A planned ‘ultimate stress test’ for the bitcoin network has passed without major incident, though the company organizing the event said it was only able to send 15% of the transaction volume it had intended.
Bitcoin brokerage CoinWallet.eu had planned the stress test starting 22 June at 13:00 GMT, to last for 100 blocks. It had sought to spend 20 BTC (around $5,000) on transactions totaling about 200MB in data.
The original intention was to flood the bitcoin network with transactions to see if the current 1MB block size was adequate at such volume, and if the network could recover quickly from a surge.
CoinWallet’s stated aim was to demonstrate that the 1MB block size is inadequate if bitcoin is to become “anything more than a costly science project”.
Ten bitcoin servers would send transactions at two per second, each approximately 3KB in size and each sending to 10–20 addresses. Outputs from those transactions (totalling larger transactions of around 15-30KB) would then be combined and sent back to the original servers.
This followed a series of shorter tests CoinWallet performed over the past few days.
However, CoinWallet’s servers failed and the test could not be completed at the volumes planned.
The company posted on Reddit:
“At 17:00 GMT our BitcoinD servers crashed. The servers were restarted but failed to meet our planned transaction volume. The max pending backlog was 15MB. Far short of the 200MB goal. By late evening the test was considered complete. Approximately 15% of our planned transaction volume occurred.”
CoinWallet said it will launch another test in seven days.
CoinWallet had faced criticism for launching the test live on an unsuspecting bitcoin network, with some calling it an ‘attack’.
Others though, like bitcoin developer Peter Todd, were curious to see its effects and posted instructions on how users could ensure their transactions still went though.
For most, the test period appears to have passed without incident.
There were some anecdotal reports of transaction delays posted online, though it has not been ascertained that they were a result of the test or other circumstances.
CoinWallet attached a range of different fees to its test transactions to see whether everyday more bitcoin transactions would be delayed. Some of the test fees exceeded 10,000 satoshi (0.0001 BTC) per KB of transaction data.
Wallet software maker MultiBit posted on its website that transactions with only 1,000 satoshi (0.00001 BTC) fee per KB were delayed during the test, some of them taking as much as 87 blocks to confirm.
The MultiBit HD wallet has a default setting of 3,000 satoshi (0.00003 BTC) per KB, which is user-adjustable between 1,000 and 10,000 satoshi.
Transactions set at the default 3,000 satoshi took from 11–80 blocks to confirm, and those set at the maximum 10,000 satoshi took an average of nine.
The team concluded:
“When the Bitcoin network is flooded with transactions at a particular fee level, transactions with lower fees do not get confirmed in a timely manner.”
While most in bitcoin have argued recently for the size of a single 10-minute transaction block to increase to either 8MB or 20MB to handle larger volumes, others prefer the status quo.
BitTorrent protocol creator Bram Cohen wrote in a blog post titled Bitcoin’s Ironic Crisis yesterday that 1MB block sizes are preferable, as this would lead to more competition in transaction fees and in turn benefit the miners who keep the network secure.
Image via Shutterstock