arXiv subjects

Joseph M. Hellerstein

Publications and source records attributed to Joseph M. Hellerstein.

51 records · Page 3Linked to original sources

arXiv2019

Looking Back at Postgres

This is a recollection of the UC Berkeley Postgres project, which was led by Mike Stonebraker from the mid-1980's to the mid-1990's. The article was solicited for Stonebraker's Turing Award book, as one of many personal/historical recollections. As a result it focuses on Stonebraker's design ideas and leadership. But Stonebraker was never a coder, and he stayed out of the way of his development team. The Postgres codebase was the work of a team of brilliant students and the occasional university "staff programmers" who had little more experience (and only slightly more compensation) than the students. I was lucky to join that team as a student during the latter years of the project. I got helpful input on this writeup from some of the more senior students on the project, but any errors or omissions are mine. If you spot any such, please contact me and I will try to fix them.

cs.DB↗

arXiv2018

Serverless Computing: One Step Forward, Two Steps Back

Serverless computing offers the potential to program the cloud in an autoscaling, pay-as-you go manner. In this paper we address critical gaps in first-generation serverless computing, which place its autoscaling potential at odds with dominant trends in modern computing: notably data-centric and distributed computing, but also open source and custom hardware. Put together, these gaps make current serverless offerings a bad fit for cloud innovation and particularly bad for data systems innovation. In addition to pinpointing some of the main shortfalls of current serverless architectures, we raise a set of challenges we believe must be met to unlock the radical potential that the cloud---with its exabytes of storage and millions of cores---should offer to innovative developers.

cs.DC↗

arXiv2018

Eliminating Boundaries in Cloud Storage with Anna

In this paper, we describe how we extended a distributed key-value store called Anna into an elastic, multi-tier service for the cloud. In its extended form, Anna is designed to overcome the narrow cost-performance limitations typical of current cloud storage systems. We describe three key aspects of Anna's new design: multi-master selective replication of hot keys, a vertical tiering of storage layers with different cost-performance tradeoffs, and horizontal elasticity of each tier to add and remove nodes in response to load dynamics. Anna's policy engine uses these mechanisms to balance service-level objectives around cost, latency and fault tolerance. Experimental results explore the behavior of Anna's mechanisms and policy, exhibiting orders of magnitude efficiency improvements over both commodity cloud KVS services and research systems.

cs.DB↗

arXiv2017

A Berkeley View of Systems Challenges for AI

With the increasing commoditization of computer vision, speech recognition and machine translation systems and the widespread deployment of learning-based back-end technologies such as digital advertising and intelligent infrastructures, AI (Artificial Intelligence) has moved from research labs to production. These changes have been made possible by unprecedented levels of data and computation, by methodological advances in machine learning, by innovations in systems software and architectures, and by the broad accessibility of these technologies. The next generation of AI systems promises to accelerate these developments and increasingly impact our lives via frequent interactions and making (often mission-critical) decisions on our behalf, often in highly personalized contexts. Realizing this promise, however, raises daunting challenges. In particular, we need AI systems that make timely and safe decisions in unpredictable environments, that are robust against sophisticated adversaries, and that can process ever increasing amounts of data across organizations and individuals without compromising confidentiality. These challenges will be exacerbated by the end of the Moore's Law, which will constrain the amount of data these technologies can store and process. In this paper, we propose several open research directions in systems, architectures, and security that can address these challenges and help unlock AI's potential to improve lives and society.

cs.AI↗

arXiv2016

Naughton's Wisconsin Bibliography: A Brief Guide

Over nearly three decades at the University of Wisconsin, Jeff Naughton has left an indelible mark on computer science. He has been a global leader of the database research field, deepening its core and pushing its boundaries. Many of Naughton's ideas were translated directly into practice in commercial and open-source systems. But software comes and goes. In the end, it is the ideas themselves that have had impact, ideas written down in papers. Naughton has been a prolific scholar over the last thirty years, with over 175 publications in his bibliography, covering a wide range of topics. This document does not attempt to enumerate or even summarize the wealth of ideas that Naughton has published over the course of his academic career--the task is too daunting. Instead, the best this short note aims to do is to serve as a rough map of the territory: something to help other researchers navigate the wide spaces of Naughton's work.

cs.GL↗

arXiv2015

Asynchronous Complex Analytics in a Distributed Dataflow Architecture

Scalable distributed dataflow systems have recently experienced widespread adoption, with commodity dataflow engines such as Hadoop and Spark, and even commodity SQL engines routinely supporting increasingly sophisticated analytics tasks (e.g., support vector machines, logistic regression, collaborative filtering). However, these systems' synchronous (often Bulk Synchronous Parallel) dataflow execution model is at odds with an increasingly important trend in the machine learning community: the use of asynchrony via shared, mutable state (i.e., data races) in convex programming tasks, which has---in a single-node context---delivered noteworthy empirical performance gains and inspired new research into asynchronous algorithms. In this work, we attempt to bridge this gap by evaluating the use of lightweight, asynchronous state transfer within a commodity dataflow engine. Specifically, we investigate the use of asynchronous sideways information passing (ASIP) that presents single-stage parallel iterators with a Volcano-like intra-operator iterator that can be used for asynchronous information passing. We port two synchronous convex programming algorithms, stochastic gradient descent and the alternating direction method of multipliers (ADMM), to use ASIPs. We evaluate an implementation of ASIPs within on Apache Spark that exhibits considerable speedups as well as a rich set of performance trade-offs in the use of these asynchronous algorithms.

cs.DB↗

arXiv2015

Putting Logic-Based Distributed Systems on Stable Grounds

In the Declarative Networking paradigm, Datalog-like languages are used to express distributed computations. Whereas recently formal operational semantics for these languages have been developed, a corresponding declarative semantics has been lacking so far. The challenge is to capture precisely the amount of nondeterminism that is inherent to distributed computations due to concurrency, networking delays, and asynchronous communication. This paper shows how a declarative, model-based semantics can be obtained by simply using the well-known stable model semantics for Datalog with negation. We show that the model-based semantics matches previously proposed formal operational semantics.

cs.LO↗

arXiv2014

Coordination Avoidance in Database Systems (Extended Version)

Minimizing coordination, or blocking communication between concurrently executing operations, is key to maximizing scalability, availability, and high performance in database systems. However, uninhibited coordination-free execution can compromise application correctness, or consistency. When is coordination necessary for correctness? The classic use of serializable transactions is sufficient to maintain correctness but is not necessary for all applications, sacrificing potential scalability. In this paper, we develop a formal framework, invariant confluence, that determines whether an application requires coordination for correct execution. By operating on application-level invariants over database states (e.g., integrity constraints), invariant confluence analysis provides a necessary and sufficient condition for safe, coordination-free execution. When programmers specify their application invariants, this analysis allows databases to coordinate only when anomalies that might violate invariants are possible. We analyze the invariant confluence of common invariants and operations from real-world database systems (i.e., integrity constraints) and applications and show that many are invariant confluent and therefore achievable without coordination. We apply these results to a proof-of-concept coordination-avoiding database prototype and demonstrate sizable performance gains compared to serializable execution, notably a 25-fold improvement over prior TPC-C New-Order performance on a 200 server cluster.

cs.DB↗

arXiv2013

Blazes: Coordination Analysis for Distributed Programs

Distributed consistency is perhaps the most discussed topic in distributed systems today. Coordination protocols can ensure consistency, but in practice they cause undesirable performance unless used judiciously. Scalable distributed architectures avoid coordination whenever possible, but under-coordinated systems can exhibit behavioral anomalies under fault, which are often extremely difficult to debug. This raises significant challenges for distributed system architects and developers. In this paper we present Blazes, a cross-platform program analysis framework that (a) identifies program locations that require coordination to ensure consistent executions, and (b) automatically synthesizes application-specific coordination code that can significantly outperform general-purpose techniques. We present two case studies, one using annotated programs in the Twitter Storm system, and another using the Bloom declarative language.

cs.DC↗

arXiv2013

Highly Available Transactions: Virtues and Limitations (Extended Version)

To minimize network latency and remain online during server failures and network partitions, many modern distributed data storage systems eschew transactional functionality, which provides strong semantic guarantees for groups of multiple operations over multiple data items. In this work, we consider the problem of providing Highly Available Transactions (HATs): transactional guarantees that do not suffer unavailability during system partitions or incur high network latency. We introduce a taxonomy of highly available systems and analyze existing ACID isolation and distributed data consistency guarantees to identify which can and cannot be achieved in HAT systems. This unifies the literature on weak transactional isolation, replica consistency, and highly available systems. We analytically and experimentally quantify the availability and performance benefits of HATs--often two to three orders of magnitude over wide-area networks--and discuss their necessary semantic compromises.

cs.DB↗

arXiv2013

Learning and Verifying Quantified Boolean Queries by Example

To help a user specify and verify quantified queries --- a class of database queries known to be very challenging for all but the most expert users --- one can question the user on whether certain data objects are answers or non-answers to her intended query. In this paper, we analyze the number of questions needed to learn or verify qhorn queries, a special class of Boolean quantified queries whose underlying form is conjunctions of quantified Horn expressions. We provide optimal polynomial-question and polynomial-time learning and verification algorithms for two subclasses of the class qhorn with upper constant limits on a query's causal density.

cs.DB↗

arXiv2012

Distributed GraphLab: A Framework for Machine Learning in the Cloud

While high-level data parallel frameworks, like MapReduce, simplify the design and implementation of large-scale data processing systems, they do not naturally or efficiently support many important data mining and machine learning algorithms and can lead to inefficient learning systems. To help fill this critical void, we introduced the GraphLab abstraction which naturally expresses asynchronous, dynamic, graph-parallel computation while ensuring data consistency and achieving a high degree of parallel performance in the shared-memory setting. In this paper, we extend the GraphLab framework to the substantially more challenging distributed setting while preserving strong data consistency guarantees. We develop graph based extensions to pipelined locking and data versioning to reduce network congestion and mitigate the effect of network latency. We also introduce fault tolerance to the GraphLab abstraction using the classic Chandy-Lamport snapshot algorithm and demonstrate how it can be easily implemented by exploiting the GraphLab abstraction itself. Finally, we evaluate our distributed implementation of the GraphLab abstraction on a large Amazon EC2 deployment and show 1-2 orders of magnitude performance gains over Hadoop-based implementations.

cs.DB↗

arXiv2012

Probabilistically Bounded Staleness for Practical Partial Quorums

Data store replication results in a fundamental trade-off between operation latency and data consistency. In this paper, we examine this trade-off in the context of quorum-replicated data stores. Under partial, or non-strict quorum replication, a data store waits for responses from a subset of replicas before answering a query, without guaranteeing that read and write replica sets intersect. As deployed in practice, these configurations provide only basic eventual consistency guarantees, with no limit to the recency of data returned. However, anecdotally, partial quorums are often "good enough" for practitioners given their latency benefits. In this work, we explain why partial quorums are regularly acceptable in practice, analyzing both the staleness of data they return and the latency benefits they offer. We introduce Probabilistically Bounded Staleness (PBS) consistency, which provides expected bounds on staleness with respect to both versions and wall clock time. We derive a closed-form solution for versioned staleness as well as model real-time staleness for representative Dynamo-style systems under internet-scale production workloads. Using PBS, we measure the latency-consistency trade-off for partial quorum systems. We quantitatively demonstrate how eventually consistent systems frequently return consistent data within tens of milliseconds while offering significant latency benefits.

cs.DB↗

arXiv2010

Multiresolution Cube Estimators for Sensor Network Aggregate Queries

In this work we present in-network techniques to improve the efficiency of spatial aggregate queries. Such queries are very common in a sensornet setting, demanding more targeted techniques for their handling. Our approach constructs and maintains multi-resolution cube hierarchies inside the network, which can be constructed in a distributed fashion. In case of failures, recovery can also be performed with in-network decisions. In this paper we demonstrate how in-network cube hierarchies can be used to summarize sensor data, and how they can be exploited to improve the efficiency of spatial aggregate queries. We show that query plans over our cube summaries can be computed in polynomial time, and we present a PTIME algorithm that selects the minimum number of data requests that can compute the answer to a spatial query. We further extend our algorithm to handle optimization over multiple queries, which can also be done in polynomial time. We discuss enriching cube hierarchies with extra summary information, and present an algorithm for distributed cube construction. Finally we investigate node and area failures, and algorithms to recover query results.

cs.DB↗

arXiv2010

GraphLab: A New Framework for Parallel Machine Learning

Designing and implementing efficient, provably correct parallel machine learning (ML) algorithms is challenging. Existing high-level parallel abstractions like MapReduce are insufficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges. By targeting common patterns in ML, we developed GraphLab, which improves upon abstractions like MapReduce by compactly expressing asynchronous iterative algorithms with sparse computational dependencies while ensuring data consistency and achieving a high degree of parallel performance. We demonstrate the expressiveness of the GraphLab framework by designing and implementing parallel versions of belief propagation, Gibbs sampling, Co-EM, Lasso and Compressed Sensing. We show that using GraphLab we can achieve excellent parallel performance on large scale real-world problems.

cs.LG↗