Distributed Complex Event Processing

Distributed Complex Event Processing

Details
  • Print
Partner:

FZI
Short description

DCEP is the complex event processor in PLAY. Our approach is based on declarative (logic) rules. We will bring this approach to the cloud creating a large-scale, elastic CEP service which dynamically adapts to fluctuating event frequencies.


Main research challenges

•    Algorithms for dynamic operator placement in an elastic Event Processing Network for the cloud
•    Finding ways of dealing with semantically described events to detect complex situations in an efficient, scalable way


Beyond state of the art:

•    Processing events in a dynamically changing topology in the cloud
•    Providing an efficient implementation of a intelligent event reasoning system incorporating real-time and historic data.

 
The role in the whole platform

Complex event processing is a technology which allows correlating basic events to complex events closely aligned with the semantics of services and processes in PLAY where the events originate. It is the role of DCEP to detect context and patterns in events which resemble situations of interest to notify users and services.


More infromation

 

Component Description:
Web applications today create a lot of events. To make sense and react to these events, generic services are scarcely available. Mashup services only provide limited capabilities of detecting and aggregating such events. Thus, in this section we propose a generic Complex Event Processing (CEP) service for the Web. This service should be available to multiple tenants and be able to deal with a high throughput of events in an elastic manner, utilizing computing resources in a cloud. This section collects challenges in the current state of the art.

The role of CEP is to detect complex events using events from various distributed sources such as federated SOAs.

The distributed Complex Event Processing engine will use virtualized computing resources to meet changing demand for event processing. Thereby, complex event-driven communication is provided for large, highly distributed and heterogeneous service systems.

Our approach to CEP is based on declarative (logic) rules. Using rules for CEP has various advantages. First, they are expressive enough and convenient to represent diverse complex event patterns. They come with a formal declarative semantics. Moreover, declarative rules are free of side-effects (e.g. confluence problem). Second, integration of query processing with event processing is easy and natural (e.g. processing of recursive queries). Third, our experience with use of logic rules in implementation of the main constructs in CEP as well as in providing extensibility of a CEP system is very positive and encouraging (e.g. number of code lines in logic programming is significantly smaller than in procedural programming). Ultimately, a logic-based event model allows for reasoning over events, their relationships, entire state, and possible contextual knowledge available for a particular domain. Reasoning about temporal knowledge (i.e., events) and static or evolving knowledge (i.e., facts, rules and ontologies) is a feature beyond of the state of the art in CEP. Taking inference capability into account, logic-based CEP provides a powerful combination of deductive capabilities and temporal features.

 

 

Related Papers

 

Related Documents

Files:
Name Date File size Hits    
PLAY D3.1 Requirements for DCEP 2011-11-17 219.81 KB 483

PLAY D3.2 Distributed CEP - Specification and Prototype V1 2012-04-02 440.98 KB 357

paul

Hello!
Would you like
to PLAY?

Login Form