Random Read Research Articles

Tiziana Fuoco

Angewandte ChemieVolume 133, Issue 34 p. 18526-18526 Introducing …Free Access Tiziana Fuoco First published: 09 July 2021 https://doi.org/10.1002/ange.202107180AboutSectionsPDF ToolsRequest permissionAdd to favorites ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onEmailFacebookTwitterLinked InRedditWechat Abstract “My motto is: dare to be different … My favorite time of day is late night when I take time for unspecific thinking and random reading …” Find out more about Tiziana Fuoco in her Introducing … Profile. Tiziana Fuoco The author presented on this page has published her first article as a submitting corresponding author in Angewandte Chemie: “Degradation in order: simple and versatile one-pot combination of two macromolecular concepts to encode diverse and spatially regulated degradability functions”: T. Fuoco, Angew. Chem. Int. Ed. 2021, 60, doi.org/10.1002/anie.202103143; Angew. Chem. 2021, 133, doi.org/10.1002/ange.202103143. Position: Researcher, Department of Fibre and Polymer Technology, KTH Royal Institute of Technology Homepage: https://www.kth.se/profile/tiziana ORCID: orcid.org/0000-0001-7135-9158 Education: 2012 Master's Degree in Chemistry, University of Salerno (Italy) 2016 PhD in Chemistry with Daniela Pappalardo and Marina Lamberti, University of Salerno 2016–2018 Postdoctoral position with Anna Finne-Wistrand, KTH Royal Institute of Technology (Sweden) Research: Degradable and functional polymers, copolymers, ring-opening (co)polymerization, macromolecular design and catalysis Hobbies: Reading, art, travel I am waiting for the day when someone will discover innovative and robust synthetic strategies to mimic the modus operandi of Nature. The biggest challenge facing my generation of scientists is to balance curiosity-driven research with societal needs. The greatest scientific advance of the last decade was the development of vaccines against COVID-19. I chose chemistry as a career because I wanted to create new molecules, indeed I started with organic chemistry and later I “scaled-up” to macromolecules. In five years, I hope to be an established research leader who is an inspiring supervisor and an enthusiastic, passionate colleague. If I were not a scientist, I would be a veterinarian. The advice I wish I had received is strive for excellence by following the pathway that is yours. My favorite time of day is late night when I take time for unspecific thinking and random reading. If I could be described as an animal it would be a cat according to my family and friends, but I prefer the Ugly Duckling. My motto is dare to be different. The most important quality of a role model is a quiet demeanor that hides a great intellect, preparation, patience, and discipline. My favorite way to spend a holiday is wandering around fishing villages and small islands in Southern Italy. My favorite composer is Ennio Morricone. My favorite quote is “Luck favors the prepared mind” (Louis Pasteur). My favorite painter is Sandro Botticelli together with Perugino and Caravaggio. Behind the Science References 1The idea of this contribution came out by building on different models and concepts that apply to diverse areas of chemistry, and on a good dose of reading and curiosity. Recently, I could dedicate myself to that idea. Indeed, I had been wondering and elaborating for some time now, on how to develop innovative ways to properly design the polymer's primary structure and endow specific performance, yet the simplicity and versatility of the strategy presented here were what surprised me the most on the way to the results. Google Scholar Volume133, Issue34August 16, 2021Pages 18526-18526 This is the German version of Angewandte Chemie. Note for articles published since 1962: Do not cite this version alone. Take me to the International Edition version with citable page numbers, DOI, and citation export. We apologize for the inconvenience. ReferencesRelatedInformation

Tiziana Fuoco.

"My motto is: dare to be different … My favorite time of day is late night when I take time for unspecific thinking and random reading …" Find out more about Tiziana Fuoco in her Introducing … Profile.

COPS: A complete oblivious processing system

The communication link between the processor and the off-chip memory in a modern computing system is vulnerable to hardware-level security attacks. An adversary can snoop the memory access behavior over the communication link to retrieve secret information even when the data transferred over the link are encrypted. The Oblivious RAM (ORAM) mechanism is a promising solution for such a problem. ORAM hides a true memory access with a group of random memory read and write operations. To reliably estimate the cost and performance overhead incurred by ORAM in real scenarios, a complete oblivious processing system where a fully functional ORAM is integrated into a real processor system that can run an Operating System (OS) and execute real-world application workloads is required. In this paper, we present a generic design for ORAM integration into a general-purpose processing system that can be used for the integration of any ORAM design. Our design targets the ORAM integration into high-performance memory bus (AXI). The AXI-compatibility of our design makes it usable with any processor system, thus, making the ORAM integration independent of processor architecture. As a case study, we develop a complete oblivious processing system (we call it COPS) by integrating an ORAM into a general-purpose RISC-V based processing system. We evaluate ORAM performance in real-world execution environment using realistic workloads with full Operating System support. Our elevation demonstrates that our ORAM prototype, on average, can return a 64 bytes requested block to the processor in 1.46μs, which is ∼ 13x faster than the prior-art implementation.

594-P: Use of Connected Glucometer with Real-Time Feedback and Counseling for Glycemic Control in Patients with Diabetes in India: BeatO's Connected Diabetes Care Program

Background: BeatO® offers connected diabetes care program including smartphone connected glucometer and mobile app with real time feedback and counselling support. This study aims to evaluate the effectiveness of BeatO's connected diabetes care program with real time contextualized feedback loop and certified diabetes educator (CDE) counselling on glycemic control of patients with diabetes in India. Method: A retrospective data-based evaluation study was conducted on BeatO® cloud database. Medical records of patients with diabetes, digitally engaged between September 2019 to July 2020 and received at least one CDE counselling were evaluated. Intervention included SMBG using smartphone connected glucometer, real-time feedback, and contextualized counselling by CDE based on SMBG findings. Pre-counselling (BG values recorded for 30 days period before first counselling) and post counselling (BG values recorded between 15th to 45th day after last counselling) impact on glycemic control and in range BG values were assessed. Results: Out of total 7514 diabetes patients, majority of them were aged >40 years (76%). Most of the patients had T2D (95%) followed by T1D (4.5%), GDM (0.3%), LADA (0.1%) and MODY (0.1%). A total of 10,667 counselling sessions were conducted out of which majority had one session (74%) followed by 2 (17%), 3 (5%), 4 (2%) and ≥5 (1%) sessions. Intervention was effective in significantly reducing mean BG values across fasting (by 9.5%), pre-prandial (9.3%), post prandial (8.9%) and random (6%) BG readings (p value <.05). Proportion of patients with in-range BG values post intervention increased by 9.6% (post prandial <180 mg/dL), 9.1% (random <200 mg/dL), 5.8% (fasting 80-130 mg/dL) and 3.3% (pre-prandial 80-130 mg/dL). Conclusion: BeatO's smartphone connected SMBG with real time feedback, nudges, alerts and CDE counselling is effective in glycemic control in real world patients with diabetes. Disclosure M. Sabharwal: Employee; Self; Health Arx Technologies Pvt. Ltd. M. Mohit: Consultant; Self; Health Arx Technologies Pvt. Ltd. V. K. Kolukula: None. R. Jangid: None. G. Chopra: Employee; Spouse/Partner; Health Arx Technologies Private Limited, Other Relationship; Self; Health Arx Technologies Private Limited.

Memory-Adjustable Navigation Piles with Applications to Sorting and Convex Hulls

We consider space-bounded computations on a random-access machine, where the input is given on a read-only random-access medium, the output is to be produced to a write-only sequential-access medium, and the available workspace allows random reads and writes but is of limited capacity. The length of the input is N elements, the length of the output is limited by the computation, and the capacity of the workspace is O ( S ) bits for some predetermined parameter S ≥ lg N . We present a state-of-the-art priority queue—called an adjustable navigation pile —for this restricted model. This priority queue supports M inimum in O (1) time, C onstruct in O ( N ) time, and E xtract - min in O ( N / S + lg S ) time for any S ≥ lg N . The priority queue can be further augmented in O ( N ) time to deal with a batch of at most S elements in a specified range of values at a time, and allow to I nsert (activate) or E xtract (deactivate) an element among these elements, such that I nsert and E xtract take O ( N / S + lg S ) time for any S ≥ lg N . We show how to use our data structure to sort N elements and to compute the convex hull of N points in the Euclidean plane in O ( N 2 / S + N lg S ) time for any S ≥ lg N . Following a known lower bound for the space-time product of any branching program for finding unique elements, both our sorting and convex-hull algorithms are optimal. The adjustable navigation pile has turned out to be useful when designing other space-efficient algorithms, and we expect that it will find its way to yet other applications.

Hard-to-Detect Fault Analysis in FinFET SRAMs

Manufacturing defects can cause hard-to-detect (HTD) faults in fin field-effect transistor (FinFET) static random access memories (SRAMs). Detection of these faults, such as random read outputs and out-of-spec parametric deviations, is essential when testing FinFET SRAMs. Undetected HTD faults result in test escapes, which lead to early in-field failures. This article presents a detailed analysis of HTD faults in FinFET SRAMs by exploring their sensitization and discussing solutions to improve HTD fault coverage during manufacturing testing. We first define the fault space for SRAMs and classify all faults in the space. Following this, we perform a systematic fault analysis based on injecting resistive defects in a memory cell, inspecting its behavior, and identifying HTD faults. Furthermore, we survey existing test solutions and discuss their HTD fault coverage and limitations. Based on our analysis, it is clear that no single test solution can fully detect all HTD faults, thus leading to test escapes. Hence, there is a need for new and more efficient test solutions. Improved detection of HTD faults could be achieved by using parametric test solutions, proposing solutions that cover yet-untargeted HTD faults, combining multiple test approaches into a single solution, and further exploring stress conditions. These new approaches would reduce test escapes and therefore improve the quality of FinFET SRAMs.

Approaching DRAM performance by using microsecond-latency flash memory for small-sized random read accesses

For applications in which small-sized random accesses frequently occur for datasets that exceed DRAM capacity, placing the datasets on SSD can result in poor application performance. For the read-intensive case we focus on in this paper, low latency flash memory with microsecond read latency is a promising solution. However, when they are used in large numbers to achieve high IOPS (Input/Output operations Per Second), the CPU processing involved in IO requests is an overhead. To tackle the problem, we propose a new access method combining two approaches: 1) optimizing issuance and completion of the IO requests to reduce the CPU overhead. 2) utilizing many contexts with lightweight context switches by stackless coroutines. These reduce the CPU overhead per request to less than 10 ns, enabling read access with DRAM-like overhead, while the access latency longer than DRAM can be hidden by the context switches. We apply the proposed method to graph algorithms such as BFS (Breadth First Search), which involves many small-sized random read accesses. In our evaluation, the large graph data is placed on microsecond-latency flash memories within prototype boards, and it is accessed by the proposed method. As a result, for the synthetic and real-world graphs, the execution times of the graph algorithms are 88--141% of those when all the data are placed in DRAM.

The Use of Artificial Intelligence Combined with Cloud Computing in the Design of Education Information Management Platform

Based on cloud computing theory and service-oriented architecture (SOA) design pattern, a smart education management platform is designed by using cloud computing and artificial intelligence technology. The platform is deployed in the server cluster environment, with Hadpoop managed storage cluster as the data storage center. For the education management, education portal and remote classroom, the corresponding services are launched, which has the advantages of intelligent and efficient, massive data access and intelligent collaborative management. The test results show that the platform can be used normally in the ports of computer, mobile phone and tablet computer, and can successfully complete the basic operations such as user registration and login, educational administration, and content storage. The results of database stress test show that the total time consumption of sequential read, sequential write, random read and random write is 752s, 312s, 968s and 211s, respectively when millions of simulated data are inserted into the database, which indicates that the database can support large-scale data access. The stress test results of content storage service show that when the number of clients is adjusted to 500, 1000 and 3000, the output quantity of each read-write interface can be maintained at about 3500 pages/min in 60s, indicating that the system can still run stably under the condition of high concurrency. The management platform discussed has practical significance to promote the development of intelligent and information-based education management.

Balanced Multi-Cell Modulation for Flash Memory

The cell (i.e., the storage unit) in flash memory is a floating-gate MOSFET where data is stored in the form of trapped charge into its floating gate. The flash memory die is organized into blocks where each block consists of multiple wordlines that can be individually programmed. Each wordline consists of tens of thousands of cells and is organised into log <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> (q) pages where q is the number of threshold voltage levels a cell has. The pages are separately encoded to enable reading them individually. A key parameter of flash memory is its random read speed where small chunks of data are read from random addresses. The disparity in random read speed across pages can be minimized by using a balanced Gray labelling (BGL) to map the binary data into threshold voltages which allows reading different pages with almost the same number of sensing operations. We provide new BGL instances whereby multi-cell modulation (MCM) is used to map binary data into threshold voltages of multiple cells. The proposed modulations are found through computer search routines and are shown to improve the flash memory reliability via numerical simulations.

Improving in-memory file system reading performance by fine-grained user-space cache mechanisms

Nowadays, as the memory capacity of servers become larger and larger, distributed in-memory file systems, which enable applications to interact with data at fast speed, have been widely used. However, the existing distributed in-memory file systems still face the problem of low data access performance in small data reading, which seriously reduce their usefulness in many important big data scenarios. In this paper, we analyze the factors that affect the performance of reading in-memory files and propose a two-layer user space cache management mechanism: in the first layer, we cache data packet references to reduce frequent page fault interruptions (packet-level cache); in the second layer, we cache and manage small file data units to avoid redundant inter-process communications (object-level cache). We further design a fine-grained caching model based on the submodular function optimization theory, for efficiently managing the variable-length cache units with partially overlapping fragments on the client side. Experimental results on synthetic and real-world workloads show that compared with the existing cutting-edge systems, the first level cache can double the reading performance on average, and the second level cache can improve random reading performance by more than 4 times. Our caching strategies also outperform the cutting-edge cache algorithms over 20% on hit ratio. Furthermore, the proposed client-side caching framework idea has been adopted by the Alluxio open source community, which shows the practical benefits of this work. • A packet-level cache policy for reducing page faults during sequential data reading. • An in-memory cache layer to cache the variable-length hot file fragments. • Proved that the variable-length cache problem is a submodular optimization problem theoretically. • Comprehensive performance evaluation of the client-side cache strategies.

Accelerated Isotropic Multiparametric Imaging by High Spatial Resolution 3D-QALAS With Compressed Sensing: A Phantom, Volunteer, and Patient Study.

The aims of this study were to develop an accelerated multiparametric magnetic resonance imaging method based on 3D-quantification using an interleaved Look-Locker acquisition sequence with a T2 preparation pulse (3D-QALAS) combined with compressed sensing (CS) and to evaluate the effect of CS on the quantitative mapping, tissue segmentation, and quality of synthetic images. A magnetic resonance imaging system phantom, containing multiple compartments with standardized T1, T2, and proton density (PD) values; 10 healthy volunteers; and 12 patients with multiple sclerosis were scanned using the 3D-QALAS sequence with and without CS and conventional contrast-weighted imaging. The scan times of 3D-QALAS with and without CS were 5:56 and 11:11, respectively. For healthy volunteers, brain volumetry and myelin estimation were performed based on the measured T1, T2, and PD. For patients with multiple sclerosis, the mean T1, T2, PD, and the amount of myelin in plaques and contralateral normal-appearing white matter (NAWM) were measured. Simple linear regression analysis and Bland-Altman analysis were performed for each metric obtained from the datasets with and without CS. To compare overall image quality and structural delineations on synthetic and conventional contrast-weighted images, case-control randomized reading sessions were performed by 2 neuroradiologists in a blinded manner. The linearity of both phantom and volunteer measurements in T1, T2, and PD values obtained with and without CS was very strong (R2 = 0.9901-1.000). The tissue segmentation obtained with and without CS also had high linearity (R2 = 0.987-0.999). The quantitative tissue values of the plaques and NAWM obtained with CS showed high linearity with those without CS (R2 = 0.967-1.000). There were no significant differences in overall image quality between synthetic contrast-weighted images obtained with and without CS (P = 0.17-0.99). Multiparametric imaging of the whole brain based on 3D-QALAS can be accelerated using CS while preserving tissue quantitative values, tissue segmentation, and quality of synthetic images.

A 128Gb 1-bit/Cell 96-Word-Line-Layer 3D Flash Memory to Improve the Random Read Latency With tProg = 75 μs and tR = 4 μs

A 128-Gb 1-bit/cell 3-D flash memory chip has been developed with 96-word-line-layer technology. A novel chip floorplan architecture with less time constants of wordline and bitline realizes fast read access time. A newly introduced program sequence achieves for higher reliability with less read-retry even after write/erase cycles. External VPP supply (12 V), current mode reference distribution, and auto temperature code refresh are also adopted to boost the performance of the chip. A new duty cycle corrector succeeds in obtaining wider unit interval of the DQS. Consequently, the proposed chip has a read access latency of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4~\mu \text{s}$ </tex-math></inline-formula> and a program time of 75 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{s}$ </tex-math></inline-formula> , which is 12–13 and 4–5 times faster than conventional 3-D flash memory with the same technology [Maejima <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> , (2018)]. The random read latency (tRRL) is estimated to be less than 50 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{s}$ </tex-math></inline-formula> , which enables in the reduction of total read access time of solid state drive (SSD) system.

Boafft: Distributed Deduplication for Big Data Storage in the Cloud

As data progressively grows within data centers, the cloud storage systems continuously facechallenges in saving storage capacity and providing capabilities necessary to move big data within an acceptable time frame. In this paper, we present the Boafft, a cloud storage system with distributed deduplication. The Boafft achieves scalable throughput and capacity usingmultiple data servers to deduplicate data in parallel, with a minimal loss of deduplication ratio. Firstly, the Boafft uses an efficient data routing algorithm based on data similarity that reduces the network overhead by quickly identifying the storage location. Secondly, the Boafft maintains an in-memory similarity indexing in each data server that helps avoid a large number of random disk reads and writes, which in turn accelerates local data deduplication. Thirdly, the Boafft constructs hot fingerprint cache in each data server based on access frequency, so as to improve the data deduplication ratio. Our comparative analysis with EMC's stateful routing algorithm reveals that the Boafft can provide a comparatively high deduplication ratio with a low network bandwidth overhead. Moreover, the Boafft makes better usage of the storage space, with higher read/write bandwidth and good load balance.

Achieving equity in diabetes care in BAME communities in Hampshire, UK: the DEBI‐H project

Achieving equity in diabetes care in <scp>BAME</scp> communities in Hampshire, <scp>UK</scp>: the <scp>DEBI‐H</scp> project

MzMLb: A Future-Proof Raw Mass Spectrometry Data Format Based on Standards-Compliant mzML and Optimized for Speed and Storage Requirements.

With ever-increasing amounts of data produced by mass spectrometry (MS) proteomics and metabolomics, and the sheer volume of samples now analyzed, the need for a common open format possessing both file size efficiency and faster read/write speeds has become paramount to drive the next generation of data analysis pipelines. The Proteomics Standards Initiative (PSI) has established a clear and precise extensible markup language (XML) representation for data interchange, mzML, receiving substantial uptake; nevertheless, storage and file access efficiency has not been the main focus. We propose an HDF5 file format “mzMLb” that is optimized for both read/write speed and storage of the raw mass spectrometry data. We provide an extensive validation of the write speed, random read speed, and storage size, demonstrating a flexible format that with or without compression is faster than all existing approaches in virtually all cases, while with compression is comparable in size to proprietary vendor file formats. Since our approach uniquely preserves the XML encoding of the metadata, the format implicitly supports future versions of mzML and is straightforward to implement: mzMLb’s design adheres to both HDF5 and NetCDF4 standard implementations, which allows it to be easily utilized by third parties due to their widespread programming language support. A reference implementation within the established ProteoWizard toolkit is provided.

Perceived Satisfaction of A&amp;E Staff on the Level of Integration of Patient Care Operations with Emergency Ambulance Service, Sri Lanka

Integration in operations between 1990 Suwa Seriya ambulance service and A&amp;E services of hospitals in Sri Lanka was assessed qualitatively using the Mini Focus Group Discussion (Mini-FGD) technique and applying a Likert scale survey. Agreed group-perceived satisfaction of the A&amp;E staff was recorded based on identified thematic areas. The technique of patient handling, reliability of vital sign recording, providing random blood sugar reading, and professional relationship with hospital staff of Suwa Seriya emergency ambulance service were the thematic areas that recorded positive agreed group-perceived satisfaction indicating a potential to broaden the integration process with the A&amp;E operations of the country. Quality of providing clinical history, administration of medication, and awareness of A&amp;E operations of emergency ambulance service indicated least agreed group-perceived satisfaction. Moreover, providing clinical notes and investigation records such as electrocardiogram, sharing feedback about the patient, prehospital triaging, and pre-alert system were found to be non-operational. The health-care policymakers of Sri Lanka should exploit the positives and explore the negatives of the A&amp;E staff perception on 1990 Suwa Seriya ambulance service, focusing on widening the integration process.

Probability-Based Address Translationfor Flash SSDs

Thanks to the advance of NAND scaling technologies, an ultra-scale SSD (e.g., $>$ > 100 TB) is introduced to markets. This rapid increase of SSD capacity, however, comes at the cost of more DRAM which resides in an SSD controller for logical-to-physical (L2P) address translation. Many have proposed various address translation algorithms to reduce DRAM, but they fail to provide short read latency, in particular when a workload has weak locality. This letter proposes a novel probability-based address translation algorithm, called ProbFTL . In contrast to existing translation techniques that maintain exact L2P mapping, ProbFTL employs a probability-based data structure, a bloom filter, for address translation. By leveraging a space-efficient nature of a bloom filter, ProbFTL reduces the amount of DRAM for address translation to 20 percent of the existing techniques. The read latency of ProbFTL is not affected from locality of a workload; ProbFTL guarantees a read amplification factor of 1.1 even under a random read workload. ProbFTL exhibits slightly worse garbage collection efficiency, but its write amplification factor is maintained sufficiently low.

Evaluating Encryption Algorithms for Sensitive Data Using Different Storage Devices

Sensitive data need to be protected from being stolen and read by unauthorized persons regardless of whether the data are stored in hard drives, flash memory, laptops, desktops, and other storage devices. In an enterprise environment where sensitive data is stored on storage devices, such as financial or military data, encryption is used in the storage device to ensure data confidentiality. Nowadays, the SSD-based NAND storage devices are favored over HDD and SSHD to store data because they offer increased performance and reduced access latency to the client. In this paper, the performance of different symmetric encryption algorithms is evaluated on HDD, SSHD, and SSD-based NAND MLC flash memory using two different storage encryption software. Based on the experiments we carried out, Advanced Encryption Standard (AES) algorithm on HDD outperforms Serpent and Twofish algorithms in terms of random read speed and write speed (both sequentially and randomly), whereas Twofish algorithm is slightly faster than AES in sequential reading on SSHD and SSD-based NAND MLC flash memory. By conducting full range of evaluative tests across HDD, SSHD, and SSD, our experimental results can give better idea for the storage consumers to determine which kind of storage device and encryption algorithm is suitable for their purposes. This will give them an opportunity to continuously achieve the best performance of the storage device and secure their sensitive data.

Automated performance tuning of distributed storage system based on deep reinforcement learning

Automated performance tuning is a tricky task for a large scale storage system. Traditional methods highly reply on experience of system administrators and cannot adapt to changes of working load and system configurations. Reinforcement learning is a promising machine learning paradigm which learns an optimized strategy from the trials and errors between agents and environments. Combining with the strong feature learning capability of deep learning, deep reinforcement learning has showed its success in many fields. We implemented a performance parameter tuning engine based on deep reinforcement learning for Lustre file system, a distributed file system widely used in HEP data centres. Three reinforcement learning algorithms: Deep Q-learning, A2C, and PPO are enabled in the tuning engine. Experiments show that, in a small test bed, with IOzone workload, this method can increase the random read throughput by about 30% compared to default settings of Lustre. In the future, it is possible to apply this method to other parameter tuning use cases of data centre operations.

File Type and Access Pattern Aware Buffer Cache Management for Rendering Systems

Rendering is the process of generating high-resolution images by software, which is widely used in animation, video games and visual effects in movies. Although rendering is a computation-intensive job, we observe that storage accesses may become another performance bottleneck in desktop-rendering systems. In this article, we present a new buffer cache management scheme specialized for rendering systems. Unlike general-purpose computing systems, rendering systems exhibit specific file access patterns, and we show that this results in significant performance degradation in the buffer cache system. To cope with this situation, we collect various file input/output (I/O) traces of rendering workloads and analyze their access patterns. The results of this analysis show that file I/Os in rendering processes consist of long loops for configuration, short loops for texture input, random reads for input, and single-writes for output. Based on this observation, we propose a new buffer cache management scheme for improving the storage performance of rendering systems. Experimental results show that the proposed scheme improves the storage I/O performance by an average of 19% and a maximum of 55% compared to the conventional buffer cache system.