TY - JOUR
T1 - Toward Secure Microfluidic Fully Programmable Valve Array Biochips
AU - Shayan, Mohammed
AU - Bhattacharjee, Sukanta
AU - Song, Yong Ak
AU - Chakrabarty, Krishnendu
AU - Karri, Ramesh
N1 - Funding Information:
Manuscript received February 1, 2019; revised May 21, 2019; accepted June 14, 2019. Date of publication July 16, 2019; date of current version November 22, 2019. This work was supported in part by the Army Research Office under Grant W911NF-17-1-0320, in part by NSF under Award CNS-1833622 and Award CNS-1833624, in part by the NYU Center for Cyber Security (CCS), and the NYU Abu Dhabi Center for Cyber Security, Abu Dhabi (CCS-AD). A preliminary version of this paper has appeared in the Proceedings of VLSI Design 2019. (Corresponding author: Mohammed Shayan.) M. Shayan and R. Karri are with the Department of Electrical and Computer Engineering, New York University, Brooklyn, NY 11201 USA (e-mail: mos283@nyu.edu; rkarri@nyu.edu).
Publisher Copyright:
© 1993-2012 IEEE.
PY - 2019/12
Y1 - 2019/12
N2 - The fully programmable valve array (FPVA) is a general-purpose programmable flow-based microfluidic platform, akin to the VLSI field-programmable gate array (FPGA). FPVAs are dynamically reconfigurable and, hence, are suitable in a broad spectrum of applications involving immunoassays and cell analysis. Since these applications are safety critical, addressing security concerns is vital for the success and adoption of FPVAs. This study evaluates the security of FPVA biochips. We show that FPVAs are vulnerable to malicious operations similar to digital and flow-based microfluidic biochips. FPVAs are further prone to new classes of attacks-tunneling and deliberate aging. This study establishes security metrics and describes possible attacks on real-life bioassays. Furthermore, we study the use of machine learning (ML) techniques to detect and classify attacks based on the golden and real-time biochip state. In order to boost the classifier's performance, we propose a smart checkpointing mechanism. Experimental results are presented to showcase: 1) best-fit ML model classifier; 2) performance of different tradeoffs in checkpointing; and 3) effectiveness of the proposed smart checkpointing scheme.
AB - The fully programmable valve array (FPVA) is a general-purpose programmable flow-based microfluidic platform, akin to the VLSI field-programmable gate array (FPGA). FPVAs are dynamically reconfigurable and, hence, are suitable in a broad spectrum of applications involving immunoassays and cell analysis. Since these applications are safety critical, addressing security concerns is vital for the success and adoption of FPVAs. This study evaluates the security of FPVA biochips. We show that FPVAs are vulnerable to malicious operations similar to digital and flow-based microfluidic biochips. FPVAs are further prone to new classes of attacks-tunneling and deliberate aging. This study establishes security metrics and describes possible attacks on real-life bioassays. Furthermore, we study the use of machine learning (ML) techniques to detect and classify attacks based on the golden and real-time biochip state. In order to boost the classifier's performance, we propose a smart checkpointing mechanism. Experimental results are presented to showcase: 1) best-fit ML model classifier; 2) performance of different tradeoffs in checkpointing; and 3) effectiveness of the proposed smart checkpointing scheme.
KW - Computer security
KW - microfluidics
KW - microvalves
KW - statistical learning
UR - http://www.scopus.com/inward/record.url?scp=85075637957&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85075637957&partnerID=8YFLogxK
U2 - 10.1109/TVLSI.2019.2924915
DO - 10.1109/TVLSI.2019.2924915
M3 - Article
AN - SCOPUS:85075637957
SN - 1063-8210
VL - 27
SP - 2755
EP - 2766
JO - IEEE Transactions on Very Large Scale Integration (VLSI) Systems
JF - IEEE Transactions on Very Large Scale Integration (VLSI) Systems
IS - 12
M1 - 8764604
ER -