T. V. Vijay Kumar

A Reduced Lattice Greedy Algorithm for Selecting Materialized Views

View selection generally deals with selecting an optimal set of beneficial views for materialization subject to constraints like space, response time, etc. The problem of view selection has been shown to be in NP. Several greedy view selection algorithms exist in literature, most of which are focused around algorithm HRU, which uses a multidimensional lattice framework to determine a good set of views to materialize. Algorithm HRU exhibits a high run time complexity. One reason for it may be the high number of re-computations of benefit values needed for selecting views for materialization. This problem has been addressed by the algorithm Reduced Lattice Greedy Algorithm (RLGA) proposed in this paper. Algorithm RLGA selects beneficial views greedily over a reduced lattice, instead of the complete lattice as in the case of HRU algorithm. The use of the reduced lattice, containing a reduced number of dependencies among views, would lead to overall reduction in the number of re-computations required for selecting materialized views. Further, it was also experimentally found that RLGA, in comparison to HRU, was able to select fairly good quality views with fewer re-computations and an improved execution time.

Proposing Candidate Views for Materialization

View selection concerns selection of appropriate set of views for materialization subject to constraints like size, space, time etc. However, selecting optimal set of views for a higher dimensional data set is an NP-Hard problem. Alternatively, views can be selected by exploring the search space in a greedy manner. Several greedy algorithms for view selection exist in literature among which HRUA is considered the most fundamental. HRUA exhibits high run time complexity primarily because the number of possible views that it needs to evaluate is exponential in the number of dimensions. As a result, it would become infeasible to select views for higher dimensional data sets. The Proposed Views Greedy Algorithm (PVGA), presented in this paper, addresses this problem by selecting views from a smaller set of proposed views, instead of all the views in the lattice as in case of HRUA. This would make view selection more efficient and feasible for higher dimensional data. Further, it was experimentally found that PVGA trades significant improvement in time to evaluate all views with a slight drop in the quality of views selected for materialization.

A View Recommendation Greedy Algorithm for Materialized Views Selection

View selection is one of the key problems in view materialization. Several algorithms exist in literature for view selection, most of them are greedy based. The greedy algorithms, in each iteration, select the most beneficial view for materialization. Most of these algorithms are focused around algorithm HRUA. HRUA exhibits high run time complexity. As a result, it becomes infeasible to select views for higher dimensions. This scalability problem is addressed by the greedy algorithm VRGA proposed in this paper. Unlike HRUA, VRGA selects views from a smaller search space, comprising of recommended views, instead of all the views in the lattice. This enables VRGA to select views efficiently for higher dimensional data sets. Further, experimental results show that VRGA, in comparison to HRUA, requires significantly lesser benefit computations, view evaluation time and memory. Alternatively, HRUA has a slight edge over VRGA as regards to the total cost of evaluating all the views.

A query answering greedy algorithm for selecting materialized views

Materialized views aim to improve the response time of analytical queries posed on a data warehouse. This entails that they contain information that provides answers to most future queries. The selection of such information from the data warehouse is referred to as view selection. View selection deals with selection of appropriate sets of views to improve the query response time. Several view selection algorithms exist in literature, most of them being greedy based. The greedy algorithm HRUA, which selects top-k views from a multidimensional lattice, is considered the most fundamental greedy based algorithm. It selects views having the highest benefit, computed in terms of size, for materialization. Though the views selected using HRUA are beneficial with respect to size, they may not account for a large number of future queries and may hence become an unnecessary overhead. This problem is addressed by the Query Answering Greedy Algorithm (QAGA) proposed in this paper. QAGA uses both the size of the view, and the frequency of previously posed queries answered by each view, to compute the profits of all views in each iteration. Thereafter it selects, from among them, the most profitable view for materialization. QAGA is able to select views which are beneficial with respect to size and have a greater likelihood of answering future queries. Further, experimental results show that QAGA, as compared to HRUA, is able to select views capable of answering greater number of queries. Though HRUA incurs a lower total cost of evaluating all the views, QAGA has a lower total cost of answering all the queries leading to an improvement in the average query response time. This in turn facilitates decision making.

Materialized View Selection Using Genetic Algorithm

A data warehouse stores historical information, integrated from several large heterogeneous data sources spread across the globe, for the purpose of supporting decision making. The queries for decision making are usually analytical and complex in nature and their response time is high when processed against a large data warehouse. This query response time can be reduced by materializing views over a data warehouse. Since all views cannot be materialized, due to space constraints, and optimal selection of subsets of views is an NP-complete problem, there is a need for selecting appropriate subsets of views for materialization. An approach for selecting such subsets of views using Genetic Algorithm is proposed in this paper. This approach computes the top-T views from a multidimensional lattice by exploring and exploiting the search space containing all possible views. Further, this approach, in comparison to the greedy algorithm, is able to comparatively lower the total cost of evaluating all the views.

Mining information for constructing materialised views

A materialised view is constructed to improve response time for complex analytical queries posed on a large data warehouse. Most existing approaches use all the queries posed on the data warehouse for constructing materialised views. It is generally observed that, among all the queries posed on the data warehouse in the past, queries that are similar and more frequently posed have high likelihood of being posed again in future and are therefore, appropriate for constructing materialised views. The approach presented in this paper, attempts to select such frequently posed queries from among all the queries posed on the data warehouse. Further, since the materialised views are required to fit within the available storage space, the approach selects a subset of profitable frequent queries that conforms to the space constraint. The information accessed by these queries has high likelihood of being accessed again by future queries. Furthermore, it is experimentally shown that use of this information for constructing materialised views reduces query response time. This in turn would facilitate decision-making.

Greedy Views Selection Using Size and Query Frequency

Greedy view selection, in each iteration, selects the most beneficial view for materialization. Algorithm HRUA, the most fundamental greedy based algorithm, uses the size of the views to select the top-k beneficial views from a multidimensional lattice. HRUA does not take into account the query frequency of each view and as a consequence it may select views which may not be beneficial in respect of answering future queries. As a result, the selected views may not contain relevant and required information for answering queries leading to an unnecessary space overhead. This problem is addressed by the algorithm proposed in this paper, which considers both the size and the query frequency of each view to select the top-k views. The views so selected are profitable with respect to size and are capable of answering large number of queries. Further, experiments show that the views selected using the proposed algorithm, in comparison to those selected using HRUA, are able to answer comparatively greater number of queries at the cost of a slight drop in the total cost of evaluating all the views. This in turn aids in reducing the query response time and facilitates decision making.

Materialized View Selection Using Simulated Annealing

A data warehouse is designed for the purpose of answering decision making queries. These queries are usually long and exploratory in nature and have high response time, when processed against a continuously expanding data warehouse leading to delay in decision making. One way to reduce this response time is by using materialized views, which store pre-computed summarized information for answering decision queries. All views cannot be materialized due to their exponential space overhead. Further, selecting optimal subset of views is an NP-Complete problem. Alternatively, several view selection algorithms exist in literature, out of which most are empirical or based on heuristics like greedy, evolutionary etc. It has been observed that most of these view selection approaches find it infeasible to select good quality views for materialization for higher dimensional data sets. In this paper, a randomized view selection algorithm based on simulated annealing, for selecting Top-K views from amongst all possible sets of views in a multidimensional lattice, is presented. It is shown that the simulated annealing based view selection algorithm, in comparison to the better known greedy view selection algorithm, is able to select better quality views for higher dimensional data sets.

Selection of Views for Materialization Using Size and Query Frequency

View selection is concerned with selecting a set of views that improves the query response time while fitting within the available space for materialization. The most fundamental view selection algorithm HRUA uses the view size, and ignores the query answering ability of the view, while selecting views for materialization. As a consequence, the view selected may not account for large numbers of queries. This problem is addressed by the proposed algorithm, which aims to select views by considering query frequency along with the size of the view. The proposed algorithm, in each iteration, computes the profit of each view, using the query frequency and size of views, and then selects from amongst them, the most profitable view for materialization. The views so selected would be able to answer a greater number of queries resulting in improvement in the average query response time. Further, experimental based comparison of the proposed algorithm with HRUA showed that the proposed algorithm was able to select views capable of answering significantly greater number of queries at the cost of a slight increase in the total cost of evaluating all the views.

Materialized views selection for answering queries

A data warehouse stores historical data to support analytical query processing. These analytical queries are long and complex and processing these against a large data warehouse consumes a lot of time. As a result, the query response time is high. One way to reduce this time is by selecting views that are likely to answer a large number of future queries and storing them in a data warehouse. This problem is referred to as view selection. Several view selection algorithms have been proposed with most of these being focused around HRUA. HRUA considers the size of the views to select the most beneficial view for materialization. The views selected using HRUA, though beneficial with respect to size, may be unable to account for large numbers of queries and thus making them an unnecessary overhead. The algorithm proposed in this paper attempts to address this problem by considering query frequency, along with the size, of the view to select Top-K views for materialization. The proposed algorithm, in each iteration, computes the profit, defined in terms of size and query frequency, and then selects the most profitable view for materialization. As a result, the views selected are beneficial with respect to size and have the ability to answer future queries. Further, experimental results show that the proposed algorithm, in comparison to HRUA, is able to select views capable of answering larger number of queries against a slight increase in the total cost of evaluating all the views. This in turn would result in efficient decision making.