Introduction to persistency and Berkeley DB
- 1. Introduction to Persistency and Berkeley DB Philip Johnson Collaborative Software Development Laboratory Information and Computer Sciences University of Hawaii
- 2. Data Model: Hierarchical Tree of nodes and subnodes "Cousins" have same structure Limited data access language (traversal) Data model for XML
- 3. Data Model: Relational Set of tables containing fields Limited columns, unlimited rows Primary, secondary keys for access Powerful data access language
- 4. Data model: Object Similar to hierarchical no limitation on node structure. Nodes can contain methods to guide retrieval.
- 5. Data Model: NoSQL Generic term for all non-relational DBs key-value stores, document DBs, etc. A rejection of the RDBMS "orthodoxy" Examples: Google's BigTable, CouchDB
- 6. Object-relational "impedance mismatch" A primary difficulty with relational DBs How to map table-based data to objects? Issues: Objects nested, tables flat Objects traversed, tables searched Conversion back and forth Approaches: ORM (Object-relational mapping) frameworks Object databases
- 7. Client-server vs. embedded Client-server: Database runs in one process Supports multiple connections Relatively complicated to set up, administer. Embedded: Database runs "within" the client. Only one client connects to DB. Relatively simple to set up, administer.
- 8. How to choose? Beyond the scope of this lecture! Take ICS 321 Talk to ICS Professor Lipyeow Lim
- 9. BerkeleyDB Open source, embedded, object-oriented, key-value, non-relational, highly concurrent, transactional, high performance, cross-platform Benefits for us: Easy to set up and use. No impedance mismatch. High performance Constraints: Embedded No use of SQL language
- 10. On to the demo