Nonoutsourceable puzzle and secure applications of blockchain

Abstract

Blockchain has gone through ups and downs for more than 10 years since its first implementation Bitcoin was launched. We can hear kinds of voices from the public, sometimes with the worries about its security, sometimes with questions about its applications other than coins and sometimes with the announcements of a new era coming. That drives me to explore this new technique in this thesis. We first focus on one of the security issues, centralized mining pool problem. In Bitcoin-like blockchains, the miners (i.e. block creators) tend to gather together to create a new block so that they can earn rewards more steadily. It leads to the appearance of centralized mining pools. With the fast development, the largest mining pool owns more than 20% computational resources among the whole network, which is a threat to the security of Bitcoin-like blockchains. Researchers proposed nonoutsourceable puzzles to solve it. However, the existing solutions cannot work well with a high-rate transaction processing protocol, GHOST (Greedy Heaviest-Observed Sub-Tree protocol). Thus, we propose a new puzzle construction. In short, we replace the random number in the original puzzle with a nondeterministic signature. We investigate the different signature schemes and one of the advantages of our solution is that it saves much space compared with existing solutions. Besides, we also attempt to use blockchain in some secure applications. Here, we introduce two applications: a blockchain-based verifiable framework and a blockchain-based e-voting system. Since cloud computing is widely used in kinds of applications and people have raised the questions if the remote computation is correct. Taking the secure compressed pattern matching protocol for example, the computation in each round depends on the result in the last round. Thus, to verify the correctness of results in each round is important. Inspired by previous works, we designed a blockchain-based verifiable framework for it. In short, we take blockchain as a public verifier. The clients and the servers upload the required data to the chain. The chain helps to verify the computation and the data owner only need query the result from the chain afterward. Another application is e-voting. The topic has been studied for a long time in the cryptography and security community. Recently, researchers found blockchain is an alternative choice to be the bulletin board in the e-voting systems for its tamper resistance. In this thesis, we propose a self-tallying e-voting system while there are no powerful central authorities nor trusted third parties. In other words, the public can verify the validity of the ballots and tally the ballots. To conclude, the current blockchain technique is not perfect and it is a developing technique, but it has huge potential and a good application foreground.