The goal of the course is

• to enable students to deepen the knowledge of the architecture and functionalities of relational database management systems (internal structure, modularity, capabilities, ...)
• to achieve a well-established awareness of the implications of the transactional ACID properties on concurrency and reliability
• to show to the students how relational data are stored within a DBMS and how the availability of suitable indexing structures and data statistics profoundly affect query execution
• to learn how a rule-based computational paradigm can help to build incrementally additional features for a DBMS such as business rules, checking of general constraints, view maintenance of derived data, etc. via the trigger concept
• to learn new data models (semistructured, ordered) and new languages to handle data expressed in such models (e.g., XPath, XQuery).

All in all, the main purpose of this course is to build on consolidated data management technologies to provide the learners with the ability to understand the opportunities and limitations of the plethora of new emerging technologies for data management. 

• Transactional Systems. Why transactional systems are relevant. Examples of transactional systems: financial applications, banks, online order-entry (e-commerce), online booking, "wall-street applications". Notion of "transaction". ACID properties: atomicity, consistency, isolation, durability.
• Concurrency control theory. Histories (or schedules), serializability, various notions of equivalence, complexity of testing. View-serializability, conflict-serializability. Two phase locking. Hierarchical locking. Deadlock analysis and resolution. Timestamp-based concurrency control. Multi-version concurrency control. Implementation of locking in commercial systems.
• Reliability control theory. Notion of: stable storage, logging, checkpointing, write-ahead log rule, commit rule. Recovery protocols: warm restart, cold restart. Implementation of reliability control in commercial systems. Commit protocols, theory of two-phase-commit, presumed-abort and read-only optimisations, non-blocking protocols (3 and 4 phase commit protocols). Implementation of commit protocols with heterogeneous DB servers in the X-open standard.
• Database architectures. Distributed databases. Notion of fragmentation, allocation, transparent access. Query optimisation. Distributed transactions. Parallelism in database servers. Shared-memory vs shared-nothing approaches. Scale-up, speed-up, benchmarking of performance. Replicated databases. Synchronous vs asynchronous methods. Symmetric vs primary-secondary approaches. 
• Internal structure of a database server. Buffer Management. Page management. Data organisations according to the sequential, direct, and indexed data structures. B and B+ trees. Hashing functions. Access methods: scans, ordering, joins. Join methods. Query optimisation fundamentals. Cost models and optimal query plan selection (branch&bound method for execution plan selection). Database administration in commercial systems. Hints to physical database design (index selection, primary storage method selection).
• XML Databases. XML as a data modeling paradigm. Native vs relational storage. Query languages for XML. XPath, XQuery.
• Active databases: The ECA Paradigm (event/condition/action) and data management. Execution methods for active databases. Trigger languages and systems. Formal properties of active rule sets. Termination, confluence, observable determinism. Rule analysis. Design of active rules for integrity maintenance, automatic data derivation, application business rules, etc.
• Data analysis: OLTP vs OLAP. Data warehouses. Multidimensional model. Data cubes. Data analysis operators: cube, rollup, pivot.
• Hints to evolutions of database technology: noSQL, data streams.

In order to fully understand and be ready to apply the notions of the course, students are required to possess: 

• Programming skills in any general-purpose language (design of data structure, simple classical algorithms, recursion, stack of activation records) 
• Very good knowledge of the relational data model and of the SQL language (table creation, formulation of complex queries, update commands, integrity constraints)
• Basic knowledge of computing system architectures (memory management, caching algorithms, file system) and operating systems
• Reading Entity-Relationship diagrams 

The exam consists in a written verification covering all the topics of the course.

English version of an older edition of the Italian textbook. PDF available at http://dbbook.dia.uniroma3.it/

General database textbook, for students who are not able to read the textbook in Italian. Many other textbooks can be used.