論文

査読有り
2014年

A trusted knowledge management system for multi-layer threat analysis

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
  • Thanasis Petsas
  • ,
  • Kazuya Okada
  • ,
  • Hajime Tazaki
  • ,
  • Gregory Blanc
  • ,
  • Paweł Pawliński

8564
開始ページ
214
終了ページ
215
記述言語
英語
掲載種別
研究論文(国際会議プロシーディングス)
DOI
10.1007/978-3-319-08593-7_18
出版者・発行元
Springer Verlag

In recent years, we have seen a surge of cybersecurity incidents ranging fromwidespread attacks (e.g., large-scale attacks against infrastructures or end points [1]) to new technological advances (i.e., new generations of malicious code are increasingly stealthy, powerful and pervasive [2]). Facing these incidents, the European Union, Japan, the United States or China have developed national cybersecurity programs, including training of professionals, development of roadmaps for new tools and services, and organization of national interest groups on the topic. There is thus a shared need for a better understanding of this kind of large-scale threats. Some of the basic requirements to better understand these large-scale incidents include handling large volumes of data collected from distributed probes and performing efficient cross-layer analysis. © 2014 Springer International Publishing.

リンク情報
DOI
https://doi.org/10.1007/978-3-319-08593-7_18
DBLP
https://dblp.uni-trier.de/rec/conf/trust/PetsasOTBP14
URL
http://dblp.uni-trier.de/db/conf/trust/trust2014.html#conf/trust/PetsasOTBP14
ID情報
  • DOI : 10.1007/978-3-319-08593-7_18
  • ISSN : 1611-3349
  • ISSN : 0302-9743
  • DBLP ID : conf/trust/PetsasOTBP14
  • SCOPUS ID : 84904193730

エクスポート
BibTeX RIS