Such systems recently started to appear and are sometimes denoted as NewSQL. In the new hardware era, all these components could be implemented to reside in the main memory. (2008) and are related to ACID transactions (i.e., logging, locking, and latching), as well as buffer management operations. The main components responsible for the performance bottleneck of current RDBMS systems have been identified in Harizopoulos et al. The needed adaptations have to consider the evolution of hardware that has happened during the last few years-for example, the cost of main memory is decreasing so rapidly that servers with hundreds of gigabytes is not uncommon SSDs are getting less expensive and are starting to replace disks in some situations faster CPUs and networks are arising computing with graphics processing units (GPUs) is easier through APIs and programming languages and dominance of shared-nothing architecture is being confirmed. In fact, the paper states that if they are not adapting rapidly, they could even lose their leading position in their OLTP niche market. (2007), the authors argue that the one-size-fits-all property of RDBMSs is over. To cope with the goals of NoSQL-that is, storing and processing large data sets on machine clusters-RDBMSs may have to rearchitecture at least some of their main components. NoSQL can be considered the latest threat for RDBMS dominance. These adaptations never involved deep architectural modifications and most of the main components of RDBMSs still rely on the design choices of the 1970s and 1980s. Each time, they have adapted to the situation by introducing novel functionalities and retained their market dominance. Nevertheless, due to space limitations, we do not consider specific fields such as stream processing and scientific database systems.ĭuring their long history, RDBMSs have faced several contenders, such as object databases in the 1990s and XML databases in the 2000s. In this section, we focus on evolutions that impact the two kinds of systems presented in this chapter and that will probably influence future solutions in the management of RDF data. In RDF Database Systems, 2015 2.3 Evolutions of RDBMS and NoSQL systemsĭue to the Big Data phenomenon, data management systems are almost obliged to evolve to cope with new needs.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |