Differential privacy is a promising privacy-preserving paradigm for statistical query processing over sensitive data. It works by injecting random noise into each query result, such that it is provably hard for the adversary to infer the presence or absence of any individual record from the published noisy results. The main objective in differentially private query processing is to maximize the accuracy of the query results, while satisfying the privacy guarantees. Previous work, notably \cite{LHR+10}, has suggested that with an appropriate strategy, processing a batch of correlated queries as a whole achieves considerably higher accuracy than answering them individually. However, to our knowledge there is currently no practical solution to find such a strategy for an arbitrary query batch; existing methods either return strategies of poor quality (often worse than naive methods) or require prohibitively expensive computations for even moderately large domains. Motivated by this, we propose low-rank mechanism (LRM), the first practical differentially private technique for answering batch linear queries with high accuracy. LRM works for both exact (i.e., $ε$-) and approximate (i.e., ($ε$, $δ$)-) differential privacy definitions. We derive the utility guarantees of LRM, and provide guidance on how to set the privacy parameters given the user's utility expectation. Extensive experiments using real data demonstrate that our proposed method consistently outperforms state-of-the-art query processing solutions under differential privacy, by large margins.
Workloads that comb through vast amounts of data are gaining importance in the sciences. These workloads consist of "needle in a haystack" queries that are long running and data intensive so that query throughput limits performance. To maximize throughput for data-intensive queries, we put forth LifeRaft: a query processing system that batches queries with overlapping data requirements. Rather than scheduling queries in arrival order, LifeRaft executes queries concurrently against an ordering of the data that maximizes data sharing among queries. This decreases I/O and increases cache utility.…
Differential privacy is a promising privacy-preserving paradigm for statistical query processing over sensitive data. It works by injecting random noise into each query result, such that it is provably hard for the adversary to infer the presence or absence of any individual record from the published noisy results. The main objective in differentially private query processing is to maximize the accuracy of the query results, while satisfying the privacy guarantees. Previous work, notably the matrix mechanism, has suggested that processing a batch of correlated queries as a whole can potentiall…
Many real-world applications operate on dynamic graphs that undergo rapid changes in their topological structure over time. However, it is challenging to design dynamic algorithms that are capable of supporting such graph changes efficiently. To circumvent the challenge, we propose a batch-dynamic framework for answering distance queries, which combines offline labelling and online searching to leverage the advantages from both sides - accelerating query processing through a partial distance labelling that is of limited size but provides a good approximation to bound online searches. We devise…
Most work on query optimization has concentrated on loop-free queries. However, data science and machine learning workloads today typically involve recursive or iterative computation. In this work, we propose a novel framework for optimizing recursive queries using methods from program synthesis. In particular, we introduce a simple yet powerful optimization rule called the "FGH-rule" which aims to find a faster way to evaluate a recursive program. The solution is found by making use of powerful tools, such as a program synthesizer, an SMT-solver, and an equality saturation system. We demonstr…
Differential privacy is a rigorous privacy condition achieved by randomizing query answers. This paper develops efficient algorithms for answering multiple queries under differential privacy with low error. We pursue this goal by advancing a recent approach called the matrix mechanism, which generalizes standard differentially private mechanisms. This new mechanism works by first answering a different set of queries (a strategy) and then inferring the answers to the desired workload of queries. Although a few strategies are known to work well on specific workloads, finding the strategy which m…
In practice, differentially private data releases are designed to support a variety of applications. A data release is fit for use if it meets target accuracy requirements for each application. In this paper, we consider the problem of answering linear queries under differential privacy subject to per-query accuracy constraints. Existing practical frameworks like the matrix mechanism do not provide such fine-grained control (they optimize total error, which allows some query answers to be more accurate than necessary, at the expense of other queries that become no longer useful). Thus, we desi…
Differential privacy is a promising privacy-preserving paradigm for statistical query processing over sensitive data. It works by injecting random noise into each query result, such that it is provably hard for the adversary to infer the presence or absence of any individual record from the published noisy results. The main objective in differentially private query processing is to maximize the accuracy of the query results, while satisfying the privacy guarantees. Previous work, notably \cite{LHR+10}, has suggested that with an appropriate strategy, processing a batch of correlated queries as…
Recent advances in neural networks (NNs) have enabled automatic querying of large volumes of video data with high accuracy. While these deep NNs can produce accurate annotations of an object's position and type in video, they are computationally expensive and require complex, imperative deployment code to answer queries. Prior work uses approximate filtering to reduce the cost of video analytics, but does not handle two important classes of queries, aggregation and limit queries; moreover, these approaches still require complex code to deploy. To address the computational and usability challen…
Answering queries over a federation of SPARQL endpoints requires combining data from more than one data source. Optimizing queries in such scenarios is particularly challenging not only because of (i) the large variety of possible query execution plans that correctly answer the query but also because (ii) there is only limited access to statistics about schema and instance data of remote sources. To overcome these challenges, most federated query engines rely on heuristics to reduce the space of possible query execution plans or on dynamic programming strategies to produce optimal plans. Never…
Graph simulation (using graph schemata or data guides) has been successfully proposed as a technique for adding structure to semistructured data. Design patterns for description (such as meta-classes and homomorphisms between schema layers), which are prominent in the object-oriented programming community, constitute a generalization of this graph simulation approach. In this paper, we show description applicable to a wide range of data models that have some notion of object (-identity), and propose to turn it into a data model primitive much like, say, inheritance. We argue that such an exten…
$k$d-trees are widely used in parallel databases to support efficient neighborhood/similarity queries. Supporting parallel updates to $k$d-trees is therefore an important operation. In this paper, we present BDL-tree, a parallel, batch-dynamic implementation of a $k$d-tree that allows for efficient parallel $k$-NN queries over dynamically changing point sets. BDL-trees consist of a log-structured set of $k$d-trees which can be used to efficiently insert or delete batches of points in parallel with polylogarithmic depth. Specifically, given a BDL-tree with $n$ points, each batch of $B$ updates …
In the last decade, many business applications have moved into the cloud. In particular, the "database-as-a-service" paradigm has become mainstream. While existing multi-tenant data management systems focus on single-tenant query processing, we believe that it is time to rethink how queries can be processed across multiple tenants in such a way that we do not only gain more valuable insights, but also at minimal cost. As we will argue in this paper, standard SQL semantics are insufficient to process cross-tenant queries in an unambiguous way, which is why existing systems use other, expensive …
The Sloan Digital Sky Survey science database is approaching 2TB. While the vast majority of queries normally execute in seconds or minutes, this interactive execution time can be disproportionately increased by a small fraction of queries that take hours or days to run; either because they require non-index scans of the largest tables or because they request very large result sets. In response to this, we added a multi-queue job submission and tracking system. The transfer of very large result sets from queries over the network is another serious problem. Statistics suggested that much of thi…
We consider accelerating machine learning (ML) inference queries on unstructured datasets. Expensive operators such as feature extractors and classifiers are deployed as user-defined functions(UDFs), which are not penetrable with classic query optimization techniques such as predicate push-down. Recent optimization schemes (e.g., Probabilistic Predicates or PP) assume independence among the query predicates, build a proxy model for each predicate offline, and rewrite a new query by injecting these cheap proxy models in the front of the expensive ML UDFs. In such a manner, unlikely inputs that …
Differential privacy is a robust privacy standard that has been successfully applied to a range of data analysis tasks. Despite much recent work, optimal strategies for answering a collection of correlated queries are not known. We study the problem of devising a set of strategy queries, to be submitted and answered privately, that will support the answers to a given workload of queries. We propose a general framework in which query strategies are formed from linear combinations of counting queries, and we describe an optimal method for deriving new query answers from the answers to the strate…
Hybrid complex analytics workloads typically include (i) data management tasks (joins, selections, etc. ), easily expressed using relational algebra (RA)-based languages, and (ii) complex analytics tasks (regressions, matrix decompositions, etc.), mostly expressed in linear algebra (LA) expressions. Such workloads are common in many application areas, including scientific computing, web analytics, and business recommendation. Existing solutions for evaluating hybrid analytical tasks - ranging from LA-oriented systems, to relational systems (extended to handle LA operations), to hybrid systems …
Characteristic sets (CS) organize RDF triples based on the set of properties characterizing their subject nodes. This concept is recently used in indexing techniques, as it can capture the implicit schema of RDF data. While most CS-based approaches yield significant improvements in space and query performance, they fail to perform well in the presence of schema heterogeneity, i.e., when the number of CSs becomes very large, resulting in a highly partitioned data organization. In this paper, we address this problem by introducing a novel technique, for merging CSs based on their hierarchical st…
In this work we describe the High-Dimensional Matrix Mechanism (HDMM), a differentially private algorithm for answering a workload of predicate counting queries. HDMM represents query workloads using a compact implicit matrix representation and exploits this representation to efficiently optimize over (a subset of) the space of differentially private algorithms for one that is unbiased and answers the input query workload with low expected error. HDMM can be deployed for both $ε$-differential privacy (with Laplace noise) and $(ε, δ)$-differential privacy (with Gaussian noise), although the cor…
The optimization of query execution plans is known to be crucial for reducing the query execution time. In particular, query optimization has been studied thoroughly for relational databases over the past decades. Recently, the Resource Description Framework (RDF) became popular for publishing data on the Web. As a consequence, federations composed of different data models like RDF and relational databases evolved. One type of these federations are Semantic Data Lakes where every data source is kept in its original data model and semantically annotated with ontologies or controlled vocabularie…
The Sloan Digital Sky Survey (SDSS) science database describes over 140 million objects and is over 1.5 TB in size. The SDSS Catalog Archive Server (CAS) provides several levels of query interface to the SDSS data via the SkyServer website. Most queries execute in seconds or minutes. However, some queries can take hours or days, either because they require non-index scans of the largest tables, or because they request very large result sets, or because they represent very complex aggregations of the data. These "monster queries" not only take a long time, they also affect response times for ev…
Differentially private algorithms for answering sets of predicate counting queries on a sensitive database have many applications. Organizations that collect individual-level data, such as statistical agencies and medical institutions, use them to safely release summary tabulations. However, existing techniques are accurate only on a narrow class of query workloads, or are extremely slow, especially when analyzing more than one or two dimensions of the data. In this work we propose HDMM, a new differentially private algorithm for answering a workload of predicate counting queries, that is espe…
As data sets grow in size, analytics applications struggle to get instant insight into large datasets. Modern applications involve heavy batch processing jobs over large volumes of data and at the same time require efficient ad-hoc interactive analytics on temporary data. Existing solutions, however, typically focus on one of these two aspects, largely ignoring the need for synergy between the two. Consequently, interactive queries need to re-iterate costly passes through the entire dataset (e.g., data loading) that may provide meaningful return on investment only when data is queried over a l…
Differential computation (DC) is a highly general incremental computation/view maintenance technique that can maintain the output of an arbitrary and possibly recursive dataflow computation upon changes to its base inputs. As such, it is a promising technique for graph database management systems (GDBMS) that support continuous recursive queries over dynamic graphs. Although differential computation can be highly efficient for maintaining these queries, it can require a prohibitively large amount of memory. This paper studies how to reduce the memory overhead of DC with the goal of increasing …
The emergence of Linked Data on the WWW has spawned research interest in an online execution of declarative queries over this data. A particularly interesting approach is traversal-based query execution which fetches data by traversing data links and, thus, is able to make use of up-to-date data from initially unknown data sources. The downside of this approach is the delay before the query engine completes a query execution. In this paper, we address this problem by proposing an approach to return as many elements of the result set as soon as possible. The basis of this approach is a traversa…
Data provenance is essential for debugging query results, auditing data in cloud environments, and explaining outputs of Big Data analytics. A well-established technique is to represent provenance as annotations on data and to instrument queries to propagate these annotations to produce results annotated with provenance. However, even sophisticated optimizers are often incapable of producing efficient execution plans for instrumented queries, because of their inherent complexity and unusual structure. Thus, while instrumentation enables provenance support for databases without requiring any mo…
While the Web of Data in principle offers access to a wide range of interlinked data, the architecture of the Semantic Web today relies mostly on the data providers to maintain access to their data through SPARQL endpoints. Several studies, however, have shown that such endpoints often experience downtime, meaning that the data they maintain becomes inaccessible. While decentralized systems based on Peer-to-Peer (P2P) technology have previously shown to increase the availability of knowledge graphs, even when a large proportion of the nodes fail, processing queries in such a setup can be an ex…
We study ranked enumeration of join-query results according to very general orders defined by selective dioids. Our main contribution is a framework for ranked enumeration over a class of dynamic programming problems that generalizes seemingly different problems that had been studied in isolation. To this end, we extend classic algorithms that find the k-shortest paths in a weighted graph. For full conjunctive queries, including cyclic ones, our approach is optimal in terms of the time to return the top result and the delay between results. These optimality properties are derived for the widel…
Apache Flink is an open-source system for scalable processing of batch and streaming data. Flink does not natively support efficient processing of spatial data streams, which is a requirement of many applications dealing with spatial data. Besides Flink, other scalable spatial data processing platforms including GeoSpark, Spatial Hadoop, etc. do not support streaming workloads and can only handle static/batch workloads. To fill this gap, we present GeoFlink, which extends Apache Flink to support spatial data types, indexes and continuous queries over spatial data streams. To enable the efficie…
Recent advances in computer vision-in the form of deep neural networks-have made it possible to query increasing volumes of video data with high accuracy. However, neural network inference is computationally expensive at scale: applying a state-of-the-art object detector in real time (i.e., 30+ frames per second) to a single video requires a $4000 GPU. In response, we present NoScope, a system for querying videos that can reduce the cost of neural network video analysis by up to three orders of magnitude via inference-optimized model search. Given a target video, object to detect, and referenc…
National University of Singapore（新加坡国立大学）；GrabTaxi Holdings（GrabTaxi控股公司）
The use of deep learning models for forecasting the resource consumption patterns of SQL queries have recently been a popular area of study. With many companies using cloud platforms to power their data lakes for large scale analytic demands, these models form a critical part of the pipeline in managing cloud resource provisioning. While these models have demonstrated promising accuracy, training them over large scale industry workloads are expensive. Space inefficiencies of encoding techniques over large numbers of queries and excessive padding used to enforce shape consistency across diverse…
The B-tree is a fundamental secondary index structure that is widely used for answering one-dimensional range reporting queries. Given a set of $N$ keys, a range query can be answered in $O(\log_B \nm + \frac{K}{B})$ I/Os, where $B$ is the disk block size, $K$ the output size, and $M$ the size of the main memory buffer. When keys are inserted or deleted, the B-tree is updated in $O(\log_B N)$ I/Os, if we require the resulting changes to be committed to disk right away. Otherwise, the memory buffer can be used to buffer the recent updates, and changes can be written to disk in batches, which si…
University of Ulm（乌尔姆大学）；National Technical University of Athens（雅典国家技术大学）
The SPARQL query language is currently being extended by the World Wide Web Consortium (W3C) with so-called entailment regimes. An entailment regime defines how queries are evaluated under more expressive semantics than SPARQLs standard simple entailment, which is based on subgraph matching. The queries are very expressive since variables can occur within complex concepts and can also bind to concept or role names. In this paper, we describe a sound and complete algorithm for the OWL Direct Semantics entailment regime. We further propose several novel optimizations such as strategies for deter…
