Security testing of a secure cache design

Cache side channel attacks are attacks that leak secret information through physical implementation of cryptographic operations, nullifying cryptographic protection. Recently, these attacks have received great interest. Previous research found that software countermeasures alone are not enough to defend against cache side channel attacks. Secure cache designs can thwart the root causes of cache side channels and are more efficient. For instance, Newcache is a cache design that can enhance security, performance and power efficiency simultaneously through dynamic memory-cache remapping and eviction randomization. However, these cache designs seldom had their security verified experimentally by mounting cache side channel attacks on them. In this paper, we test the security of Newcache using representative classes of cache side channel attacks proposed for conventional set-associative caches. The results show that Newcache can defeat all these attacks. However, what if a very knowledgeable attacker crafted the attack strategy targeting the secure caches design? We redesign the attacks specifically for Newcache. The results show that Newcache can defeat even crafted access-driven attacks specifically targeted at it but sometimes succumbs to the specifically crafted timing attacks, which is due to a very subtle vulnerability in its replacement algorithm. We further secure Newcache by modifying its replacement algorithm slightly, thus defeating these specifically crafted timing attacks. In addition, the improved Newcache simplifies the replacement algorithm in the original Newcache design.

[1]  Hovav Shacham,et al.  Are AES x86 cache timing attacks still feasible? , 2012, CCSW '12.

[2]  Ruby B. Lee,et al.  A novel cache architecture with enhanced performance and security , 2008, 2008 41st IEEE/ACM International Symposium on Microarchitecture.

[3]  Stephan Krenn,et al.  Cache Games -- Bringing Access-Based Cache Attacks on AES to Practice , 2011, 2011 IEEE Symposium on Security and Privacy.

[4]  Daniel J. Bernstein,et al.  Cache-timing attacks on AES , 2005 .

[5]  Siva Sai Yerubandi,et al.  Differential Power Analysis , 2002 .

[6]  Jean-Pierre Seifert,et al.  Hardware-software integrated approaches to defend against software cache-based side channel attacks , 2009, 2009 IEEE 15th International Symposium on High Performance Computer Architecture.

[7]  Joseph Bonneau,et al.  Cache-Collision Timing Attacks Against AES , 2006, CHES.

[8]  Ruby B. Lee,et al.  New cache designs for thwarting software cache-based side channel attacks , 2007, ISCA '07.

[9]  Robert Könighofer,et al.  A Fast and Cache-Timing Resistant Implementation of the AES , 2008, CT-RSA.

[10]  Onur Aciiçmez,et al.  Trace-Driven Cache Attacks on AES , 2006, IACR Cryptol. ePrint Arch..

[11]  Dan Page,et al.  Partitioned Cache Architecture as a Side-Channel Defence Mechanism , 2005, IACR Cryptology ePrint Archive.

[12]  Colin Percival.  CACHE MISSING FOR FUN AND PROFIT , 2005 .

[13]  Adi Shamir,et al.  Cache Attacks and Countermeasures: The Case of AES , 2006, CT-RSA.

[14]  Nael B. Abu-Ghazaleh,et al.  Non-monopolizable caches: Low-complexity mitigation of cache side channel attacks , 2012, TACO.

[15]  Ruby B. Lee,et al.  Processor accelerator for AES , 2010, 2010 IEEE 8th Symposium on Application Specific Processors (SASP).

[16]  Jean-Jacques Quisquater,et al.  A Practical Implementation of the Timing Attack , 1998, CARDIS.

[17]  Dan Page,et al.  Theoretical Use of Cache Memory as a Cryptanalytic Side-Channel , 2002, IACR Cryptol. ePrint Arch..

[18]  Michael K. Reiter,et al.  Cross-VM side channels and their use to extract private keys , 2012, CCS.