Caching for Object Database Management Systems
Caching is the retention of data, usually in the application, to minimize network traffic flow and/or disk access. With transparent persistence, caching is often set up as part of the application work space. With this cache, there is no need for the application to explicitly move objects from disk. They move from disk storage automatically into program memory. This is either as a result of a query or from traversing the graph structure. For an animation that demonstrates this caching, see transparent persistence.
Caching for an ODBMS is similar in many respects as is caching for object-relational mapping. The exception is that that caching for object-relational mapping can have problems with cache synchronization. This problem not occur with ODBMSs because the cache is tightly integrated with the ODBMS server.
ODBMSs provide the lowest cost for development and best performance combination when using objects because they store objects on disk and have the transparent program integration with object programming languages. Performance is enhanced by storing objects directly on disk which eliminates impedance mismatch. Development costs are reduced because there is no need to program the caching for the application programs and there is only one model to develop.
More on the general topic: Object Database Articles
- Object-Oriented Database Management System (OODBMS) Definition
- When an Object Database Should Be Used
- ODBMS FAQ
- Architectures That Use Object Database Products
- Everyday Uses of Object Databases
- ODBMS Myths
- Object Database Problems?