Performance Measurements of Supercomputing and Cloud Storage Solutions
Michael Jones, Jeremy Kepner, William Arcand, David Bestor, Bill Bergeron, Vijay Gadepally, Michael Houle, Matthew Hubbell, Peter Michaleas, Andrew Prout, Alber… 展开作者
Michael Jones, Jeremy Kepner, William Arcand, David Bestor, Bill Bergeron, Vijay Gadepally, Michael Houle, Matthew Hubbell, Peter Michaleas, Andrew Prout, Albert Reuther, Siddharth Samsi, Paul Monticiollo
Increasing amounts of data from varied sources, particularly in the fields of machine learning and graph analytics, are causing storage requirements to grow rapidly. A variety of technologies exist for storing and sharing these data, ranging from parallel file systems used by supercomputers to distributed block storage systems found in clouds. Relatively few comparative measurements exist to inform decisions about which storage systems are best suited for particular tasks. This work provides these measurements for two of the most popular storage technologies: Lustre and Amazon S3. Lustre is an open-source, high performance, parallel file system used by many of the largest supercomputers in the world. Amazon's Simple Storage Service, or S3, is part of the Amazon Web Services offering, and offers a scalable, distributed option to store and retrieve data from anywhere on the Internet. Parallel processing is essential for achieving high performance on modern storage systems. The performance tests used span the gamut of parallel I/O scenarios, ranging from single-client, single-node Amazon S3 and Lustre performance to a large-scale, multi-client test designed to demonstrate the capabilities of a modern storage appliance under heavy load. These results show that, when parallel I/O is used correctly (i.e., many simultaneous read or write processes), full network bandwidth performance is achievable and ranged from 10 gigabits/s over a 10 GigE S3 connection to 0.35 terabits/s using Lustre on a 1200 port 10 GigE switch. These results demonstrate that S3 is well-suited to sharing vast quantities of data over the Internet, while Lustre is well-suited to processing large quantities of data locally.
Background: Photo-realistic terapixel visualization is computationally intensive and to date there have been no such visualizations of urban digital twins, the few terapixel visualizations that exist have looked towards space rather than earth. Objective: our aims are: creating a scalable cloud supercomputer software architecture for visualization; a photo-realistic terapixel 3D visualization of urban IoT data supporting daily updates; a rigorous evaluation of cloud supercomputing for our application. Method: we migrated the Blender Cycles path tracer to the public cloud within a new software …
As primary provider for research computing services at the University of Minnesota, the Minnesota Supercomputing Institute (MSI) has long been responsible for serving the needs of a user-base numbering in the thousands. In recent years, MSI---like many other HPC centers---has observed a growing need for self-service, on-demand, data-intensive research, as well as the emergence of many new controlled-access datasets for research purposes. In light of this, MSI constructed a new on-premise cloud service, named Stratus, which is architected from the ground up to easily satisfy data-use agreements…
Descartes Labs（笛卡尔实验室）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
We present our experiences using cloud computing to support data-intensive analytics on satellite imagery for commercial applications. Drawing from our background in high-performance computing, we draw parallels between the early days of clustered computing systems and the current state of cloud computing and its potential to disrupt the HPC market. Using our own virtual file system layer on top of cloud remote object storage, we demonstrate aggregate read bandwidth of 230 gigabytes per second using 512 Google Compute Engine (GCE) nodes accessing a USA multi-region standard storage bucket. Thi…
The latest generation of radio astronomy interferometers will conduct all sky surveys with data products consisting of petabytes of spectral line data. Traditional approaches to identifying and parameterising the astrophysical sources within this data will not scale to datasets of this magnitude, since the performance of workstations will not keep up with the real-time generation of data. For this reason, it is necessary to employ high performance computing systems consisting of a large number of processors connected by a high-bandwidth network. In order to make use of such supercomputers subs…
Growing data volumes and velocities are driving exciting new methods across the sciences in which data analytics and machine learning are increasingly intertwined with research. These new methods require new approaches for scientific computing in which computation is mobile, so that, for example, it can occur near data, be triggered by events (e.g., arrival of new data), or be offloaded to specialized accelerators. They also require new design approaches in which monolithic applications can be decomposed into smaller components, that may in turn be executed separately and on the most efficient…
We have used the Montage image mosaic engine to investigate the cost and performance of processing images on the Amazon EC2 cloud, and to inform the requirements that higher-level products impose on provenance management technologies. We will present a detailed comparison of the performance of Montage on the cloud and on the Abe high performance cluster at the National Center for Supercomputing Applications (NCSA). Because Montage generates many intermediate products, we have used it to understand the science requirements that higher-level products impose on provenance management technologies.…
Storage systems are essential building blocks for cloud computing infrastructures. Although high performance storage servers are the ultimate solution for cloud storage, the implementation of inexpensive storage system remains an open issue. To address this problem, the efficient cloud storage system is implemented with inexpensive and commodity computer nodes that are organized into PC cluster based datacenter. Hadoop Distributed File System (HDFS) is an open source cloud based storage platform and designed to be deployed in low-cost hardware. PC Cluster based Cloud Storage System is implemen…
In past years, cloud storage systems saw an enormous rise in usage. However, despite their popularity and importance as underlying infrastructure for more complex cloud services, today's cloud storage systems do not account for compliance with regulatory, organizational, or contractual data handling requirements by design. Since legislation increasingly responds to rising data protection and privacy concerns, complying with data handling requirements becomes a crucial property for cloud storage systems. We present PRADA, a practical approach to account for compliance with data handling require…
Cloud storage plays an important role in social computing. This paper aims to develop a cloud storage management system for mobile devices to support an extended set of file operations. Because of the limit of storage, bandwidth, power consumption, and other resource restrictions, most existing cloud storage apps for smartphones do not keep local copies of files. This efficient design, however, limits the application capacities. In this paper, we attempt to extend the available file operations for cloud storage service to better serve smartphone users. We develop an efficient and secure file m…
Enterprises are driving towards less cost, more availability, agility, managed risk - all of which is accelerated towards Cloud Computing. Cloud is not a particular product, but a way of delivering IT services that are consumable on demand, elastic to scale up and down as needed, and follow a pay-for-usage model. Out of the three common types of cloud computing service models, Infrastructure as a Service (IaaS) is a service model that provides servers, computing power, network bandwidth and Storage capacity, as a service to their subscribers. Cloud can relate to many things but without the fun…
Design and architecture of cloud storage system plays a vital role in cloud computing infrastructure in order to improve the storage capacity as well as cost effectiveness. Usually cloud storage system provides users to efficient storage space with elasticity feature. One of the challenges of cloud storage system is difficult to balance the providing huge elastic capacity of storage and investment of expensive cost for it. In order to solve this issue in the cloud storage infrastructure, low cost PC cluster based storage server is configured to be activated for large amount of data to provide …
Implementing big data storage at scale is a complex and arduous task that requires an advanced infrastructure. With the rise of public cloud computing, various big data management services can be readily leveraged. As a critical part of Twitter's "Project Partly Cloudy", the cold storage data and analytics systems are being moved to the public cloud. This paper showcases our approach in designing a scalable big data storage and analytics management framework using BigQuery in Google Cloud Platform while ensuring security, privacy, and data protection. The paper also discusses the limitations o…
With the advent of virtualization technology, cloud computing realizes on-demand computing. The capability of dynamic resource provisioning is a fundamental driving factor for users to adopt the cloud technology. The aspect is important for cloud service providers to optimize the expense for running the infrastructure as well. Despite many technological advances in related areas, however, it is still the case that the infrastructure providers must decide hardware configuration before deploying a cloud infrastructure, especially from the storage's perspective. This static nature of the storage …
Cloud computing is a powerful new technology that is widely used in the business world. Recently, we have been investigating the benefits it offers to scientific computing. We have used three workflow applications to compare the performance of processing data on the Amazon EC2 cloud with the performance on the Abe high-performance cluster at the National Center for Supercomputing Applications (NCSA). We show that the Amazon EC2 cloud offers better performance and value for processor- and memory-limited applications than for I/O-bound applications. We provide an example of how the cloud is well…
We describe a cloud based infrastructure that we have developed that is optimized for wide area, high performance networks and designed to support data mining applications. The infrastructure consists of a storage cloud called Sector and a compute cloud called Sphere. We describe two applications that we have built using the cloud and some experimental studies.
Due to its advantages over traditional data centers, there has been a rapid growth in the usage of cloud infrastructures. These include public clouds (e.g., Amazon EC2), or private clouds, such as clouds deployed using OpenStack. A common factor in many of the well known infrastructures, for example OpenStack and CloudStack, is that networked storage is used for storage of persistent data. However, traditional Big Data systems, including Hadoop, store data in commodity local storage for reasons of high performance and low cost. We present an architecture for supporting Hadoop on Openstack usin…
Today's cloud storage services must offer storage reliability and fast data retrieval for large amount of data without sacrificing storage cost. We present SEARS, a cloud-based storage system which integrates erasure coding and data deduplication to support efficient and reliable data storage with fast user response time. With proper association of data to storage server clusters, SEARS provides flexible mixing of different configurations, suitable for real-time and archival applications. Our prototype implementation of SEARS over Amazon EC2 shows that it outperforms existing storage systems i…
Cloud block storage systems support diverse types of applications in modern cloud services. Characterizing their I/O activities is critical for guiding better system designs and optimizations. In this paper, we present an in-depth comparative analysis of production cloud block storage workloads through the block-level I/O traces of billions of I/O requests collected from two production systems, Alibaba Cloud and Tencent Cloud Block Storage. We study their characteristics of load intensities, spatial patterns, and temporal patterns. We also compare the cloud block storage workloads with the not…
Hyper-converged cloud refers to an architecture that an operator runs compute and storage services on the same set of physical servers. Although the hyper-converged design comes with a number of benefits, it makes crucial operational tasks, such as capacity planning and cost analysis, fairly complicated. The problem becomes more onerous if we consider a complex distributed system, such as Ceph, for the cloud with the proliferation of SSD drives. In this paper, we aim to answer some of these questions based on comprehensive microbenchmarks, and consequently better understand the behavior of Cep…
In cloud data center, shared storage with good management is a main structure used for the storage of virtual machines (VM). In this paper, we proposed Hybrid VM storage (HVSTO), a privacy preserving shared storage system designed for the virtual machine storage in large-scale cloud data center. Unlike traditional shared storage, HVSTO adopts a distributed structure to preserve privacy of virtual machines, which are a threat in traditional centralized structure. To improve the performance of I/O latency in this distributed structure, we use a hybrid system to combine solid state disk and distr…
Cloud storage systems have been introduced to provide a scalable, secure, reliable, and highly available data storage environment for the organizations and end-users. Therefore, the service provider should grow in a geographical extent. Consequently, extensive storage service provision requires a replication mechanism. Replication imposes many costs on the cloud storage, including the synchronization, communications, storage, etc., costs among the replicas. Moreover, the synchronization process among replicas is a major challenge in cloud storage. Therefore, consistency can be defined as the c…
In cloud storage, the digital data is stored in logical storage pools, backed by heterogeneous physical storage media and computing infrastructure that are managed by a Cloud Service Provider (CSP). To balance the tradeoff between service performance and cost, CSPs often employ different storage tiers, for instance, cold storage and hot storage. Storing data in hot storage incurs high storage cost yet delivers low access latency, whereas cold storage is able to inexpensively store massive amounts of data and thus provides lower cost with higher latency. In this paper, we address a major challe…
We present CDStore, which disperses users' backup data across multiple clouds and provides a unified multi-cloud storage solution with reliability, security, and cost-efficiency guarantees. CDStore builds on an augmented secret sharing scheme called convergent dispersal, which supports deduplication by using deterministic content-derived hashes as inputs to secret sharing. We present the design of CDStore, and in particular, describe how it combines convergent dispersal with two-stage deduplication to achieve both bandwidth and storage savings and be robust against side-channel attacks. We eva…
We present a simple, efficient, and secure data-oblivious randomized shuffle algorithm. This is the first secure data-oblivious shuffle that is not based on sorting. Our method can be used to improve previous oblivious storage solutions for network-based outsourcing of data.
With the wide adoption of large-scale Internet services and big data, the cloud has become the ideal environment to satisfy the ever-growing storage demand, thanks to its seemingly limitless capacity, high availability and faster access time. In this context, data replication has been touted as the ultimate solution to improve data availability and reduce access time. However, replica placement systems usually need to migrate and create a large number of data replicas over time between and within data centers, incurring a large overhead in terms of network load and availability. In this paper,…
In this paper we present the Chelonia storage cloud middleware. It was designed to fill the requirements gap between those of large, sophisticated scientific collaborations which have adopted the grid paradigm for their distributed storage needs, and of corporate business communities which are gravitating towards the cloud paradigm. The similarities to and differences between Chelonia and several well-known grid- and cloud-based storage solutions are commented. The design of Chelonia has been chosen to optimize high reliability and scalability of an integrated system of heterogeneous, geograph…
It is very challenging part to keep safely all required data that are needed in many applications for user in cloud. Storing our data in cloud may not be fully trustworthy. Since client doesn't have copy of all stored data, he has to depend on Cloud Service Provider. But dynamic data operations, Read-Solomon and verification token construction methods don't tell us about total storage capacity of server allocated space before and after the data addition in cloud. So we have to introduce a new proposed system of efficient storage measurement and space comparison algorithm with time management f…
Cloud computing has been envisioned as the next generation architecture of IT Enterprise. Using Cloud Storage,users can remotely store their data and enjoy the on demand high quality applications and services from a shared pool of configurable computing resources, without the burden local copy data storage and maintenance. It moves the application of software data stored to the centralized large data centers, where the management of the data stored services may not be completely trusted. There are many new security challenges and the problems taken into account for ensuring the integrity of da…
Cloud computing provides scientists a platform that can deploy computation and data intensive applications without infrastructure investment. With excessive cloud resources and a decision support system, large generated data sets can be flexibly 1 stored locally in the current cloud, 2 deleted and regenerated whenever reused or 3 transferred to cheaper cloud service for storage. However, due to the pay for use model, the total application cost largely depends on the usage of computation, storage and bandwidth resources, hence cutting the cost of cloud based data storage becomes a big concern f…
First responders and other forward deployed essential workers can benefit from advanced analytics. Limited network access and software security requirements prevent the usage of standard cloud based microservice analytic platforms that are typically used in industry. One solution is to precompute a wide range of analytics as files that can be used with standard preinstalled software that does not require network access or additional software and can run on a wide range of legacy hardware. In response to the COVID-19 pandemic, this approach was tested for providing geo-spatial census data to al…
A common task in scientific computing is the derivation of data. This workflow extracts the most important information from large input data and stores it in smaller derived data objects. The derived data objects can then be used for further analysis tasks. Typically, those workflows use distributed storage and computing resources. A straightforward configuration of storage media would be low cost tape storage and higher cost disk storage. The large, infrequently accessed input data is stored on tape storage. The smaller, frequently accessed derived data is stored on disk storage. In a best ca…
SAGE (Percipient StorAGe for Exascale Data Centric Computing) is a European Commission funded project towards the era of Exascale computing. Its goal is to design and implement a Big Data/Extreme Computing (BDEC) capable infrastructure with associated software stack. The SAGE system follows a "storage centric" approach as it is capable of storing and processing large data volumes at the Exascale regime. SAGE addresses the convergence of Big Data Analysis and HPC in an era of next-generation data centric computing. This convergence is driven by the proliferation of massive data sources, such as…
Cloud object storage such as AWS S3 is cost-effective and highly elastic but relatively slow, while high-performance cloud storage such as AWS ElastiCache is expensive and provides limited elasticity. We present a new cloud storage service called ServerlessMemory, which stores data using the memory of serverless functions. ServerlessMemory employs a sliding-window-based memory management strategy inspired by the garbage collection mechanisms used in the programming language to effectively segregate hot/cold data and provides fine-grained elasticity, good performance, and a pay-per-access cost …
The ICT (Information Communication Technologies) ecosystem is estimated to be responsible, as of today, for 10% of the total worldwide energy demand - equivalent to the combined energy production of Germany and Japan. Cloud storage, mainly operated through large and densely-packed data centers, constitutes a non-negligible part of it. However, since the cloud is a fast-inflating market and the energy-efficiency of data centers is mostly an insensitive issue for the collectivity, its carbon footprint shows no signs of slowing down. In this paper, we analyze a novel paradigm for cloud storage (i…
Cloud storage is fast securing its role as a major repository for both consumers and business customers. Many companies now offer storage solutions, sometimes for free for limited amounts of capacity. We have surveyed the pricing plans of a selection of major cloud providers and compared them using the unit price as the means of comparison. All the providers, excepting Amazon, adopt a bundling pricing scheme; Amazon follows instead a block-declining pricing policy. We compare the pricing plans through a double approach: a pointwise comparison for each value of capacity, and an overall comparis…
As a Distributed Hash Table (DHT), Skip Graph routing overlays are exploited in several peer-to-peer (P2P) services, including P2P cloud storage. The fully decentralized replication algorithms that are applicable to the Skip Graph-based P2P cloud storage fail on improving the performance of the system with respect to both the availability of replicas as well as their response time. Additionally, they presume the system as homogeneous with respect to the nodes' latency distribution, availability behavior, bandwidth, or storage. In this paper, we propose Pyramid, which is the first fully decentr…
Network codes designed specifically for distributed storage systems have the potential to provide dramatically higher storage efficiency for the same availability. One main challenge in the design of such codes is the exact repair problem: if a node storing encoded information fails, in order to maintain the same level of reliability we need to create encoded information at a new node. One of the main open problems in this emerging area has been the design of simple coding schemes that allow exact and low cost repair of failed nodes and have high data rates. In particular, all prior known expl…
Achieving high performance in virtualized data centers requires both deploying high throughput storage clusters, i.e. based on Solid State Disks (SSDs), as well as optimally consolidating the workloads across storage nodes. Nowadays, the only practical solution for cloud storage providers to offer guaranteed performance is to grossly over-provision the storage nodes. The current workload scheduling mechanisms used in production do not have the intelligence to optimally allocate block storage volumes based on the performance of SSDs. In this paper, we introduce Serifos, an autonomous performanc…
We describe the design and implementation of a high performance cloud that we have used to archive, analyze and mine large distributed data sets. By a cloud, we mean an infrastructure that provides resources and/or services over the Internet. A storage cloud provides storage services, while a compute cloud provides compute services. We describe the design of the Sector storage cloud and how it provides the storage services required by the Sphere compute cloud. We also describe the programming paradigm supported by the Sphere compute cloud. Sector and Sphere are designed for analyzing large dat…
In the immediate future holographic technology will be available to store a very large amount of data in HVD (Holographic Versatile Disk) devices. This technology make extensive use of the WORM (Write-Once-Read-Many) paradigm: this means that such devices allow for a simultaneous and parallel reading of millions of volumetric pixels (i.e. voxels). This characteristic will make accessible wherever the acquired data from a telescope (or satellite) in a quite-simultaneous way. With the support of this new technology the aim of this paper is to identify the guidelines for the implementation of a d…
Cloud computing has demonstrated that processing very large datasets over commodity clusters can be done simply given the right programming model and infrastructure. In this paper, we describe the design and implementation of the Sector storage cloud and the Sphere compute cloud. In contrast to existing storage and compute clouds, Sector can manage data not only within a data center, but also across geographically distributed data centers. Similarly, the Sphere compute cloud supports User Defined Functions (UDF) over data both within a data center and across data centers. As a special case, Ma…
Given a cloud-based anonymous data storage system, there are two ways for managing the nodes involved in file transfers. One of them is using reputations and the other uses a micropayment system. In reputation-based approach, each node has a reputation associated with it, which is used as a currency or feedback collection for file exchange operations. There have been several attempts over the years to develop a strong and efficient reputation system that provides credibility, fairness, and accountability. One such attempt was the Free Haven Project that provides a strong foundation for cloud-b…
Cloud computing has emerged as a popular computing paradigm in recent years. However, today's cloud computing architectures often lack support for computer forensic investigations. A key task of digital forensics is to prove the presence of a particular file in a given storage system. Unfortunately, it is very hard to do so in a cloud given the black-box nature of clouds and the multi-tenant cloud models. In clouds, analyzing the data from a virtual machine instance or data stored in a cloud storage only allows us to investigate the current content of the cloud storage, but not the previous co…
As the cost-per-byte of storage systems dramatically decreases, SSDs are finding their ways in emerging cloud infrastructure. Similar trend is happening for main memory subsystem, as advanced DRAM technologies with higher capacity, frequency and number of channels are deploying for cloud-scale solutions specially for non-virtualized environment where cloud subscribers can exactly specify the configuration of underling hardware. Given the performance sensitivity of standard workloads to the memory hierarchy parameters, it is important to understand the role of memory and storage for data intens…
Cloud computing provides a powerful yet low-cost environment for distributed deep learning workloads. However, training complex deep learning models often requires accessing large amounts of data, which can easily exceed the capacity of local disks. Prior research often overlooks this training data problem by implicitly assuming that data is available locally or via low latency network-based data storage. Such implicit assumptions often do not hold in a cloud-based training environment, where deep learning practitioners create and tear down dedicated GPU clusters on demand, or do not have the …
As cars are ubiquitous they could play a major role in a next generation communication and computation framework. In the last years, the development of vehicle-to-vehicle communication and vehicle-to-infrastructure communication took huge steps forward and therefore gives us the tools to build "mobile computing service" on cars equipped with computation capabilities. Recently, several groups of researchers independently proposed the design of "vehicular clouds" that materializes the concept. In this paper, we introduce a new paradigm of the vehicular clouds, followed by a case study of data st…
From home appliances to industrial enterprises, the Information and Communication Technology (ICT) industry is revolutionizing the world. We are witnessing the emergence of new technologies (e.g, Cloud computing, Fog computing, Internet of Things (IoT), Artificial Intelligence (AI) and Block-chain) which proves the growing use of ICT (e,g. business, education, health, and home appliances), resulting in massive data generation. It is expected that more than 175 ZB data will be processed annually by 75 billion devices by 2025. The 5G technology (i.e. mobile communication technology) dramatically…
With the growing reliance on the ubiquitous availability of IT systems and services, these systems become more global, scaled, and complex to operate. To maintain business viability, IT service providers must put in place reliable and cost efficient operations support. Artificial Intelligence for IT Operations (AIOps) is a promising technology for alleviating operational complexity of IT systems and services. AIOps platforms utilize big data, machine learning and other advanced analytics technologies to enhance IT operations with proactive actionable dynamic insight. In this paper we share our…
Eliminating duplicate data in primary storage of clouds increases the cost-efficiency of cloud service providers as well as reduces the cost of users for using cloud services. Existing primary deduplication techniques either use inline caching to exploit locality in primary workloads or use post-processing deduplication running in system idle time to avoid the negative impact on I/O performance. However, neither of them works well in the cloud servers running multiple services or applications for the following two reasons: Firstly, the temporal locality of duplicate data writes may not exist i…
One key requirement for storage clouds is to be able to retrieve data quickly. Recent system measurements have shown that the data retrieving delay in storage clouds is highly variable, which may result in a long latency tail. One crucial idea to improve the delay performance is to retrieve multiple data copies by using parallel downloading threads. However, how to optimally schedule these downloading threads to minimize the data retrieving delay remains to be an important open problem. In this paper, we develop low-complexity thread scheduling policies for several important classes of data do…
Edge computing moves the computation closer to the data and the data closer to the user to overcome the high latency communication of cloud computing. Storage at the edge allows data access with high speeds that enable latency-sensitive applications in areas such as autonomous driving and smart grid. However, several distributed services are typically designed for the cloud and building an efficient edge-enabled storage system is challenging because of the distributed and heterogeneous nature of the edge and its limited resources. In this paper, we propose EdgeKV, a decentralized storage syste…
A benchmark study of modern distributed databases is an important source of information to select the right technology for managing data in the cloud-edge paradigms. To make the right decision, it is required to conduct an extensive experimental study on a variety of hardware infrastructures. While most of the state-of-the-art studies have investigated only response time and scalability of distributed databases, focusing on other various metrics (e.g., energy, bandwidth, and storage consumption) is essential to fully understand the resources consumption of the distributed databases. Also, exis…
Maastricht University Medical Centre+（马斯特里赫特大学医学中心+）；University Medical Center Groningen（格罗宁根大学医学中心）；University of Groningen（格罗宁根大学）
Artificial intelligence (AI), especially deep learning, requires vast amounts of data for training, testing, and validation. Collecting these data and the corresponding annotations requires the implementation of imaging biobanks that provide access to these data in a standardized way. This requires careful design and implementation based on the current standards and guidelines and complying with the current legal restrictions. However, the realization of proper imaging data collections is not sufficient to train, validate and deploy AI as resource demands are high and require a careful hybrid …
