Date: Oct 4th, 2019, Friday
Location: Room E4229 (da Vinci room),
College of Engineering East Hall, VCU
401 W. Main Street, VA 23284
(Please see below for directions and parking instructions. )
Program (subject to change):
- 9:30am – 10:00am: Breakfast
- 10:00am – 10:45am: Thang Dinh (VCU): BackPackers: A New Network Paradigm for Secure and High-performance Blockchains
- 10:45am – 11:00am: coffee break
- 11am – 11:45am: Arkady Yerukhimovich (GWU): Stormy: Statistics in Tor by Measuring Securely
- 11:45am – 12:30pm: Arka Rai Choudhuri (JHU): Round Optimal Secure Multiparty Computation from Minimal Assumptions
- 12:30pm – 1:45pm: Lunch (on your own)
- 1:45pm – 2:30pm: Qiang Tang (NJIT): Correcting Subverted Random Oracles
- 2:30pm – 3:00pm: coffee break
- 3:00pm – 3:45pm: David Wu (UVA): Watermarking PRFs from Lattices: Stronger Security via Extractable PRFs
Directions and Parking
Please see the below detail regarding our location and parking information. The JL Lot will be the location for you to park.
Virginia Commonwealth University
College of Engineering (East Hall)
401 West Main Street
Richmond, Virginia 23284-3068
Directions to the Engineering Building
Arriving from the North/West by Interstate 95S/64E
Take Exit 76B Belvidere Street. Stay in the middle lane on the exit. Turn left, then get in the right lane for an immediate right turn onto Belvidere Street. When you pass Monroe Park on the right, get in the left lane turning left onto West Cary Street (the next block after Main Street.). Go one block and turn left at Madison Street. The JL parking lot is on the right behind the 7-11.
Arriving from the South/East by Interstate 95N/64W
Take Exit 190 for Fifth Street and Downtown/Coliseum. Turn right at the eighth traffic light onto Main Street. Follow Main Street to Belvidere Street. (Landmarks: 7-11, then the Snead Building on the left). Turn left at Belvidere Street and stay in the left hand lane turning left on West Cary Street (the next light). Go one block and turn left at Madison Street. The JL parking lot is on the right behind the 7-11.
Once in the Eng East Building, make your way up to the 4th floor near the Belvidere and Main St side/corner of the building. Room E4229.
Thang Dinh (VCU): BackPackers: A New Network Paradigm for Secure and High-performance Blockchains
Abstract: Despite many scaling proposals for Bitcoin protocols, existing permissionless approaches, including Bitcoin-Compact, Bitcoin-NG, and Conflux, achieve very low efficiency in terms of networking. In our large-scale peer-to-peer blockchain simulation, no existing permissionless protocols can achieve more than 4% bandwidth utilization, the fraction of bandwidth used for transmitting confirmed transactions. We propose BackPackers, a cross-layer paradigm that optimizes concurrently both consensus (layer 1) and network communication (layer 0) protocols. BackPackers introduces a new node role, called packers, who form a secure and decentralized network backbone. Without any trust assumption, independent packers work together to effectively distribute transactions to all miners, eliminating a major network bottleneck in broadcasting transactions. In exchange for their networking service, each packer receives a portion of transaction fees that it distributes. Through theoretical analysis, we show rigorous proofs for security properties, namely, consistency and liveness. Most importantly, we prove that BackPackers achieves $(1-\epsilon)$-optimality in throughput, with respect to the network limit, and $O(1)$-optimality in block propagation time, even when the network is heterogeneous. Through experimental studies, we show that BackPackers can achieve up to 80\% bandwidth utilization, achieving 7,000+ tps and 0.8s block propagation time for 1,000 nodes with 20Mbps bandwidth. Under the same networking condition, BackPackers achieves an order of magnitude higher throughput comparing to the state-of-the-arts permissionless blockchains.
Arkady Yerukhimovich (GWU): Stormy: Statistics in Tor by Measuring Securely
Abstract: The prevalence of large-scale, Internet-wide distributed systems such as the Tor network call for the development of large-scale secure multi-party computation (MPC). MPC designed for such settings must run over thousands of parties, where the parties have unequal resources (e.g., bandwidth, and processing power), and must be resilient to party failure. In this talk, we present Stormy, an MPC protocol for performing secure measurements over the Tor network that maximizes throughput through optimal utilization of the available bandwidth on Tor routers. Moreover, Stormy’s security requires no additional assumptions over what is already necessary for the secure operation of the Tor network. We describe experimental results showing that Stormy enables important statistics to be evaluated securely over the entire Tor network.
Joint work with Ryan Wails, Aaron Johnson, Daniel Starin, and Dov Gordon
Arka Rai Choudhuri (JHU): Round Optimal Secure Multiparty Computation from Minimal Assumptions
Abstract: We construct a four round secure multiparty computation (MPC) protocol in the plain model that achieves security against any dishonest majority. The security of our protocol only relies on the existence of four round oblivious transfer. This fully resolves the round complexity of MPC (w.r.t. black-box simulation) based on minimal assumptions.
All previous results required either a larger number of rounds or stronger assumptions.
Joint with with Michele Ciampi, Vipul Goyal, Abhishek Jain and Rafail Ostrovsky.
Qiang Tang (NJIT): How to Securely Deploy a Blockchain: Correcting Subverted Random Oracles
Abstract: Hash function is a fundamental primitive for many security applications including blockchain, password login, digital signatures and more. In this talk we focus on the basic problem of correcting faulty—or adversarially corrupted—random oracles, so that they can be confidently applied for such cryptographic purposes.
We prove that a simple construction can transform a “subverted” random oracle—which disagrees with the original one at a negligible fraction of inputs—into a construction that is indifferentiable from a random function. Our results permit future designers of cryptographic primitives in typical kleptographic settings (i.e., with adversaries who may subvert the implementation of cryptographic algorithms but undetectable via black-box testing) to use random oracles as a trusted black box, in spite of not trusting the implementation. Our analysis relies on a general rejection re-sampling lemma which is a tool of possible independent interest.
Bio: Qiang Tang is currently an assistant professor of New Jersey Institute of Technology and also the director of JD-NJIT-ISCAS Joint Blockchain Lab. He was a postdoc at Cornell before joining NJIT and obtained his Ph.D from the University of Connecticut. His research interests are applied and theoretical cryptography and blockchain technology, including post-Snowden cryptography, accountability among others.
David Wu (UVA): Watermarking PRFs from Lattices: Stronger Security via Extractable PRFs
Abstract: A software watermarking scheme enables one to embed a “mark” (i.e., a message) within a program while preserving the program’s functionality. Moreover, there is an extraction algorithm that recovers an embedded message from a program. The main security goal is that it should be difficult to remove the watermark without destroying the functionality of the program. Existing constructions of watermarking focus on watermarking cryptographic functions like pseudorandom functions (PRFs). Even in this setting, realizing watermarking from standard assumptions remains difficult. For example, existing constructions from standard assumptions become insecure in the presence of a mark-extraction oracle or require fully trusting a central watermarking authority (that has the ability to break security of even unmarked keys).
In this talk, I describe a new lattice-based secret-key watermarking scheme for PRFs that provides unremovability against adversaries with access to the mark-extraction oracle and offers a strong and meaningful notion of pseudorandomness even against the watermarking authority. Security of our new schemes can be based on the hardness of computing nearly polynomial approximations to worst-case lattice problems, a qualitatively weaker assumption than that needed for existing constructions of (message-embedding) watermarking. Along the way, I will introduce the notion of an extractable PRF, which offers a new intermediary primitive and approach for constructing cryptographic watermarking schemes.
Joint work with Sam Kim