Home Company News & Events Break DES in less than a single day
Break DES in less than a single day

SciEngines is proud to announce that their massively parallel architecture RIVYERA has achieved the highest-known speeds for 56-bit DES decryption for a single hardware accelerated server since 2008.

Started 2006 with a third party scientific publication achieving a throughput of more than 65 billion keys per second using Xilinx Spartan-3 1000 FPGAs on a SciEngines COPACOBANA. SciEngines reported on July 15th 2008 to break DES in an average time of a single day using a setup of COPACOBANA. 2009 we achieved a throughput of over 280 billion keys using 128 Xilinx Spartan-3 5000 FPGAs in a RIVYERA single 3 HU hardware accelerated server. Nevertheless we are proud to announce that we managed to reach the 292 billion keys per second and lower the total power consumption of our RIVYERA machines to 75%.

SciEngines' massively parallel high performance computing platform consists of 16 RIVYERA S3-5000 FPGA cards, each equipped with 8 FPGAs. The machine reportedly consumes less than 850 peak watts of power. It can be fit in any off-the-shelf rack, occupying not more than 3HU.

The system is quality 'made in Germany'.

 

In general, the Data Encryption Standard (DES) is known to be broken since the first a custom hardware attack in 1998 and has been replaced by more modern standards like Triple-DES, AES, RSA. Nevertheless DES is still in use. Providing the first known affordable solution for non government customers attacking crypt-algorithms using a attack brute force in 2006, SciEngines COPACOBANA has an important role in cryptographic research as a reference system.

"Our current and upcoming products are still the bar to measure current cryptanalytic algorithms and third-parts products.“ said Gerd Pfeiffer, SciEngines GmbH, CEO of SciEngines GmbH. “Our solutions are fully scalable in hard and software providing more performance and consuming less energy than standard architectures. We managed to eliminate the bottlenecks of classical cluster based computing and of hybrid computing accelerators using PC based bus systems like PCI-Express. DES and Cryptanalysis is one of the markets SciEngines machines have proved their market unique performance, others are natural sciences, like weather forecast and physics as well as the simulation of the ocean. FPGAs provide in these markets best cost performance ratio, compared with GPUs and CPU Clusters.“

„DES is a good example, 2006 a single COPACOBANA S3-1000 provided the performance of more than 3000 PCs in a brute force attack against DES. 2008 with COPACOBANA V4-SX35 and the first prototype of RIVYERA we were able to give this performance more than a factor of 4.3“ said Stefan Baumgart, chief hardware engineer an expert in massively parallel hardware. „2009 we were the number one in of-the-shelf hardware accelerated one click solution for brute force personal key recovery. Now introducing RIVYERA to a bigger audience, 2010, RIVYERA is accessible to every one providing the most simple to use personal key recovery tools ever designed. Our method allows to recover a lost DES Key with one simple click. Our hardware accelerated algorithm automatically analyses the 56 bit encrypted file, searches for the key using brute-force and returns the key candidate which has been tested to bet he correct key.“

SciEngines' FPGA based platform technology is capable of processing more than 292 billion keys per second utilizing 128 FPGAs. This means a key recovery which has been known to take more than a hundred years on a single PC or a couple of months on a PC Cluster or a hybrid computer cluster equipped with GPUs or FPGA-PCI-Express cards can be accomplished in a single day.

 

About SciEngines

SciEngines was founded on the vision that we can satisfy the computational needs in science and engineering. Although High Performance Computing (HPC) and FPGA technology are not new, we believe that the combination of both is an attractive alternative for unaffordable or inaccessible super computers. Our FPGA based massive parallel high performance computers are designed to scale up the advantageous cost-performance-ratio and power-performance-ratio of single FPGAs into the massive parallel level.
We are specialists in Field Programmable Gate Arrays (FPGAs) based development. SciEngines products address Applications like natural science, cryptography, signal processing and scientific computing. For more information visit www.sciengines.com.