Complex event processing (CEP) is the analysis of event data in real-time to detect patterns and respond to these event patterns. The area of CEP is an emerging one but it is growing - so much so that Forrester Research has recently conducted a report on the industry (See http://www.destinationcrm.com/Articles/CRM-News/Daily-News/Forrester-Gives-a-Welcoming-Wave-to-Complex-Event-Processing-55492.aspx). It cites hedge-fund trading and algorithmic trading as common applications for CEP. Other CEP applications include credit card fraud detection, business activity monitoring, and security monitoring. The principles of CEP were first defined by Dr David Luckham in his book "The Power of Events".

A complex event is inferred from a series of simple events. Within an organisation, events occur all the time at every level of the firm. The goal of CEP is to discover information contained in these events and analyse their collective impact. CEP will then respond to the occurrence of a particular complex event.

The website complexevents.com (http://complexevents.com) describes CEP as an emerging technology for building and managing information systems including Enterprise Application Integration, Network and Business Security, Regulatory Compliance, Activity Monitoring and Event-Driven Architectures. CEP enables organisations to keep track and react to the information being produced by their systems. The overriding goal of CEP is to make the information contained in events occurring within the organisation's Information Systems available as well as to detect its impact on the organisation and act upon this information in real time. complexevents.com specifically refers to RFID as a technology that can provide information for CEP. CEP is enabled using techniques such as event stream processing, event relationship detection and complex pattern detection. Using CEP, event data can be filtered, aggregated, correlated and analysed. CEP is enabled using the Rapide programming language. Rapide is an event driven system simulation language and analysis toolset. It uses complex event patterns and event processing agents (EPAs) to model dynamic, hierarchical systems.  It is free to download at http://complexevents.com/stanford/rapide/.

A number of vendors offer CEP solutions. Oracle (http://www.oracle.com/technologies/soa/complex-event-processing.html) offers a CEP solution that uses real-time pattern matching to define and identify complex event patterns. It has been voted number 1 CEP provider by Waters Ranking. Microsoft announced their entry into the CEP market in May 2009 (See http://geekswithblogs.net/cyoung/archive/2009/05/11/microsoft_does_wombles-again.aspx) while IBM released their CEP software in May (http://searchdatamanagement.techtarget.com/news/article/0,289142,sid91_gci1356275,00.html). There are also some smaller vendors in the CEP space such as Aleri (http://www.aleri.com/) and Coral8 (http://www.coral8.com), both of whom recently merged.

CEP vendors are mostly focussing on delivering solutions for financial markets. There have also been some applications in Customer Relationship Management (CRM) and solutions that leverage RFID and sensors for industries such as Telecommunications and Utilities. The question we are going to address here is the role CEP could potentially play in the management and securing of devices such as sensors, smart meters and wind turbines.

Coral8's platform offer solutions for the utilities industry so that is a good starting point for assessing the role CEP can play in Green IT. As Coral8 point out, their platform can be used to highlight problems in the delivery of service to consumers. CEP can play an even greater role in the smart grids being developed by electricity utilities as the whole aim of smart meters is to better assess consumption and spread demand for electricity. As a richer level of detail is being provided to utility companies, CEP can be used to analyse this data and identify critical patterns in electricity usage. In addition, the event correlation functionality of CEP can be used to manage the network as device utilisation can be monitored and alert conditions can be flagged when they occur. This equipment monitoring also facilitates predictive maintenance of equipment. 

CEP also gives real-time visibility of events that impact smart grids. Key Performance Indicators for smart grid performance can be detected as they occur and the flow of data across a utility company's IT infrastructure - SCADA systems, ERP systems, Billing systems etc. - can be monitored as they occur. Early trials of smart meters are also presenting problems that CEP can assist. Data can be unreliable and intermittent and CEP can quickly detect the lack of data quality and/or integrity. The functionality of CEP gives a finer level of control over operations as well an enhancing business decision making.

Wind Farms can also benefit from using CEP in the management of their turbines. An individual wind turbine produces a wealth of data and events. From a maintenance perspective, single individual events that occur could represent a pattern that is in fact an alert condition or maintenance requirement. CEP can be used to analyse this stream of events that take place within an individual wind turbine. Furthermore, while data is made available using a SCADA  system and in many cases is accessible from an OPC (OLE Process Control) server, this data is relevant for other systems and software within the organisation. The amount of energy a wind turbine produces, for example, ultimately represents the revenue a company is making. CEP can trace the flow of this data across the Information Systems of a Wind Farm. Availability is another key metric for wind turbines. The time that a wind turbine is unavailable and consequently the amount of revenue being lost can be analysed using CEP and persistent unavailability patterns can be detected and analysed. This is a cardinal point as many Wind Farms have Service Level Agreements with their suppliers where the latter are penalised if a wind turbine is unavailable for more than an agreed period.

Sensors, and Wireless Sensor Network (WSN) technology in particular, can also leverage CEP to detect patterns in the vast amounts of data that can be potentially produced. RFID has already being cited as a technology that can enable CEP and WSNs can play a similar role. WSNs enable a context aware environment for equipment, environments and buildings and produce data that can analysed and transformed into meaningful information using CEP. This is not just applicable to Green IT but also for building, security and supply chain management applications among many others.

CEP then has great potential in many areas including Green IT. Given the rich data and events pools these new technologies provide, CEP should play a major role in analysing these data and events in the coming years.