arXiv subjects

Lucas C. Villa Real

Publications and source records attributed to Lucas C. Villa Real.

4 recordsLinked to original sources

arXiv2022

On Approximate Sequencing Policies for Linear Storage Devices

This paper investigates sequencing policies for file reading requests in linear storage devices, such as magnetic tapes. Tapes are the technology of choice for long-term storage in data centers due to their low cost and reliability. However, their physical structure imposes challenges to data retrieval operations reflected in classic optimization and operations research problems. In this work, we provide a theoretical and numerical performance analysis of low-complexity algorithms under deterministic, stochastic, and online settings, which are key in practice due to their interpretability and the large scale of existing data services. In the deterministic setting, we show that traditional policies, such as first-in first-out (FIFO), have arbitrarily poor performance, and we develop and investigate new constant-factor approximations. For the stochastic setting, we present a fully polynomial-time approximation scheme that weighs files based on their access frequencies. Finally, we investigate an online extension and propose a new algorithm with constant competitive-factor guarantees. Our numerical analysis on synthetic and real-world data suggest that the proposed algorithms may significantly outperform policies currently adopted in practice with respect to average reading times.

cs.DS↗

arXiv2021

User-Defined Functions for HDF5

Scientific datasets are known for their challenging storage demands and the associated processing pipelines that transform their information. Some of those processing tasks include filtering, cleansing, aggregation, normalization, and data format translation -- all of which generate even more data. In this paper, we present an infrastructure for the HDF5 file format that enables dataset values to be populated on the fly: task-related scripts can be attached into HDF5 files and only execute when the dataset is read by an application. We provide details on the software architecture that supports user-defined functions (UDFs) and how it integrates with hardware accelerators and computational storage. Moreover, we describe the built-in security model that limits the system resources a UDF can access. Last, we present several use cases that show how UDFs can be used to extend scientific datasets in ways that go beyond the original scope of this work.

cs.DC↗

arXiv2018

Theoretical and Practical Aspects of the Linear Tape Scheduling Problem

Magnetic tapes have been playing a key role as means for storage of digital data for decades, and their unsurpassed cost-effectiveness still make them the technology of choice in several industries, such as media and entertainment. Tapes are mostly used for cold storage nowadays, and therefore the study of scheduling algorithms for read requests tailored for these devices has been largely neglected in the literature. In this article, we investigate the Linear Tape Scheduling Problem (LTSP), in which read requests associated with files stored on a single-tracked magnetic tape should be scheduled in a way that the sum of all response times are minimized. LTSP has many similarities with classical combinatorial optimization problems such as the Traveling Repairmen Problem and the Dial-a-Ride Problem restricted to the real line; nevertheless, significant differences on structural properties and strict time-limit constraints of real-world scenarios make LTSP challenging and interesting on its own. In this work, we investigate several properties and algorithms for LTSP and some of its extensions. The results allowed for the identification of 3-approximation algorithms for LTSP and efficient exact algorithms for some of its special cases. We also show that LTSPR, the version of the problem with heterogeneous release times for requests, is NP-complete. OLTSP, the online extension of LTSPR, does not admit c-competitive algorithms for any constant factor c, but we nevertheless introduce an algorithm for the problem and show through extensive computational experiments on synthetic and real-world datasets that different embodiments of the proposed strategy are computationally efficient and over-perform by orders of magnitude an algorithm being currently used by real-world tape file systems.

cs.DS↗

arXiv2018

Full Speed Ahead: 3D Spatial Database Acceleration with GPUs

Many industries rely on visual insights to support decision- making processes in their businesses. In mining, the analysis of drills and geological shapes, represented as 3D geometries, is an important tool to assist geologists on the search for new ore deposits. Aeronautics manipulate high-resolution geometries when designing a new aircraft aided by the numerical simulation of aerodynamics. In common, these industries require scalable databases that compute spatial relationships and measurements so that decision making can be conducted without lags. However, as we show in this study, most database systems either lack support for handling 3D geometries or show poor performance when given a sheer volume of data to work with. This paper presents a pluggable acceleration engine for spatial database systems that can improve the performance of spatial operations by more than 3000x through GPU offloading. We focus on the design and evaluation of our plug-in for the PostgreSQL database system.

cs.DC↗