Single Logical View
Single Logical View
Objectivity/DB implements a tiered storage hierarchy, shown in Figure 1 (below). Regular C++, Java or other object-oriented languages create objects that can be made persistent using Objectivity/DB Application Programming Interfaces. An object may have multiple varying length arrays. Objects are stored in logical clusters, called containers, which are analogous to a file. The containers are stored in a database that can consist of a single file or a file for each container. The database files can be stored anywhere in a network of computers and storage devices. The databases are catalogued in a federated database, which also contains the schemas that describe the various objects that can be stored in the federation.
Figure 1 – The Objectivity/DB Logical Storage Hierarch
Each object instance has its own unique Object Identifier [OID] that is preserved for the life of the object. The Objectivity/DB OID is a 32 or 64 bit composite logical structure that can address millions of Terabytes of information. The largest publicly documented Objectivity/DB deployment stored over one Petabyte (100 TB) of objects.
Configuration Parameters
Applications can create ‘one-or-many’ to ‘one-or-many’ relationships between objects. These links, which use OIDs, can cross container and database boundaries, enabling ultra fast navigation across networks of objects.
Objectivity/DB is a true distributed DBMS, so applications do not need to know where objects are stored. Objectivity/DB automatically finds the correct object, fetches it into a client-side cache and formats it to suit the purposes of the local user (language, operating system and hardware variant).
Traversing relationships between objects, using indices, hash tables or other collection structures, can make searches faster. The application has a “Single Logical View” of all of the objects in a federation and a choice of mechanisms for accessing them.
