BICEP · Benchmarking complex event processing systems
6РП — Действия „Мария Кюри“
- Период
- 2007-04-01 → 2009-03-31
- Финансиране от ЕС
- 80 000 €
- Участници
- 1
- Схема
- EIF
Линиите свързват координатора с партньорите. За проекти отпреди 2014 г. CORDIS не винаги дава точни координати. Тези точки са на ниво град или държава.
Накратко на български
Системите за обработка на сложни събития анализират данни в реално време, например при мониторинг в интензивни отделения или финансови пазари. Тестовете им помагат да се избере най-подходящият софтуер според сложността на данните и критичността на ситуацията.
Кратко обяснение, генерирано от езиков модел по текста на CORDIS. Оригиналът е по-долу.
Резултати накратко
Final Activity Report Summary - BICEP (Benchmarking Complex Event Processing Systems)
Funding for the BICEP project was awarded to Pedro Bizarro, a Computer Science researcher living and working in the USA in 2006. The project had two main goals. As a research project, the first goal was to produce a performance evaluation of complex event processing (CEP) systems. CEP is a new class of data management systems that determine what is happening in real-time and is being used in life- or mission-critical scenarios such as hospital intensive care units or financial markets. Interestingly, most CEP scenarios have very different characteristics in terms of types of data, or frequency and complexity of the operations performed. % Do CEP systems degrade gracefully with increasing loads? Are there situations where some engines can be used safely (e.g., financial fraud detection, where throughput is high) and other situations where they should not be used (e.g., emergency care units, where throughput is low but the cost of false-positives is high)? To answer these and other questions, the BICEP project set out to produce novel performance metrics, measuring techniques, and benchmarks. This goal was clearly achieved: a novel benchmark infrastructure was developed (demonstrated at DEBS'2008), the benchmarks and results were presented at TPC-TP'2009, and a tutorial is going to be presented at WOSP/SIPEW'2010 Intl. Conf. on Performance Engineering. The results are clear: no engine dominates all others, performance varies wildly with data or query complexity, and stability depends on the engine used. These findings will allow to better decide which CEP engines to use in life- or mission-critical scenarios. As a Marie Curie International Re-integration Grant, the second major goal of the project was to help a European researcher return to Europe and establish his research team. This goal was also clearly achieved. Since the start of the project, the researcher created a team of two Ph.D. and five Master's students, collaborated with professors in his department, in Europe or abroad, and has taken a leadership role in the international Event Processing Technical Society. The project produced another important system: SICU. Done in collaboration with researchers from Oracle Corporation (CA, USA), Utah Medical Center (UT, USA), and Whitehead/MIT Center for Genome Research (MA, USA), SICU is, as far as we know, the first system worldwide that integrates functionality from databases, data warehouses, medical sensors, and data mining. This integration, together with its highly customizable interface, lets SICU do two important things. First, it greatly reduces the number of false-positive and -negative alarms. Second, SICU is able to predict, with 68% accuracy, if a patient is going to have a cardiac arrest in the following 24 hours. This system, presented in multiple conferences or symposiums, has the potential to revolutionise Event Processing as well as Intensive Care Units systems. And can save many lives in the process. SICU has won the BES Innovation Award, a yearly Portuguese prize that rewards innovation in five different areas (Energy, Health Technologies, New Materials and Industrial Processes, Biotechnology and Agro-industry, and IT and Services). Organized by 11 Universities, one newspaper, one radio, BES bank, the Portuguese Science and Technology Foundation and some other organizations, the BES Innovation Award is a highly prestigious award. SICU won both its category award as well as the overall award for a total of over 60K in prizes. Finally, BiCEP has also generated a spin-off company named FeedZai. FeedZai is a result of the direct conclusions of the project: CEP engines in the market are either too complex to use or change, cannot be used, are too expensive, or have a poor, unstable performance. FeedZai, founded in late 2008 by Pedro and two partners, is building FeedZai Pulse, to fill this gap. The company has already secured close to 500K in funding and consulting and is developing monitoring tools in the Utilities market. For its entrepreneurial, integration, scientific results, and awards, BICEP is proving to be an aptly named strong project.
Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз
Цел на проекта
Many IT systems consist of several distributed pieces of software processing thousands of events (or messages) per second. Complex Event Processing (CEP) is a new area at the intersection of relational databases and data stream systems that promises to discover, in near-real time, relevant information hidden in the million of events processed by IT systems.As an example, a CEO of one CEP company described an application that, using RFID technology, detects and locates missing children in amusement parks even before the parents notify the child is missing. A number of companies in the US have just made available their first CEP products.However, the products were developed by people from different backgrounds: arbitrage companies, query language researchers, data stream developers, and database vendors. The different backgrounds and the rush to beat the competition lead to a fragmented market: essentially every CEP product has a different query language, semantics, feature set, and performance concerns. This market fragmentation is not just unfortunate; it also stifles research.The CEP field looks today as the database field looked 30 years ago: competing data models and lack of standards and benchmarks. In databases, SQL and the relational model were eventually accepted, benchmarks proposed and a common ground was found. Scientists proposed new algorithms, which could then be applied to a large product base and the database field exploded into a multi-billion dollar market with first-class research ideas yea r after year.This is a great time for Europe to step into this new field, and lead it. In that direction, we propose an important first step: Benchmarks for CEP systems (BiCEPs). BiCEPs can help unify the CEP market, can help the discovery of the fundamental scientific CEP issues, and can help researchers evaluate new algorithms. Developing BiCEps can also lead to collaborations with some of the CEP companies and possibly new, European, CEP products.
Оригинален текст от CORDIS (на английски).
Участници
- FACULDADE DE CIENCIAS E TECNOLOGIA DA UNIVERSIDADE DE COIMBRA · COIMBRAКоординаторНиво градПортугалия
Връзки
Данни: CORDIS, © Европейски съюз
