اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدDo the Hard Stuff First: Scheduling Dependent Computations in Data-Analytics Clusters
73
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
We present a scheduler that improves cluster utilization and job completion times by packing tasks having multi-resource requirements and inter-dependencies. While the problem is algorithmically very hard, we achieve near-optimality on the job DAGs that appear in production clusters at a large enterprise and in benchmarks such as TPC-DS. A key insight is that carefully handling the long-running tasks and those with tough-to-pack resource needs will produce good-enough schedules. However, which subset of tasks to treat carefully is not clear (and intractable to discover). Hence, we offer a search procedure that evaluates various possibilities and outputs a preferred schedule order over tasks. An online component enforces the schedule orders desired by the various jobs running on the cluster. In addition, it packs tasks, overbooks the fungible resources and guarantees bounded unfairness for a variety of desirable fairness schemes. Relative to the state-of-the art schedulers, we speed up 50% of the jobs by over 30% each.
قيم البحث
اقرأ أيضاً
Nearly twenty years after the launch of AWS, it remains difficult for most developers to harness the enormous potential of the cloud. In this paper we lay out an agenda for a new generation of cloud programming research aimed at bringing research ide
as to programmers in an evolutionary fashion. Key to our approach is a separation of distributed programs into a PACT of four facets: Program semantics, Availablity, Consistency and Targets of optimization. We propose to migrate developers gradually to PACT programming by lifting familiar code into our more declarative level of abstraction. We then propose a multi-stage compiler that emits human-readable code at each stage that can be hand-tuned by developers seeking more control. Our agenda raises numerous research challenges across multiple areas including language design, query optimization, transactions, distributed consistency, compilers and program synthesis.
We increasingly live in a data-driven world, with diverse kinds of data distributed across many locations. In some cases, the datasets are collected from multiple locations, such as sensors (e.g., mobile phones and street cameras) spread throughout a
geographic region. The data may need to be analyzed close to where they are produced, particularly when the applications require low latency, high, low cost, user privacy, and regulatory constraints. In other cases, large datasets are distributed across public clouds, private clouds, or edge-cloud computing sites with more plentiful computation, storage, bandwidth, and energy resources. Often, some portion of the analysis may take place on the end-host or edge cloud (to respect user privacy and reduce the volume of data) while relying on remote clouds to complete the analysis (to leverage greater computation and storage resources). Wide-area data analytics is any analysis of data that is generated by, or stored at, geographically dispersed entities. Over the past few years, several parts of the computer science research community have started to explore effective ways to analyze data spread over multiple locations. In particular, several areas of systems research - including databases, distributed systems, computer networking, and security and privacy - have delved into these topics. These research subcommunities often focus on different aspects of the problem, consider different motivating applications and use cases, and design and evaluate their solutions differently. To address these challenges the Computing Community Consortium (CCC) convened a 1.5-day workshop focused on wide-area data analytics in October 2019. This report summarizes the challenges discussed and the conclusions generated at the workshop.
Delivering effective data analytics is of crucial importance to the interpretation of the multitude of biological datasets currently generated by an ever increasing number of high throughput techniques. Logic programming has much to offer in this are
a. Here, we detail advances that highlight two of the strengths of logical formalisms in developing data analytic solutions in biological settings: access to large relational databases and building analytical pipelines collecting graph information from multiple sources. We present significant advances on the bio_db package which serves biological databases as Prolog facts that can be served either by in-memory loading or via database backends. These advances include modularising the underlying architecture and the incorporation of datasets from a second organism (mouse). In addition, we introduce a number of data analytics tools that operate on these datasets and are bundled in the analysis package: bio_analytics. Emphasis in both packages is on ease of installation and use. We highlight the general architecture of our components based approach. An experimental graphical user interface via SWISH for local installation is also available. Finally, we advocate that biological data analytics is a fertile area which can drive further innovation in applied logic programming.
Training Deep Neural Networks (DNNs) is resource-intensive and time-consuming. While prior research has explored many different ways of reducing DNN training time, the impact of input data pipeline, i.e., fetching raw data items from storage and perf
orming data pre-processing in memory, has been relatively unexplored. This paper makes the following contributions: (1) We present the first comprehensive analysis of how the input data pipeline affects the training time of widely-used computer vision and audio Deep Neural Networks (DNNs), that typically involve complex data preprocessing. We analyze nine different models across three tasks and four datasets while varying factors such as the amount of memory, number of CPU threads, storage device, GPU generation etc on servers that are a part of a large production cluster at Microsoft. We find that in many cases, DNN training time is dominated by data stall time: time spent waiting for data to be fetched and preprocessed. (2) We build a tool, DS-Analyzer to precisely measure data stalls using a differential technique, and perform predictive what-if analysis on data stalls. (3) Finally, based on the insights from our analysis, we design and implement three simple but effective techniques in a data-loading library, CoorDL, to mitigate data stalls. Our experiments on a range of DNN tasks, models, datasets, and hardware configs show that when PyTorch uses CoorDL instead of the state-of-the-art DALI data loading library, DNN training time is reduced significantly (by as much as 5x on a single server).
The growing adoption of IoT devices in our daily life is engendering a data deluge, mostly private information that needs careful maintenance and secure storage system to ensure data integrity and protection. Also, the prodigious IoT ecosystem has pr
ovided users with opportunities to automate systems by interconnecting their devices and other services with rule-based programs. The cloud services that are used to store and process sensitive IoT data turn out to be vulnerable to outside threats. Hence, sensitive IoT data and rule-based programs need to be protected against cyberattacks. To address this important challenge, in this paper, we propose a framework to maintain confidentiality and integrity of IoT data and rule-based program execution. We design the framework to preserve data privacy utilizing Trusted Execution Environment (TEE) such as Intel SGX, and end-to-end data encryption mechanism. We evaluate the framework by executing rule-based programs in the SGX securely with both simulated and real IoT device data.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
