgrid experimental platform. By comparing the power fluctuations of the battery and the supercapacitor in the HESS, the power distribution is directly reflected. Comparing with the traditional mixed energy storage control strategy, it shows that the optimized hybrid energy storage control strategy can save 4.3% of the cost compared with the traditional hybrid energy storage control strategy, and the performance of the power fluctuation of the renewable energy is also improved. It proves that the proposed capacity configuration of the HESS has certain theoretical significance and practical application value.

Abstract: Rapid progress of digital communications in the Big Data and Internet of things has stimulated the exploration of next-generation data storage devices. Among various candidates, resistive switching (RS) memory devices and artificial synapses are in the spotlight due to their low power consumption, downscaling potential, and fast operation speed. The exceptional electronic and mechanical characteristics of metal halide perovskites (MHPs) have greatly promoted their application in nonvolatile high density, low-cost, and flexible memory devices. Here, we distill the current state-of-the-art and highlight recent advances of MHP based RS memory devices and artificial synapses. The general structure and characteristics of RS memory devices are first introduced. Afterwards we discuss the excellent memory behaviors accompanied by detailed working mechanisms. Finally, the current challenges and future development prospects are also discussed. This review article is expected to pave the way in the rational design of MHP based memory devices and artificial synapses with unprecedented performance improvement.

Abstract: Digital polymers with precisely ordered units acting as the coded 0- or 1-bit, are introduced as a promising option for molecular data storage. However, the pursuit of better performance in terms of high storage capacity and useful functions never stops. Herein, we propose a concept of an information-coded 2D digital dendrimer. The divergent growth via thiol-maleimide Michael coupling allows precise arrangements of the 0- and 1-bits in the uniform dendrimers. A protocol for calculating the storage capacity of non-linear binary digital dendrimer is established based on data matrix barcode, generated by the tandem mass spectrometry decoding and encryption. Furthermore, the generated data matrix barcode can be read by a common hand-held device to cater the applications such as item identification, traceability and anticouterfeiting purpose. This work demonstrates the high data storage capacity of a uniform dendrimer and uncovers good opportunities for the digital polymers. Digital polymers with precisely ordered units for next-generation data storage media are continually investigated for higher storage capacity. Here the authors show the synthesis of information-coded 2D digital dendrimers and the generated data matrix barcode can be read by a common hand-held device.

Abstract: DNA-based data storage has emerged as a promising method to satisfy the exponentially increasing demand for information storage. However, practical implementation of DNA-based data storage remains a challenge because of the high cost of data writing through DNA synthesis. Here, we propose the use of degenerate bases as encoding characters in addition to A, C, G, and T, which augments the amount of data that can be stored per length of DNA sequence designed (information capacity) and lowering the amount of DNA synthesis per storing unit data. Using the proposed method, we experimentally achieved an information capacity of 3.37 bits/character. The demonstrated information capacity is more than twice when compared to the highest information capacity previously achieved. The proposed method can be integrated with synthetic technologies in the future to reduce the cost of DNA-based data storage by 50%.

Abstract: Based on decreasing the flexibility of the power grid through the integration of large-scale renewable energy, a multi-energy storage system architectural model and its coordination operational strategy with the same flexibility as in the pumped storage power station and battery energy storage system (BESS) are studied. According to the new energy fluctuation characteristics and the different peak valley parameters in the power grid, this paper proposes a electricity heat hydrogen multienergy storage system (EHH-MESS) and its coordination and optimization operational model to reduce the curtailment of wind power and photovoltaic (PV) to the power grid and improve the flexibility of the power grid. Finally, this paper studied the simulation model of an energy storage optimization control strategy after the multi-energy storage system is connected to the distribution networks, and analyzed three operational modes of the multi-energy storage system. The simulation results show that the EHH-MESS proposed in this paper has a better power grid regulation flexibility and economy, and can be used to replace the battery energy storage system based on MATLAB.

Abstract: The performance of a 2 × 500 kWhth thermal energy storage (TES) technology has been tested at the Masdar Institute Solar Platform (MISP) at temperatures up to 380 °C over a period of more than 20 months. The TES is based on a novel, modular storage system design, a new solid-state concrete-like storage medium, denoted HEATCRETE® vp1, - and has cast-in steel pipe heat exchangers. Measured data after specific intervals during various operation modes are analysed, and validation of system performance is done through direct comparison between measured values and numerically simulated performance. The demonstrated and measured long-term performance of the TES matches predictions based on performance simulations and proves the operational feasibility of the modular TES design. After accumulating close to 6 000 operational hours, inspection of extracted thermal elements prove that there is no degradation of the storage material, and no separation between steel pipes and storage material is observed. Measurements of core samples of the storage medium extracted from the TES confirms the material properties and stability. The thermal element design and storage material as demonstrated in the TES pilot has thus been proved to work in its final form with expected conditions and shows absolutely no sign of performance degradation. The modularity and simplicity of the TES design enables flexibility in scaling high temperature TES systems for among others industrial waste heat recovery, thermal power plants and concentrating solar power applications, thermal power plant.

Abstract: With the amount of data being stored increasing rapidly, there is significant interest in exploring alternative storage technologies. In this context, DNA-based storage systems can offer significantly higher storage densities (petabytes/gram) and durability (thousands of years) than current technologies. Specifically, DNA has been found to be stable over extended periods of time which has been demonstrated in the analysis of organisms long since extinct. Recent advances in DNA sequencing and synthesis pipelines have made DNA-based storage a promising candidate for the storage technology of the future.Recently, there have been multiple efforts in this direction, focusing on aspects such as error correction for synthesis/sequencing errors and erasure correction for handling missing sequences. The typical approach is to use separate codes for handling errors and erasures, but there is limited understanding of the efficiency of this framework. Furthermore, the existing techniques use short block-length codes and heavily rely on read consensus, both of which are known to be suboptimal in coding theory.In this work, we study the tradeoff between the writing and reading costs involved in DNA-based storage and propose a practical scheme to achieve an improved tradeoff between these quantities. Our scheme breaks with the traditional separation framework and instead uses a single large block-length LDPC code for both erasure and error correction. We also introduce novel techniques to handle insertion and deletion errors introduced by the synthesis process. For a range of writing costs, the proposed scheme achieves 30-40% lower reading costs than state-of-the-art techniques on experimental data obtained using array synthesis and Illumina sequencing.The code, data, and Supplementary Material is available at https://github.com/shubhamchandak94/LDPC_ DNA_storage.

Abstract: The extreme density of DNA presents a compelling advantage over current storage media; however, to reach practical capacities, new systems for organizing and accessing information are needed. Here, we use chemical handles to selectively extract unique files from a complex database of DNA mimicking 5 TB of data and design and implement a nested file address system that increases the theoretical maximum capacity of DNA storage systems by five orders of magnitude. These advancements enable the development and future scaling of DNA-based data storage systems with modern capacities and file access capabilities.

Abstract: In the current "big data" era, the state-of-the-art optical data storage (ODS) has become a front-runner in the competing data storage technologies. As one of the most promising methods for breaking the physical limitation suffered by traditional ones, the advance of optically stimulated luminescence (OSL) based optical storage technique is now still limited by the simultaneous single-level write-in and readout in a same spot. In this work, to bridge the data-capacity gap, we report for the first time a novel and promising nonphysical multidimensional OSL-based ODS flexible medium for erasable multilevel optical data recording and reading. We tailor multidimensional traps with discrete, narrowly distributed energy levels through (multi-)codoping of selective trivalent rare-earth ions into Eu2+-activated barium orthosilicate (Ba2SiO4). Upon UV/blue light illumination, information can be sequentially recorded in different traps assisted by thermal cleaning with an increase of storage capacity by orders of magnitude, which is addressable individually in the whole domain or bit-by-bit mode without the crosstalk by designed thermal/optical stimuli. Remarkably, good data retention and robust fatigue resistance have been achieved in recycle data recording. Insight is forged from charge carrier dynamics and interactions with traps for a universal method of data storage, and proof-of-concept applications are also demonstrated, thereby providing the way to not only rewritable multilevel ODS but also high-security encryption/decryption.

Abstract: Our research focuses on the storage decision in a semi‐automated storage system, where the inventory is stored on mobile storage pods. In a typical system, each storage pod carries a mixture of items, and the inventory of each item is spread over multiple storage pods. These pods are transported by robotic drives to stationary stations on the boundary of the storage zone where associates conduct pick or stow operations. The storage decision is to decide to which storage location within the storage zone to return a pod upon the completion of a pick or stow operation. The storage decision has a direct impact on the total travel time and hence the workload of the robotic drives. We develop a fluid model to analyze the performance of velocity‐based storage policies. We characterize the maximum possible improvement from applying a velocity‐based storage policy in comparison to the random storage policy. We show that class‐based storage with two or three classes can achieve most of the potential benefits and that these benefits increase with greater variation in the pod velocities. To validate the findings, we build a discrete‐time simulator with real industry data. We observe an 8% to 10% reduction in the travel distance with a 2‐class or 3‐class storage policy, depending on the parameter settings. From a sensitivity analysis, we establish the robustness of the class‐based storage policies as they continue to perform well under a broad range of warehouse settings, including different zoning strategies, resource utilization, and space utilization levels.

Abstract: With extremely high density and durable preservation, DNA data storage has become one of the most cutting-edge techniques for long-term data storage. Similar to traditional storage which impose restrictions on the form of encoded data, data stored in DNA storage systems are also subject to two biochemical constraints, i.e., maximum homopolymer run limit and balanced GC content limit. Previous studies used successive process to satisfy these two constraints. As a result, the process suffers low efficiency and high complexity. In this letter, we propose a novel content-balanced run-length limited code with an efficient code construction method, which generates short DNA sequences that satisfy both constraints at one time. Besides, we develop an encoding method to map binary data into long DNA sequences for DNA data storage, which ensures both local and global stability in terms of satisfying the biochemical constraints. The proposed encoding method has high effective code rate of 1.917 bits per nucleotide and low coding complexity.

Abstract: Solution-processed nonvolatile organic transistor memory devices are fabricated by employing semiconductor blends of p-channel 6,13-bis(triisopropylsilylethynyl)pentacene and n-channel poly{[ N, N'-bis(2-octyldodecyl)-naphthalene-1,4,5,8-bis(dicarboximide)-2,6-diyl]- alt-5,5'-(2,2'-bithiophene)} (P(NDI2OD-2T); N2200) on polystyrene-brush as a polymer electret. Electret-based memory characteristics are significantly changed depending on the frontier molecular orbitals of the active semiconductors because the charge-trapping efficiency is mainly determined by the energy barrier to transfer electrons and holes from the active channel to the electret layer. A semiconductor mixture with an optimized blending ratio results in an efficient programming and erasing process. Thus, we obtained a remarkably high ratio of ON/OFF current (memory ratio) about 107 and a large amount of shifts in the threshold voltage (memory window) between the programmed and erased states of 55 V, while single-component N2200 showed only writing-once-read-many (WORM)-type memory. Especially, the programmed data can be stably retained more than 10 years with a sufficient memory ratio of 103. Furthermore, our semiconductor blend system leads to preferable vertical phase separation, which affords good reliability under a sequential memory operation condition as well as stability in ambient air. It is expected that our memory devices can be applied for versatile data storage in printed and flexible electronic applications.

Abstract: The rapidly growing demand for data storage and processing, driven by artificial intelligence (AI) and other data-intensive applications, is posing a serious challenge for current computing devices based on the von Neumann architecture. For every calculation, data sets need to be shuffled sequentially between the processor, and multiple memory and storage units through bandwidth-limited and energy-inefficient interconnects, typically causing 40% power wastage. Phase-change materials (PCMs) show great promise to break this bottleneck by enabling nonvolatile memory devices that can optimize the complex memory hierarchy, and neuro-inspired computing devices that can unify computing with storage in memory cells. The articles in this issue of MRS Bulletin highlight recent breakthroughs in the fundamental materials science, as well as electronic and photonic implementations of these novel devices based on PCMs.

Abstract: Data storage is one of the hottest discussion topics in today’s connected world. The amount of data growth is expected to be exponential, while budget and space remain constricted. Since the transformation of storage device from planar NAND to 3D V-NAND [1], the areal density of semiconductor storage devices has continuously evolved and has surpassed the density of magnetic hard drives. By providing the largest storage capacity in the smallest footprint, 3D V-NAND has been leading the data center revolution in recent years. However, 3D-technology scaling faces several technical challenges [2]. (1) As the number of WL stacks increases the channel-hole etch process becomes a limit, since the total WL-mold height increases. (2) Interference between cells increases since the distance between WLs becomes smaller. (3) Faster data transfer speeds are required to support higher IO bandwidth.

Abstract: The Integration of renewable energy resources suffers from two fundamental issues: variability, and uncertainty of power output. These issues hinder the integration of renewable resources with the existing grid. This paper addresses these issues and proposes a new methodology to minimize the impact of intermittency by offering an alternative approach for energy storage. The concept of value storage is introduced as an alternative to energy storage to replace the typical large-scale battery energy storage system. The concept refers to the storage of excess renewable energy as products from industrial loads instead of the energy itself which enables a demand side management technique to be applied, namely, load shifting for some industrial plants. A hybrid Photovoltaic-wind turbine generator PV-WTG and storage system is proposed to penetrate the existing electric grid, with significant cost savings by displacing the conventional energy generation in a fossil fuel-rich location. A size optimization based on differential system cost is formulated and solved by an enhanced genetic algorithm technique. Uncertainty impact studies were done by incorporating multiple scenarios and comprehensive sensitivity analysis.

Abstract: Key-value stores are widely used as storage backends, due to their simple, yet flexible interface for cache, storage, file system, and database systems. However, when used with high performance NVMe devices, their high compute requirements for data management often leave the device bandwidth under-utilized. This leads to a performance mismatch of what the device is capable of delivering and what it actually delivers, and the gains derived from high speed NVMe devices is nullified. In this paper, we introduce KV-SSD (Key-Value SSD) as a key technology in a holistic approach to overcome such performance imbalance. KV-SSD provides better scalability and performance by simplifying the software storage stack and consolidating redundancy, thereby lowering the overall CPU usage and releasing the memory to user applications. We evaluate the performance and scalability of KV-SSDs over state-of-the-art software alternatives built for traditional block SSDs. Our results show that, unlike traditional key-value systems, the overall performance ofKV-SSD scales linearly, and delivers 1.6 to 57x gains depending on the workload characteristics.