SEARCH · Searcharxiv

Results for “cs.OS”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 397 records · Page 22Linked to original sources

arXiv2019

Nova -- A rainbow cloud over the Alps

A pooled and shared on-demand Infrastructure as a Service (IaaS), based on the Openstack software suite, was rolled out on the Grenoble university campus in 2018 and updated in 2019.We present the methods used to deploy and manage the infrastructure: racadm and preseed for basic system installation, then Kolla for Openstack deployment. This latter solution, based on containers for each service, enables a centralised and logged configuration (GitLab) of controllers and calculation nodes. The solution is the benchmark solution for a reproducible deployment of Openstack. We have been able to expand our cloud easily with new nodes. The change in version of the basic OS was also successfully tested despite a few small hitches... As security is a key element in the proper operation of this type of shared service, each project has been made watertight and its data perfectly isolated from other projects, thanks to the encryption of all network flows in VXLANs.This OpenStack infotainment platform is operational. What is it all for? For example, our first users use the Jupyter Notebook through the provision of Jupyterhub servers (web portal); the Distributed Health Assessment IT System (SIDES project); the continuous integration in connection with the GitLab platform; the test for the Kubernetes container scheduler or the calculation and visualisation software, etc. Highly varied uses that other platforms had difficulty offering.Nova, a new platform, was born.

cs.DC↗

arXiv2019

Faster than Flash: An In-Depth Study of System Challenges for Emerging Ultra-Low Latency SSDs

Emerging storage systems with new flash exhibit ultra-low latency (ULL) that can address performance disparities between DRAM and conventional solid state drives (SSDs) in the memory hierarchy. Considering the advanced low-latency characteristics, different types of I/O completion methods (polling/hybrid) and storage stack architecture (SPDK) are proposed. While these new techniques are expected to take costly software interventions off the critical path in ULL-applied systems, unfortunately no study exists to quantitatively analyze system-level characteristics and challenges of combining such newly-introduced techniques with real ULL SSDs. In this work, we comprehensively perform empirical evaluations with 800GB ULL SSD prototypes and characterize ULL behaviors by considering a wide range of I/O path parameters, such as different queues and access patterns. We then analyze the efficiencies and challenges of the polled-mode and hybrid polling I/O completion methods (added into Linux kernels 4.4 and 4.10, respectively) and compare them with the efficiencies of a conventional interrupt-based I/O path. In addition, we revisit the common expectations of SPDK by examining all the system resources and parameters. Finally, we demonstrate the challenges of ULL SSDs in a real SPDK-enabled server-client system. Based on the performance behaviors that this study uncovers, we also discuss several system implications, which are required to take a full advantage of ULL SSD in the future.

cs.OS↗

arXiv2020

Dispel: Byzantine SMR with Distributed Pipelining

Byzantine State Machine Replication (SMR) is a long studied topic that received increasing attention recently with the advent of blockchains as companies are trying to scale them to hundreds of nodes. Byzantine SMRs try to increase throughput by either reducing the latency of consensus instances that they run sequentially or by reducing the number of replicas that send messages to others in order to reduce the network usage. Unfortunately, the former approach makes use of resources in burst whereas the latter requires CPU-intensive authentication mechanisms. In this paper, we propose a new Byzantine SMR called Dispel (Distributed Pipeline) that allows any node to distributively start new consensus instances whenever they detect sufficient resources locally. We evaluate the performance of Dispel within a single datacenter and across up to 380 machines over 3 continents by comparing it against four other SMRs. On 128 nodes, Dispel speeds up HotStuff, the Byzantine fault tolerant SMR being integrated within Facebook's blockchain, by more than 12 times. In addition, we also test Dispel under isolated and correlated failures and show that the Dispel distributed design is more robust than HotStuff. Finally, we evaluate Dispel in a cryptocurrency application with Bitcoin transactions and show that this SMR is not the bottleneck.

cs.CR↗

arXiv2020

ARM Pointer Authentication based Forward-Edge and Backward-Edge Control Flow Integrity for Kernels

Code reuse attacks are still big threats to software and system security. Control flow integrity is a promising technique to defend against such attacks. However, its effectiveness has been weakened due to the inaccurate control flow graph and practical strategy to trade security for performance. In recent years, CPU vendors have integrated hardware features as countermeasures. For instance, ARM Pointer Authentication (PA in short) was introduced in ARMV8-A architecture. It can efficiently generate an authentication code for an address, which is encoded in the unused bits of the address. When the address is de-referenced, the authentication code is checked to ensure its integrity. Though there exist systems that adopt PA to harden user programs, how to effectively use PA to protect OS kernels is still an open research question. In this paper, we shed lights on how to leverage PA to protect control flows, including function pointers and return addresses, of Linux kernel. Specifically, to protect function pointers, we embed authentication code into them, track their propagation and verify their values when loading from memory or branching to targets. To further defend against the pointer substitution attack, we use the function pointer address as its context, and take a clean design to propagate the address by piggybacking it into the pointer value. We have implemented a prototype system with LLVM to identify function pointers, add authentication code and verify function pointers by emitting new machine instructions. We applied this system to Linux kernel, and solved numerous practical issues, e.g., function pointer comparison and arithmetic operations. The security analysis shows that our system can protect all function pointers and return addresses in Linux kernel.

cs.CR↗

arXiv2020

Virtual Gang based Scheduling of Real-Time Tasks on Multicore Platforms

We propose a virtual-gang based parallel real-time task scheduling approach for multicore platforms. Our approach is based on the notion of a virtual-gang, which is a group of parallel real-time tasks that are statically linked and scheduled together by a gang scheduler. We present a light-weight intra-gang synchronization framework, called RTG-Sync, and virtual gang formation algorithms that provide strong temporal isolation and high real-time schedulability in scheduling real-time tasks on multicore. We evaluate our approach both analytically, with generated tasksets against state-of-the-art approaches, and empirically with a case-study involving real-world workloads on a real embedded multicore platform. The results show that our approach provides simple but powerful compositional analysis framework, achieves better analytic schedulability, especially when the effect of interference is considered, and is a practical solution for COTS multicore platforms.

cs.OS↗

arXiv2020

Runtime Verification of Linux Kernel Security Module

The Linux kernel is one of the most important Free/Libre Open Source Software (FLOSS) projects. It is installed on billions of devices all over the world, which process various sensitive, confidential or simply private data. It is crucial to establish and prove its security properties. This work-in-progress paper presents a method to verify the Linux kernel for conformance with an abstract security policy model written in the Event-B specification language. The method is based on system call tracing and aims at checking that the results of system call execution do not lead to accesses that violate security policy requirements. As a basis for it, we use an additional Event-B specification of the Linux system call interface that is formally proved to satisfy all the requirements of the security policy model. In order to perform the conformance checks we use it to reproduce intercepted system calls and verify accesses.

cs.SE↗

arXiv2020

Online Scheduling with Makespan Minimization: State of the Art Results, Research Challenges and Open Problems

Online scheduling has been a well studied and challenging research problem over the last five decades since the pioneering work of Graham with immense practical significance in various applications such as interactive parallel processing, routing in communication networks, distributed data management, client-server communications, traffic management in transportation, industrial manufacturing and production. In this problem, a sequence of jobs is received one by one in order by the scheduler for scheduling over a number of machines. On arrival of a job, the scheduler assigns the job irrevocably to a machine before the availability of the next job with an objective to minimize the completion time of the scheduled jobs. This paper highlights the state of the art contributions for online scheduling of a sequence of independent jobs on identical and uniform related machines with a special focus on preemptive and non-preemptive processing formats by considering makespan minimization as the optimality criterion. We present the fundamental aspects of online scheduling from a beginner's perspective along with a background of general scheduling framework. Important competitive analysis results obtained by well-known deterministic and randomized online scheduling algorithms in the literature are presented along with research challenges and open problems. Two of the emerging recent trends such as resource augmentation and semi-online scheduling are discussed as a motivation for future research work.

cs.OS↗

arXiv2020

On Schedulability Analysis of EDF Scheduling by Considering Suspension as Blocking

During the execution of a job, it may suspend itself, i.e., its computation ceases to process until certain activities are complete to be resumed. This paper provides a counterexample of the schedulability analysis by Devi in Euromicro Conference on Real-Time Systems (ECRTS) in 2003, which is the only existing suspension-aware analysis specialized for uniprocessor systems when preemptive earliest-deadline-first (EDF) is applied for scheduling dynamic selfsuspending tasks.

cs.OS↗

arXiv2020

A New Fairness Model based on User's Objective for Multi-user Multi-processor Online Scheduling

Resources of a multi-user system in multi-processor online scheduling are shared by competing users in which fairness is a major performance criterion for resource allocation. Fairness ensures equality in resource sharing among the users. According to our knowledge, fairness based on the user's objective has neither been comprehensively studied nor a formal fairness model has been well defined in the literature. This motivates us to explore and define a new model to ensure algorithmic fairness with quantitative performance measures based on optimization of the user's objective. In this paper, we propose a new model for fairness in Multi-user Multi-processor Online Scheduling Problem(MUMPOSP). We introduce and formally define quantitative fairness measures based on user's objective by optimizing makespan for individual user in our proposed fairness model. We also define the unfairness of deprived users and absolute fairness of an algorithm. We obtain lower bound results for the absolute fairness for m identical machines with equal length jobs. We show that our proposed fairness model can serve as a framework for measuring algorithmic fairness by considering various optimality criteria such as flow time and sum of completion times.

cs.DS↗

arXiv2020

SPARTA: A Divide and Conquer Approach to Address Translation for Accelerators

Virtual memory (VM) is critical to the usability and programmability of hardware accelerators. Unfortunately, implementing accelerator VM efficiently is challenging because the area and power constraints make it difficult to employ the large multi-level TLBs used in general-purpose CPUs. Recent research proposals advocate a number of restrictions on virtual-to-physical address mappings in order to reduce the TLB size or increase its reach. However, such restrictions are unattractive because they forgo many of the original benefits of traditional VM, such as demand paging and copy-on-write. We propose SPARTA, a divide and conquer approach to address translation. SPARTA splits the address translation into accelerator-side and memory-side parts. The accelerator-side translation hardware consists of a tiny TLB covering only the accelerator's cache hierarchy (if any), while the translation for main memory accesses is performed by shared memory-side TLBs. Performing the translation for memory accesses on the memory side allows SPARTA to overlap data fetch with translation, and avoids the replication of TLB entries for data shared among accelerators. To further improve the performance and efficiency of the memory-side translation, SPARTA logically partitions the memory space, delegating translation to small and efficient per-partition translation hardware. Our evaluation on index-traversal accelerators shows that SPARTA virtually eliminates translation overhead, reducing it by over 30x on average (up to 47x) and improving performance by 57%. At the same time, SPARTA requires minimal accelerator-side translation hardware, reduces the total number of TLB entries in the system, gracefully scales with memory size, and preserves all key VM functionalities.

cs.AR↗

arXiv2019

AppStreamer: Reducing Storage Requirements of Mobile Games through Predictive Streaming

Storage has become a constrained resource on smartphones. Gaming is a popular activity on mobile devices and the explosive growth in the number of games coupled with their growing size contributes to the storage crunch. Even where storage is plentiful, it takes a long time to download and install a heavy app before it can be launched. This paper presents AppStreamer, a novel technique for reducing the storage requirements or startup delay of mobile games, and heavy mobile apps in general. AppStreamer is based on the intuition that most apps do not need the entirety of its files (images, audio and video clips, etc.) at any one time. AppStreamer can, therefore, keep only a small part of the files on the device, akin to a "cache", and download the remainder from a cloud storage server or a nearby edge server when it predicts that the app will need them in the near future. AppStreamer continuously predicts file blocks for the near future as the user uses the app, and fetches them from the storage server before the user sees a stall due to missing resources. We implement AppStreamer at the Android file system layer. This ensures that the apps require no source code or modification, and the approach generalizes across apps. We evaluate AppStreamer using two popular games: Dead Effect 2, a 3D first-person shooter, and Fire Emblem Heroes, a 2D turn-based strategy role-playing game. Through a user study, 75% and 87% of the users respectively find that AppStreamer provides the same quality of user experience as the baseline where all files are stored on the device. AppStreamer cuts down the storage requirement by 87% for Dead Effect 2 and 86% for Fire Emblem Heroes.

cs.OS↗

arXiv2020

Intel Page Modification Logging, a hardware virtualization feature: study and improvement for virtual machine working set estimation

Intel Page Modification Logging (PML) is a novel hardware feature for tracking virtual machine (VM) accessed memory pages. This task is essential in today's data centers since it allows, among others, checkpointing, live migration and working set size (WSS) estimation. Relying on the Xen hypervisor, this paper studies PML from three angles: power consumption, efficiency, and performance impact on user applications. Our findings are as follows. First, PML does not incur any power consumption overhead. Second, PML reduces by up to 10.18% both VM live migration and checkpointing time. Third, PML slightly reduces by up to 0.95% the performance degradation on applications incurred by live migration and checkpointing. Fourth, PML however does not allow accurate WSS estimation because read accesses are not tracked and hot pages cannot be identified. A naive extension of PML for addressing these limitations could lead to severe performance degradation (up to 34.8%) for the VM whose WSS is computed. This paper presents Page Reference Logging (PRL), a smart extension of PML for allowing both read and write accesses to be tracked. It does this without impacting user VMs. The paper also presents a WSS estimation system which leverages PRL and shows how this algorithm can be integrated into a data center which implements memory overcommitment. We implement PRL and the WSS estimation system under Gem5, a very popular hardware simulator. The evaluation results validate the accuracy of PRL in the estimation of WSS. They also show that PRL incurs no performance degradation for user VMs.

cs.DC↗

arXiv2020

A Recurrent Neural Network Based Patch Recommender for Linux Kernel Bugs

Software bugs in a production environment have an undesirable impact on quality of service, unplanned system downtime, and disruption in good customer experience, resulting in loss of revenue and reputation. Existing approaches to automated software bug repair focuses on known bug templates detected using static code analysis tools and test suites, and in automatic generation of patch code for these bugs. We describe the typical bug fixing process employed in the Linux kernel, and motivate the need for a new automated tool flow to fix bugs. We present an initial design of such an automated tool that uses Recurrent Neural Network (RNN) based Natural Language Processing to generate patch recommendations from user generated bug reports. At the 50th percentile of the test bugs, the correct patch occurs within the top 11.5 patch recommendations output by the model. Further, we present a Linux kernel developer's assessment of the quality of patches recommended for new unresolved kernel bugs.

cs.SE↗

arXiv2020

Safe and Efficient Remote Application Code Execution on Disaggregated NVM Storage with eBPF

With rapid improvements in NVM storage devices, the performance bottleneck is gradually shifting to the network, thus giving rise to the notion of "data movement wall". To reduce the amount of data movement over the network, researchers have proposed near-data computing by shipping operations and compute-extensions closer to storage devices. However, running arbitrary, user-provided extensions in a shared, disaggregated storage environment presents multiple challenges regarding safety, isolation, and performance. Instead of approaching this problem from scratch, in this work we make a case for leveraging the Linux kernel eBPF framework to program disaggregated NVM storage devices. eBPF offers a safe, verifiable, and high-performance way of executing untrusted, user-defined code in a shared runtime. In this paper, we describe our experiences building a first prototype that supports remote operations on storage using eBPF, discuss the limitations of our approach, and directions for addressing them.

cs.DC↗

arXiv2020

Bringing Inter-Thread Cache Benefits to Federated Scheduling -- Extended Results & Technical Report

Multiprocessor scheduling of hard real-time tasks modeled by directed acyclic graphs (DAGs) exploits the inherent parallelism presented by the model. For DAG tasks, a node represents a request to execute an object on one of the available processors. In one DAG task, there may be multiple execution requests for one object, each represented by a distinct node. These distinct execution requests offer an opportunity to reduce their combined cache overhead through coordinated scheduling of objects as threads within a parallel task. The goal of this work is to realize this opportunity by incorporating the cache-aware BUNDLE-scheduling algorithm into federated scheduling of sporadic DAG task sets. This is the first work to incorporate instruction cache sharing into federated scheduling. The result is a modification of the DAG model named the DAG with objects and threads (DAG-OT). Under the DAG-OT model, descriptions of nodes explicitly include their underlying executable object and number of threads. When possible, nodes assigned the same executable object are collapsed into a single node; joining their threads when BUNDLE-scheduled. Compared to the DAG model, the DAG-OT model with cache-aware scheduling reduces the number of cores allocated to individual tasks by approximately 20 percent in the synthetic evaluation and up to 50 percent on a novel parallel computing platform implementation. By reducing the number of allocated cores, the DAG-OT model is able to schedule a subset of previously infeasible task sets.

cs.DC↗

arXiv2020

Fissile Locks

Classic test-and-test (TS) mutual exclusion locks are simple, and enjoy high performance and low latency of ownership transfer under light or no contention. However, they do not scale gracefully under high contention and do not provide any admission order guarantees. Such concerns led to the development of scalable queue-based locks, such as a recent Compact NUMA-aware (CNA) lock, a variant of another popular queue-based MCS lock. CNA scales well under load and provides certain admission guarantees, but has more complicated lock handover operations than TS and incurs higher latencies at low contention. We propose Fissile locks, which capture the most desirable properties of both TS and CNA. A Fissile lock consists of two underlying locks: a TS lock, which serves as a fast path, and a CNA lock, which serves as a slow path. The key feature of Fissile locks is the ability of threads on the fast path to bypass threads enqueued on the slow path, and acquire the lock with less overhead than CNA. Bypass is bounded (by a tunable parameter) to avoid starvation and ensure long-term fairness. The result is a highly scalable NUMA-aware lock with progress guarantees that performs like TS at low contention and like CNA at high contention.

cs.OS↗

arXiv2020

Co-Optimizing Performance and Memory FootprintVia Integrated CPU/GPU Memory Management, anImplementation on Autonomous Driving Platform

Cutting-edge embedded system applications, such as self-driving cars and unmanned drone software, are reliant on integrated CPU/GPU platforms for their DNNs-driven workload, such as perception and other highly parallel components. In this work, we set out to explore the hidden performance implication of GPU memory management methods of integrated CPU/GPU architecture. Through a series of experiments on micro-benchmarks and real-world workloads, we find that the performance under different memory management methods may vary according to application characteristics. Based on this observation, we develop a performance model that can predict system overhead for each memory management method based on application characteristics. Guided by the performance model, we further propose a runtime scheduler. By conducting per-task memory management policy switching and kernel overlapping, the scheduler can significantly relieve the system memory pressure and reduce the multitasking co-run response time. We have implemented and extensively evaluated our system prototype on the NVIDIA Jetson TX2, Drive PX2, and Xavier AGX platforms, using both Rodinia benchmark suite and two real-world case studies of drone software and autonomous driving software.

cs.DC↗

arXiv2020

FLIC: A Distributed Fog Cache for City-Scale Applications

We present FLIC, a distributed software data caching framework for fogs that reduces network traffic and latency. FLICis targeted toward city-scale deployments of cooperative IoT devices in which each node gathers and shares data with surrounding devices. As machine learning and other data processing techniques that require large volumes of training data are ported to low-cost and low-power IoT systems, we expect that data analysis will be moved away from the cloud. Separation from the cloud will reduce reliance on power-hungry centralized cloud-based infrastructure. However, city-scale deployments of cooperative IoT devices often connect to the Internet with cellular service, in which service charges are proportional to network usage. IoT system architects must be clever in order to keep costs down in these scenarios. To reduce the network bandwidth required to operate city-scale deployments of cooperative IoT systems, FLIC implements a distributed cache on the IoT nodes in the fog. FLIC allows the IoT network to share its data without repetitively interacting with a simple cloud storage service reducing calls out to a backing store. Our results displayed a less than 2% miss rate on reads. Thus, allowing for only 5% of requests needing the backing store. We were also able to achieve more than 50% reduction in bytes transmitted per second.

cs.DC↗