Current (2014-): Provably Secure Censorship Circumvention ToolsIn this project we are working towards understanding what makes the current Tor pluggable transports detectable. Then, we will formalize what it means for a transport to be secure against a real-world adversary. Finally, we hope that the theory leads to a promising, efficient, and provably secure pluggable transport.
Rob Johnson (Stony Brook University)
Current (2013-): Website Fingerprinting Defenses
Website Fingerprinting (WF) attacks look at side channel information available to a passive "network sniffing" adversary and ask the question: "What content generated the sniffed network trace?". As part of this project we advanced the state-of-the-art of WF defenses by building the first set of information-theoretically secure defenses.
Current efforts are to understand how to lower the overhead of these defenses while making them more practical to implement for use while loading dynamic content.
Rob Johnson (Stony Brook University), Xiang Cai (Stony Brook University), Ian Goldberg (UWaterloo), Tao Wang (UWaterloo)
[C9] Xiang Cai, Rishab Nithyanand, Rob Johnson, "CS-BuFLO: A Congestion Sensitive Website Fingerprinting Defense", WPES 2014. [PDF | BibTex]
[C8] Rishab Nithyanand, Xiang Cai, Rob Johnson, "Glove: A Bespoke Website Fingerprinting Defense", WPES 2014. [PDF | BibTex]
[C7] Xiang Cai, Rishab Nithyanand, Tao Wang, Rob Johnson, Ian Goldberg, "A Systematic Approach to Developing and Evaluating Website Fingerprinting Defenses", ACM CCS 2014. [PDF | BibTex]
[C6] Tao Wang, Xiang Cai, Rishab Nithyanand, Rob Johnson, Ian Goldberg, "Effective Attacks and Provable Defenses for Website Fingerprinting", USENIX Security 2014. [PDF | BibTex]
[M6] Xiang Cai, Rishab Nithyanand, Rob Johnson, "New Approaches to Website Fingerprinting Defenses", Arxiv CoRR abs/1401.7304(2014). [PDF | BibTex]
Past (2013): Dealing with Resource Allocation Problems where "Failure" is Probable
In this project, we introduced a new combinatorial optimization problem -- the Destructive Object Handling problem. The problem models many real-world allocation problems such as: shipping explosive munitions, scheduling processes to fragile cluster nodes, sharing passwords across websites, etc. These problems have objects that have to be assigned to handlers (like every other allocation problem). However, the objects are fragile and have a probability of being destroyed during the allocation/handling process. When they are destroyed to also destroy all other objects allocated to the same handler. The question is: how does one make an allocation of objects to handlers while maximizing the expected "throughput" (survival value) of the allocation?
As part of our research, we showed that the problem was NP-complete, built an FPTAS for the case where there are a constant number of handlers, and identified many human and computer computable heuristics.
Rob Johnson (Stony Brook University), Jonathan Toohill (Google).
[M7] Rishab Nithyanand, Jonathan Toohill, Rob Johnson, "How Best to Handle a Dicey Situation", Arxiv CoRR abs/1401.6022 (2014). [PDF | BibTex]
[C5] Rishab Nithyanand, Rob Johnson, "The Password Allocation Problem: Strategies for Reusing Passwords Effectively", WPES 2013. [PDF | BibTex]
Past (2011-2012): Solving (or, Trying to Solve) the Offline Software Protection Problem with PUFsPhysical Unclonable Functions (PUFs) or Physical One Way Functions (P-OWFs) are physical systems whose responses to input stimuli are easy to
measure but hard to clone. The unclonability property is due to the accepted hardness of replicating the multitude of uncontrollable manufacturing characteristics and makes PUFs useful in solving problems such as device authentication, software protection, licensing, and certiﬁed execution. In this project, we investigated the effectiveness theoretically and practically of PUFs for software protection in hostile ofﬂine settings.
From the theoretical perspective, we showed that traditional non-computational (blackbox) PUFs cannot solve the software protection problem in this context. We provided the first two real-world adversary models (weak and strong variants) and the security deﬁnitions for each. We proposed schemes that were secure against the weak adversary and showed that no PUF based protection scheme could ever be secure against a strong adversary without the use of trusted hardware. Unfortunately, these results were largely negative in the larger context.
Since theoretical security was out of reach and we were gluttons for punishment, we continued to try to achieve some form of practical security. To this end, we formulated the concept of IP-PUFs -- PUFs which were intrinsically involved in handling the sensitive data (and therefore able to inject and remove hardware based flaws while processing the data as usual). To study the feasibility of IP-PUFs, we attempted to harness the inherent unclonability and unpredictability of standard multi-core computers as timing-PUFs. The benefit was that this could be measured via simple benchmarking tools (i.e., no specialized hardware required). We investigated several characterstics (CPU clock crystal inaccuracy and stability of RAM failures) of standard off-the-shelf computers and presented initial experimental results justifying our claim that IP-PUFs could be a practical approach for achieving offline software protection without trusted hardware.
While we were not able to build an entire software protection system, I still believe that PUFs which are intrinsically involved in computations over sensitive data (i.e., our IP-PUFs) are preferable to traditional (peripheral device) PUFs – especially for intellectual property protection and continuous device authentication.
John Solis (Sandia National Labs), Radu Sion (Stony Brook University).
[C4] Rishab Nithyanand, John Solis, "A Theoretical Analysis: Physical Unclonable Functions and the Software Protection Problem", 2012 International Workshop on Trustworthy Embedded Devices, TrustED 2012. [PDF | BibTex]
[M4] Rishab Nithyanand, Radu Sion, John Solis, "Solving the Software Protection Problem with Intrinsic Personal Physical Unclonable Functions", Sandia National Labs Technical Report, SAND2011-6603. [PDF | BibTex]
[P2] Rishab Nithyanand, Radu Sion, John Solis, "Making the Case for Intrinsic Personal Physical Unclonable Functions", 18th ACM Conference on Computer and Communications Security, CCS 2011. [PDF | Poster | BibTex]
Past (2009-2010): Enhancing the Security of Personal RFID Tags with Usable ProtocolsThe emergence of RFID tags that are capable of performing high level cryptographic operations (including public key operations) motivates new RFID applications, including electronic travel documents, identification cards, and payment instruments. This has introduced a new class of RFID tags which store sensitive owner specific data (e.g., biometrics), i.e., personal RFID tags. The primary task of these tags is to identify and authenticate their authorized holders to authorized RFID readers . In such settings, we observe an important feature that distinguishes these tags from the more traditional RFID tags used in supply chain and inventory management is the involvement of a human user and the sensitive nature of data contained in the tags.
In this project, our main strategy was to take advantage of the user's awareness and presence to construct simple, efficient, secure, feasible, and (most importantly) usable solutions for important, yet largely ignored problems in such RFID systems. These included user-to-tag authentication, RFID reader revocation status checking in RFID public key infrastructures, transaction verification, and tag-reader pairing. We also evaluated the usability and practical security of each of our solutions via usability studies which included surveys and actual tests using prototypes (thanks to NXP Semiconductors!). Our solutions also took advantage of new and soon to be practical low-power technologies such as OLED, ePaper, and other more recent advances in hardware integration on RFID tags. The end result was that we were largely able to use these emerging technologies to establish secure I/O channels for communication between the tag owner, the personal tag, and the reader.
This project was also extremely interesting because I got to look at how (and what) security protocols were implemented in RFID enabled credit cards and electronic passports! I also wrote a couple of (very informal) survey papers about these protocols and some security flaws that they had (see papers [M1] and [M2]).
Gene Tsudik (UC Irvine), Ersin Uzun (PARC), Alfred Kobsa (UC Irvine).
[J2] Alfred Kobsa, Rishab Nithyanand, Gene Tsudik, Ersin Uzun, "Can Jannie Verify? Usability of Display Equipped RFID Tags for Security Purposes", Journal of Computer Security, JoCS, Volume 21, Number 3, 2013. [PDF | BibTex] -- This is the complete version of our ESORICS 2011 paper.
[J1] Rishab Nithyanand, Gene Tsudik, Ersin Uzun, "User Aided Reader Revocation in PKI Based RFID Systems", Journal of Computer Security, JoCS, Volume 19, Number 6, 2011. [PDF | BibTex] -- This is the complete version of our ESORICS 2010 paper.
[C3] Alfred Kobsa, Rishab Nithyanand, Gene Tsudik, Ersin Uzun, "Usability of Display - Equipped RFID Tags for Security Purposes", 16th European Symposium on Research in Computer Security, ESORICS, 2011. [PDF | BibTex]
[C2] Rishab Nithyanand, Gene Tsudik, Ersin Uzun, "Readers Behaving Badly: Reader Revocation in PKI Based RFID Systems", 15th European Symposium on Research in Computer Security, ESORICS, 2010. [PDF | BibTex]
[P1] Rishab Nithyanand, Gene Tsudik, Ersin Uzun, "Check The Date: Reader Revocation in PKI Based RFID Systems", 31st IEEE Symposium on Security and Privacy, IEEE SP, 2010. [PDF | Poster| BibTex]
[M2] Rishab Nithyanand, "Securing Plastic Money Using an RFID Based Protocol Stack", IACR ePrint 2009/387. [PDF | BibTex]
[M1] Rishab Nithyanand, "A Report on the Evolution of Cryptographic Protocols in Electronic Passports", IACR ePrint 2009/200. [PDF | BibTex]
Past (2009-2010): Secure Device Pairing in the Group SettingA fairly common modern setting entails users, each in possession of a personal wireless device, wanting to communicate securely, via their devices. If these users (and their devices) have no prior association, a new security context must be established. In order to prevent potential attacks, the initial context (association) establishment process must involve only the intended devices and their users.
Prior to this work, a number of methods for initial secure association of two devices have been proposed and their usability factors had been explored and compared extensively. However, the more challenging problem of initial secure association of a large group of devices (and users) had not received much attention. Although a few secure group association methods had been proposed, their usability aspects were not studied, specifically, not in a comparative manner. In this project we discovered desirable features and evaluation criteria for secure group association, identiﬁed suitable methods and presented a comparative usability study. Results showed that some simple methods (e.g., peer- or leader-based number comparisons) were quite
attractive for small groups, being fast, reasonably secure and well-received by users.
Gene Tsudik (UC Irvine), Ersin Uzun (PARC), Nitesh Saxena (UAlabama).
[C1] Rishab Nithyanand, Nitesh Saxena, Gene Tsudik, Ersin Uzun. "Group Think: On the Usability of Secure Group Association of Wireless Devices", 12th ACM International Conference on Ubiquitous Computing, UbiComp 2010. [PDF | BibTex]