Matakuliah Tahun Versi : M0184 / Pengolahan Data Distribusi : 2005 : Session – 6 DISTRIBUTED DATABASE ARCHITECTURE OBJECTIVES • DISTRIBUTED DATABASE PRODUCTS AND PROTOTYPE • MULTIDATABASE SYSTEM TWO PRODUCTS • INGRES/STAR • ADR/D-NET INGRES/STAR • Accessible from all the nodes in the network without the user having to know where the data actually resides • Client Process paradigm • Important fundamental component Global Communication Architecture (GCA) INGRES/STAR INGRES/STAR, provides : • A global query optimizer • Vertical and horizontal fragmentation / partitioning • A global data dictionary • Several relation DB Interfaces • Heterogeneity of operating system and machines through the gateway feature ARD/D-NET • ARD/D-NET software module manages the entire DB environment, coordinating all activities within the system, including interaction with the user, filtering and routing data requests to the extent that the system support this, handling telecommunication link between nodes and maintaining the status of the distributed directory • The products is design to operate under a variety of IBM operating systems • System allows full portability between PCs and mainframes called CA-DB/PC ARD/D-NET ARD/D-NET capabilities : • Support for multiple databases at various network nodes • Processing of shared databases by many machines • Replication and partitioning data • Location transparency and independence for user and their programs • Excellent security and authorization features, and efficient scheduling of concurrent transactions. Two Multidatabase Systems • MULTIBASE • DISTRIBUTED QUERY SYSTEM MULTIBASE • Design by computer corporate of America, using a simple functional Query language (DAPLEX) • All operations which are necessary to access data relevant to a query are provided automatically and transparently to the user by multibase • The operation includes : locating the data needed to service query, knowing the local formats and the query language, breaking the global query, resolving any inconsistency that may occur between data, combining the partial result, and presenting the result. Distributed Query System (DQS) • DQS that has been develop by CRAI for IBM environments, is a retrieval only multidatabase system using the relational data model for its global data model and SQL as its query language • It work over IMS, IDMS, ADABAS and DB2 local databases. • DQS provides : transparency over languages, data models and distribution, and pre-existing DB are lft unaltered.