QONCEPT
Query Optimisation in CEP Technologies   
  

Student Projects

Students are welcome to contribute to the research project either as paid student researches or with Project, Diploma, Bachelor or Master thesis. There are the following student projects:

Topic Propagation of ECSL Cardinality Constraints with Respect to CEP Queries
State Finished
Student research and Bachelor thesis of Thanh Son Dang
Time period February - December 2011
Abstract Semantic optimization of database queries, i.e. the use of metadata (constraints) for query optimization, is well investigated and has led to significant performance gains. The algorithm semantically rewriting database queries is applicable to CEP queries with some adaptions. One of these adaptions concerns the treatment of views and the constraints for them.

To reduce the number of constraints which must be manually specified by the user, the algorithm semantically rewriting database queries leads views back to their respective base relations so that constraints have to be defined for base relations only. However, if a CEP engine involves automatic garbage collection of events and states, the relations saving these events and states may become incomplete in the meantime such that CEP views cannot be led back to them without lost of tuples in some cases. Therefore, an approach opposite to the one in databases is introduced in this work: ESCL cardinality constraints are automatically propagated from base events (states) to the derived events (states) with respect to the queries deriving them, i.e. only the ESCL cardinality constraints for base events and states must be manually specified by the user. The propagation of the other kinds of ESCL constraints is to be investigated.

The approach presented in this report is independent from an event query language. To illustrate it, CEP queries are expressed in StreamLog, a first order logic language extended with temporal aspects as required for CEP.

Goals
  1. Algorithm propagating ESCL cardinality constraints with respect to StreamLog queries independently from the complexity of queries and constraints, query evaluation time must not coincide with constraint validity time
  2. Proofs of termination, correctness, and completeness of the algorithm, investigation of its complexity
Topic Instantiating Hierarchical Timed Automata
State Finished
Project thesis of Sandro Giessl
Time period November 2010 - May 2011
Abstract The application logic of many information systems often follows a given workflow enforcing certain orders and dependencies of events. This knowledge can be used to make event processing more efficient. To this end Instantiating Hierarchical Timed Automata (IHTA) capturing the knowledge are formally defined in this work. Later, constraints will be derived from IHTA and used by the semantic query optimisation algorithm.

IHTA are nondeterministic automata the states of which correspond to the application states and the edges between them are labeled with event queries and temporal constraints. The labels of IHTA are more expressive than that of Timed Automata for the following reasons. First, the event queries of IHTA allow access to event data. Second, the temporal constraints of IHTA are expressed on the beginning and the end of the occurrence time of matched events (and not on local clocks).

Each state of IHTA is either atomic or non-atomic, i.e. is IHTA itself, so that an arbitrary level of abstraction can be reached. The idea of a hierarchy of models is not new. Hierarchical Timed Automata and Statecharts are examples of such models. However, only a fixed number of concurrent processes can be modelled by them. IHTA extend them by instantiation allowing an arbitrary number of concurrent processes (specified by the same non-atomic state) to run at the same time.

Goals
  1. Requirements to a model of application semantics of event-based systems
  2. Analysis of automata, flowcharts, Petri nets, and UML with respect to these requirements
  3. Syntax and semantics of IHTA, automaton configuration modification