Databases are very complex systems that require database system administrators to perform system tuning in order to achieve optimal performance. Memory tuning is vital to the performance of a database system because when the database workload exceeds its memory capacity, the results of the queries running on a system are delayed and can cause substantial user dissatisfaction. In order to solve this problem, this thesis presents a platform modeled after a closed control feedback loop to control the level of multi-query processing. Utilizing this platform provides two key assets. First, the system identification is acquired, which is one of two crucial steps involved in developing a closed feedback loop. Second, the platform provides a means to experimentally study database tuning problem and verify the effectiveness of research ideas related to database performance.

No description available.

Numerous applications gather increasing amounts of data, which have to be managed and queried. Different hardware developments help to meet this challenge. The grow-ing capacity of main memory enables database systems to keep all their data in memory. Additionally, the hardware landscape is becoming more diverse. A plethora of homo-geneous and heterogeneous co-processors is available, where heterogeneity refers not only to a different computing power, but also to different instruction set architectures. For instance, modern Intel® CPUs offer different instruction sets supporting the Single Instruction Multiple Data (SIMD) paradigm, e.g. SSE, AVX, and AVX512. Database systems have started to exploit SIMD to increase performance. However, this is still a challenging task, because existing algorithms were mainly developed for scalar processing and because there is a huge variety of different instruction sets, which were never standardized and have no unified interface. This requires to completely rewrite the source code for porting a system to another hardware architecture, even if those archi-tectures are not fundamentally different and designed by the same company. Moreover, operations on large registers, which are the core principle of SIMD processing, behave counter-intuitively in several cases. This is especially true for analytical query process-ing, where different memory access patterns and data dependencies caused by the com-pression of data, challenge the limits of the SIMD principle. Finally, there are physical constraints to the use of such instructions affecting the CPU frequency scaling, which is further influenced by the use of multiple cores. This is because the supply power of a CPU is limited, such that not all transistors can be powered at the same time. Hence, there is a complex relationship between performance and power, and therefore also between performance and energy consumption. This thesis addresses the specific challenges, which are introduced by the application of SIMD in general, and the heterogeneity of SIMD ISAs in particular. Hence, the goal of this thesis is to exploit the potential of heterogeneous SIMD ISAs for increasing the performance as well as the energy-efficiency.

No description available.

"Continuous queries process real-time streaming data and output results in streams for a wide range of applications. Due to the fluctuating stream characteristics, a streaming database system needs to dynamically adapt query execution. This dissertation proposes novel solutions to continuous query adaptation in three core areas, namely dynamic query optimization, dynamic plan migration and partitioned query adaptation. Runtime query optimization needs to efficiently generate plans that satisfy both CPU and memory resource constraints. Existing work focus on minimizing intermediate query results, which decreases memory and CPU usages simultaneously. However, doing so cannot assure that both resource constraints are being satisfied, because memory and CPU can be either positively or negatively correlated. This part of the dissertation proposes efficient optimization strategies that utilize both types of correlations to search the entire query plan space in polynomial time when a typical exhaustive search would take at least exponential time. Extensive experimental evaluations have demonstrated the effectiveness of the proposed strategies. Dynamic plan migration is concerned with on-the-fly transition from one continuous plan to a semantically equivalent yet more efficient plan. It is a must to guarantee the continuation and repeatability of dynamic query optimization. However, this research area has been largely neglected in the current literature. The second part of this dissertation proposes migration strategies that dynamically migrate continuous queries while guaranteeing the integrity of the query results, meaning there are no missing, duplicate or incorrect results. The extensive experimental evaluations show that the proposed strategies vary significantly in terms of output rates and memory usages given distinct system configurations and stream workloads. Partitioned query processing is effective to process continuous queries with large stateful operators in a distributed system. Dynamic load redistribution is necessary to balance uneven workload across machines due to changing stream properties. However, existing solutions generally assume static query plans without runtime query optimization. This part of the dissertation evaluates the benefits of applying query optimization in partitioned query processing and shows dramatic performance improvement of more than 300%. Several load balancing strategies are then proposed to consider the heterogeneity of plan shapes across machines caused by dynamic query optimization. The effectiveness of the proposed strategies is analyzed through extensive experiments using a cluster."

In the era of "big data revolution," marked by an exponential growth of information, extracting value from data enables analysts and businesses to address challenging problems such as drug discovery, fraud detection, and earthquake predictions. Multi-Criteria Decision Support (MCDS) queries are at the core of big-data analytics resulting in several classes of MCDS queries such as OLAP, Top-K, Pareto-optimal, and nearest neighbor queries. The intuitive nature of specifying multi-dimensional preferences has made Pareto-optimal queries, also known as skyline queries, popular. Existing skyline algorithms however do not address several crucial issues such as performing skyline evaluation over disparate sources, progressively generating skyline results, or robustly handling workload with multiple skyline over join queries. In this dissertation we thoroughly investigate topics in the area of skyline-aware query evaluation. In this dissertation, we first propose a novel execution framework called SKIN that treats skyline over joins as first class citizens during query processing. This is in contrast to existing techniques that treat skylines as an "add-on," loosely integrated with query processing by being placed on top of the query plan. SKIN is effective in exploiting the skyline characteristics of the tuples within individual data sources as well as across disparate sources. This enables SKIN to significantly reduce two primary costs, namely the cost of generating the join results and the cost of skyline comparisons to compute the final results. Second, we address the crucial business need to report results early; as soon as they are being generated so that users can formulate competitive decisions in near real-time. On top of SKIN, we built a progressive query evaluation framework ProgXe to transform the execution of queries involving skyline over joins to become non-blocking, i.e., to be progressively generating results early and often. By exploiting SKIN's principle of processing query at multiple levels of abstraction, ProgXe is able to: (1) extract the output dependencies in the output spaces by analyzing both the input and output space, and (2) exploit this knowledge of abstract-level relationships to guarantee correctness of early output. Third, real-world applications handle query workloads with diverse Quality of Service (QoS) requirements also referred to as contracts. Time sensitive queries, such as fraud detection, require results to progressively output with minimal delay, while ad-hoc and reporting queries can tolerate delay. In this dissertation, by building on the principles of ProgXe we propose the Contract-Aware Query Execution (CAQE) framework to support the open problem of contract driven multi-query processing. CAQE employs an adaptive execution strategy to continuously monitor the run-time satisfaction of queries and aggressively take corrective steps whenever the contracts are not being met. Lastly, to elucidate the portability of the core principle of this dissertation, the reasoning and query processing at different levels of data abstraction, we apply them to solve an orthogonal research question to auto-generate recommendation queries that facilitate users in exploring a complex database system. User queries are often too strict or too broad requiring a frustrating trial-and-error refinement process to meet the desired result cardinality while preserving original query semantics. Based on the principles of SKIN, we propose CAPRI to automatically generate refined queries that: (1) attain the desired cardinality and (2) minimize changes to the original query intentions. In our comprehensive experimental study of each part of this dissertation, we demonstrate the superiority of the proposed strategies over state-of-the-art techniques in both efficiency, as well as resource consumption.

Many applications are migrating or beginning to make use native XML data. We anticipate that queries will emerge that emphasize the structural semantics of XML query languages like XPath and XQuery. This brings a need for an efficient query engine and database management system tailored for XML data similar to traditional relational engines. While mapping large XML documents into relational database systems while possible, poses difficulty in mapping XML queries to the less powerful relational query language SQL and creates a data model mismatch between relational tables and semi-structured XML data. Hence native solutions to efficiently store and query XML data are being developed recently. However, most of these systems thus far fail to demonstrate scalability with large document sizes, to provide robust support for the XPath query language nor to adequately address costing with respect to query optimization. In this thesis, we propose a novel cost-driven XPath engine to support the scalable evaluation of ad-hoc XPath expressions called VAMANA. VAMANA makes use of an efficient XML repository for storing and indexing large XML documents called the Multi-Axis Storage Structure (MASS) developed at WPI. VAMANA extensively uses indexes for query evaluation by considering index-only plans. To the best of our knowledge, it is the only XML query engine that supports an index plan approach for large XML documents. Our index-oriented query plans allow queries to be evaluated while reading only a fraction of the data, as all tuples for a particular context node are clustered together. The pipelined query framework minimizes the cost of handing intermediate data during query processing. Unlike other native solutions, VAMANA provides support for all 13 XPath axes. Our schema independent cost model provides dynamically calculated statistics that are then used for intelligent cost-based transformations, further improving performance. Our optimization strategy for increasing execution time performance is affirmed through our experimental studies on XMark benchmark data. VAMANA query execution is significantly faster than leading available XML query engines.

The needs of new modern day applications such as network monitoring systems, telecommunications data management, web applications, remote medical monitoring applications and others for near real time results over continuous data streams have spurred the development of new data management systems called Data Stream Management Systems (DSMS). Unlike traditional database systems which answer one-time user queries only after the finite data has been captured on disk, DSMSs provide on-the-fly answers to user queries as data is arriving at various rates in the form of continuous, potentially infinite streams of tuples. To meet the timeliness requirements of applications, DSMSs aim to keep all data in main memory. Thus queries with multiple stateful operators pose a major strain on memory. Existing adaptation techniques designed to address this issue are ineffective when faced with continuous bursts of high data rates. When system load exceeds system capacity, a DSMS has three options: 1) discard some new data; 2) crash; or 3) spill data to disk. Only option three allows it to produce delayed, yet accurate and complete query results. However, this option involves disk access overhead and change in the natural order of tuples flowing through the query plan tree. As not all stream operators can process correctly out of order tuples, data spilling may have a negative impact on the quality of the final results. Moreover, since operators in a query plan are interconnected, changes in the order of tuple flows inevitably impact the stages of execution of affected downstream operators such as for example data purging . Data purging is necessary for processing continuous queries composed of stateful operators. The state of such operators is divided into finite non-overlapping sets of tuples called windows. Thus, after all the tuples for a window have been processed and all results output, these tuples can be discarded to free memory for new data. To address these issues, we have redesigned the state structure of continuous operators into smaller, finite, non-overlapping sets of tuples such as partitioned window groups, which incur less disk-access overhead. Second, we provide for the capability of continuous operators to correctly process out of order tuples using punctuation pointers. Third, we design methods for downstream operators to synchronize their processing stages with those of upstream operators to achieve optimized query plan throughput. Putting these techniques together, we have designed a consolidated spilling adaptation strategy which considers all aspects of operators' inter-connections in a query plan for making optimal adaptation decisions. The effectiveness of our integrated approach was empirically tested in a comparative evaluation study against several alternate spilling adaptation strategies. We conducted our experiments on CAPE, a DSMS developed at WPI, using different types of query plans composed of multiple partitioned window join operators. Our experiments prove that despite the higher overhead of a more synchronized adaptation approach, our consolidated strategy provides better query plan performance and higher plan throughput during periods of continuous bursts of high data rates.

In a growing number of information-processing applications, such as network-traffic monitoring, sensor networks, financial analysis, data mining for e-commerce, etc., data takes the form of continuous data streams rather than traditional stored databases/relational tuples. These applications have some common features like the need for real time analysis, huge volumes of data, and unpredictable and bursty arrivals of stream elements. In all of these applications, it is infeasible to process queries over data streams by loading the data into a traditional database management system (DBMS) or into main memory. Such an approach does not scale with high stream rates. As a consequence, systems that can manage streaming data have gained tremendous importance. The need to process a large number of continuous queries over bursty, high volume online data streams, potentially in real time, makes it imperative to design algorithms that should use limited resources. This dissertation focuses on processing exact results for join queries over high speed data streams using limited resources, and proposes several novel techniques for processing join queries incorporating secondary storages and non-dedicated computers. Existing approaches for stream joins either, (a) deal with memory limitations by shedding loads, and therefore can not produce exact or highly accurate results for the stream joins over data streams with time varying arrivals of stream tuples, or (b) suffer from large I/O-overheads due to random disk accesses. The proposed techniques exploit the high bandwidth of a disk subsystem by rendering the data access pattern largely sequential, eliminating small, random disk accesses. This dissertation proposes an I/O-efficient algorithm to process hybrid join queries, that join a fast, time varying or bursty data stream and a persistent disk relation. Such a hybrid join is the crux of a number of common transformations in an active data warehouse. Experimental results demonstrate that the proposed scheme reduces the response time in output results by exploiting spatio-temporal locality within the input stream, and minimizes disk overhead through disk-I/O amortization. The dissertation also proposes an algorithm to parallelize a stream join operator over a shared-nothing system. The proposed algorithm distributes the processing loads across a number of independent, non-dedicated nodes, based on a fixed or predefined communication pattern; dynamically maintains the degree of declustering in order to minimize communication and processing overheads; and presents mechanisms for reducing storage and communication overheads while scaling over a large number of nodes. We present experimental results showing the efficacy of the proposed algorithms.

In a growing number of information-processing applications, such as network-traffic monitoring, sensor networks, financial analysis, data mining for e-commerce, etc., data takes the form of continuous data streams rather than traditional stored databases/relational tuples. These applications have some common features like the need for real time analysis, huge volumes of data, and unpredictable and bursty arrivals of stream elements. In all of these applications, it is infeasible to process queries over data streams by loading the data into a traditional database management system (DBMS) or into main memory. Such an approach does not scale with high stream rates. As a consequence, systems that can manage streaming data have gained tremendous importance. The need to process a large number of continuous queries over bursty, high volume online data streams, potentially in real time, makes it imperative to design algorithms that should use limited resources. This dissertation focuses on processing exact results for join queries over high speed data streams using limited resources, and proposes several novel techniques for processing join queries incorporating secondary storages and non-dedicated computers. Existing approaches for stream joins either, (a) deal with memory limitations by shedding loads, and therefore can not produce exact or highly accurate results for the stream joins over data streams with time varying arrivals of stream tuples, or (b) suffer from large I/O-overheads due to random disk accesses. The proposed techniques exploit the high bandwidth of a disk subsystem by rendering the data access pattern largely sequential, eliminating small, random disk accesses. This dissertation proposes an I/O-efficient algorithm to process hybrid join queries, that join a fast, time varying or bursty data stream and a persistent disk relation. Such a hybrid join is the crux of a number of common transformations in an active data warehouse. Experimental results demonstrate that the proposed scheme reduces the response time in output results by exploiting spatio-temporal locality within the input stream, and minimizes disk overhead through disk-I/O amortization. The dissertation also proposes an algorithm to parallelize a stream join operator over a shared-nothing system. The proposed algorithm distributes the processing loads across a number of independent, non-dedicated nodes, based on a fixed or predefined communication pattern; dynamically maintains the degree of declustering in order to minimize communication and processing overheads; and presents mechanisms for reducing storage and communication overheads while scaling over a large number of nodes. We present experimental results showing the efficacy of the proposed algorithms.