Abstract

Bitcoin is a potentially disruptive new cryptocurrency based on a decentralized open-source cryptographic protocol. The Bitcoin network consists of nodes who contribute their computational power to approve transactions. Transactions are approved in batches that are called blocks once every 10 minutes (in expectation). Transactions need multiple such approvals to occur before they can be considered irreversible with sufficiently high probability. This implies a long waiting time for each individual transaction (a typical transaction may wait for an hour or so). Additionally, blocks are currently restricted in size to 1MB and thus limit the average number of transactions that can be processed per second.

We seek to improve both the waiting time for transaction authorization and the number of transactions processed per second by lowering the block authorization time to far below 10 minutes, and by increasing the maximal allowed block size. We analyze the effects such changes would have on the security of the protocol.

Using the typical block propagation time in the Bitcoin network (recently measured by Decker & Wattenhoffer) our findings, indicate that: 1) At today’s transaction rates, the waiting time for approval can be significantly decreased with negligible compromise with regards to the security of the protocol. 2) Bitcoin’s ability to scale up to greater transaction volumes is inherently limited by the propagation delay in the network. Our analysis allows us to derive estimates regarding this limit.