Skyline-join in distributed databases

Abstract

Abstract-Checkpointing and rollback recovery are well-known techniques for handling failures in distributed database systems. In this paper, we establish the necessary and sufficient conditions for the checkpoints on a set of data items to be part of a transaction-consistent global checkpoint of the distributed database. This can throw light on designing efficient, non-intrusive checkpointing techniques and transparent recovery techniques for distributed database systems.

PrefJoin: An efficient preference-aware join operator

“…Furthermore, two of these research efforts mainly rely on the existence of index structures and are geared towards generating progressive results, turning them not optimized for a full join result [14], [27]. Other research efforts compute skyline join for a single relation, represented as set of sorted list as in [32], a centralized environment [13], [20] or in a distrusted environment [31]. Most of approaches designed for skyline and preference join (with except of [32]) support a generalized join condition.…”

Efficiently Evaluating Skyline Queries on RDF Databases

“…This would be a very natural scenario in the case of RDF data since such data is often stored as vertically partitioning relations [7]. However, there are much fewer efforts [11] [12][16] directed at evaluating skylines over multiple relations. A common strategy, which was also proposed in the context of preference queries on the Semantic Web [8], is to first join all necessary tables in a single table and then use a single table skyline algorithm to compute the skyline result, i.e.…”

“…An important class of querying paradigm for this purpose is preference queries, and in particular, skyline queries. Skyline queries are valuable for supporting multi-criteria decision making and have been extensively investigated in the context of relational databases [1] [12] but in a very limited way for Semantic Web [8].…”

“…The reason is that for a tuple to be selected as being in the skyline, it would need to be compared against all other tuples to ensure that no other tuples [12] for computing skylines on relational databases focus on reducing the number of tuple pair comparisons. This is achieved using indexing [4][5] [6], or partitioning data into subsets [11] where some subsets would contain points that can quickly be determined to be in or pruned from the skyline result.…”

“…Identifying such tuples would allow pruning them from advancing to the skyline phase and avoid the expensive dominance checks needed for skyline computation on those tuples. Alternative techniques to the JFSL approach [4][5] [6][11] [12][16] employ specialized indexing and storage schemes which are not typical in RDF data and require preprocessing or storage in multiple formats.…”