Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

CTL: A platform-independent crypto tools library based on dataflow programming paradigm

: Ahmad, J.J.; Li, S.; Sadeghi, A.-R.; Schneider, T.


Keromytis, A.D.:
Financial cryptography and data security : 16th international conference, FC 2012, Kralendijk, Bonaire, Februray 27 - March 2, 2012. Revised selected papers
Heidelberg: Springer, 2012 (Lecture Notes in Computer Science 7397)
ISBN: 3-642-32945-4 (print)
ISBN: 978-3-642-32945-6 (print)
ISBN: 978-3-642-32946-3 (online)
International Conference on Financial Cryptography and Data Security (FC) <16, 2012, Kralendijk, Bonaire>
Conference Paper
Fraunhofer SIT ()

The diversity of computing platforms is increasing rapidly. In order to allow security applications to run on such diverse platforms, implementing and optimizing the same cryptographic primitives for multiple target platforms and heterogeneous systems can result in high costs. In this paper, we report our efforts in developing and benchmarking a platform-independent Crypto Tools Library (CTL). CTL is based on a dataflow programming framework called Reconfigurable Video Coding (RVC), which was recently standardized by ISO/IEC for building complicated reconfigurable video codecs. CTL benefits from various properties of the RVC framework including tools to 1) simulate the platform-independent designs, 2) automatically generate implementations in different target programming languages (e.g., C/C++, Java, LLVM, and Verilog/VHDL) for deployment on different platforms as software and/or hardware modules, and 3) design space exploitation such as automatic parallelization for mu lti- and many-core systems. We benchmarked the performance of the SHA-256 implementation in CTL on single-core target platforms and demonstrated that implementations automatically generated from platform-independent RVC applications can achieve a run-time performance comparable to reference implementations manually written in C and Java. For a quad-core target platform, we benchmarked a 4-adic hash tree application based on SHA-256 that achieves a performance gain of up to 300% for hashing messages of size 8 MB.