Cryptocurrency Network Measurement
Surya, Sergi, Andrew, Simon, Sid (collaboration with CMU, UMD)
Bitcoin is the most well known cryptocurrency, but there are hundreds of competitors. Built on top of a peer-to-peer network, consisting of thousands of nodes.
What is the topology of this network? Is it vulnerable to attacks? These questions can best be answered by measuring the network. We build and deploy experiments using Coinscope, a platform for connecting to, monitoring, and actively interacting with cryptocurrency networks. Coinscope project page at UMD
Kevin Lee, Andrew, Niraj (collaboration with Princeton, CMU, BU)
We analyzed transactions in the Monero network and proposed deanonymization heuristics for transactions. We investigated and proposed solutions to the security flaws in the anonymity of Monero transactions, which en hanced the security of all Cryptonote currencies and led to the adoption of the solutions still in use today. MoneroLink. Technical Report Andrew Miller, Malte Moser, Kevin Lee, Arvind Narayanan. http://monerolink.com/monerolink.pdf
Honey Badger BFT is the first *asynchronous* consensus protocol for the post-Bitcoin world. Unlike the other protocols you’ve heard of, like Raft, Paxos, and PBFT, our new protocol makes progress whenever messages are delivered, regardless of how long they are delayed. HoneyBadgerBFT just doesn’t care about the timeliness of the underlying network! HoneyBadgerBFT GitHub page The Honey Badger of BFT Protocols Andrew Miller and Yu Xia and Kyle Croman and Elaine Shi and Dawn Song. To appear, CCS 2016. https://eprint.iacr.org/2016/199
Privacy for Permissioned Blockchains
Tom, Hanyun, Andrew
Permissioned blockchains today do not provide strong privacy guarantees. Since all the nodes reach consensus about committed transactions, if even one node suffers a data breach, then privacy is lost. We are investigating how to combine threshold cryptography, homomorphic encryption, and to provide privacy.
Sprites and State Channels: Payment Networks that go Faster Than Lighting
Provable Security for Blockchains
SaUCy – Composable Cryptography Framework
Kevin Liao, Abhiram, Andrew
SmartCast – Incentive compatible consensus protocols
SmartCast is a consensus protocol that utilizes rational nodes incentivized by smart contracts. A smart contract is informed of node behavior through a reporting mechanism. If the node properly follows the protocol, the smart contract issues a reward. To ensure that following the protocol is a Nash Equilibrum, we set a bound on the reward for which the worst case reward for following the protocol exceeds the best case reward for deviating.
Hawk – Privacy Preserving Smart Contracts