2013
Autores
Santos, A; Nogueira, R; Lourenço, A;
Publicação
ADCAIJ: Advances in Distributed Computing and Artificial Intelligence Journal
Abstract
2013
Autores
Soares, J; Lourenco, J; Preguica, N;
Publicação
EURO-PAR 2013 PARALLEL PROCESSING
Abstract
Multicore processors are available for over a decade, but general purpose database management systems (DBMS) still cannot fully explore the computational resources of these platforms. This paper explores a simple and easy to deploy approach for improving DBMS performance in multicore platforms, by maintaining multiple database engines running in parallel, rather than a single instance, thus circumventing the increase in contention due to load interactions. Unlike previous works, we focus on in-memory DBMS, exploring different design solutions that combine distributed systems and concurrent programming techniques. We show that we are able to improve performance over standalone solutions, without modifying either database or application code, by up to 3 times while minimizing response times.
2013
Autores
Soares, J; Lourenco, J; Preguica, N;
Publicação
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Abstract
Programs increasingly rely on the use of complex component libraries, such as in-memory databases. As any other software, these libraries have bugs that may lead to the application failure. In this work we revisit the idea of software component replication for masking software bugs in the context of multi-core systems. We propose a new abstraction: a Macro-Component. A Macro-Component is a software component that includes several internal replicas with diverse implementations to detect and mask bugs. By relying on modern multicores processing capacity it is possible to execute the same operation in multiple replicas concurrently, thus incurring in minimal overhead. Also, by exploring the multiple existent implementations of well-known interfaces, it is possible to use the idea without incurring in additional development cost. © 2013 Springer-Verlag.
2013
Autores
Dalmazo, BL; Vilela, JP; Curado, M;
Publicação
2013 IEEE THIRD INTERNATIONAL CONFERENCE ON CLOUD AND GREEN COMPUTING (CGC 2013)
Abstract
Monitoring and managing traffic are vital elements to the operation of a network. Traffic prediction is an essential tool that captures the underlying behavior of a network and can be used, for example, to detect anomalies by defining acceptable data traffic thresholds. In this context, most current solutions are heavily based on historical time data, which makes it difficult to employ them in a dynamic environment such as cloud computing. We propose a traffic prediction approach based on a statistical model where observations are weighted with a Poisson distribution inside a sliding window. The evaluation of the proposed method is performed by assessing the Normalized Mean Square Error of predicted values over observed values from a real cloud computing dataset, collected by monitoring the utilization of Dropbox. Compared with other predictors, our solution exhibits the strongest correlation level and shows a close match with real observations.
2013
Autores
Vilela, JP; Barros, J;
Publicação
2013 IEEE 14th International Symposium on a World of Wireless, Mobile and Multimedia Networks, WoWMoM 2013
Abstract
We present a collision-free jammer selection policy for enhanced wireless secrecy. Jammers, selected from the neighbors of a source, are friendly in the sense that they are willing to help the source to transmit securely by causing interference/collisions to possible eavesdroppers. The proposed jammer selection policy results in the selection of the largest number of jammers that do not cause collisions among themselves. This enables jammers to assist the source to transmit securely by causing interference to eavesdroppers, while sending their own traffic into the network. © 2013 IEEE.
2012
Autores
Devezas, J; Alves, H; Figueira, A;
Publicação
INTERNATIONAL MULTICONFERENCE OF ENGINEERS AND COMPUTER SCIENTISTS, IMECS 2012, VOL I
Abstract
We propose a method for creating news context by taking advantage of a folksonomy of web clipping based on online news. We experiment with an ontology-based named entity recognition process and study two different ways of modeling the relationships induced by the coreference of named entities on news clips. We try to establish a context by identifying the community structure for a clip-centric network and for an entity-centric network, based on a small test set from the Breadcrumbs system. Finally, we compare both models, based on the detected news communities, and show the advantages of each network specification.
The access to the final selection minute is only available to applicants.
Please check the confirmation e-mail of your application to obtain the access code.