Bitcoin Network Stress Test: A Review of the 2015 Garbage Transaction Incident

Recently, there have been proposals to remove the policy limits on the OP_Return output size in the Bitcoin core software, which has sparked discussions about spam transactions on the blockchain. This article reviews a series of spam transaction attacks that the Bitcoin network suffered during the summer of 2015, explores the situation then compared to now, and the lessons learned from it.

The 2015 garbage transaction attack was an early skirmish in the block size debate. The attackers were from the pro-expansion side, who believed that the 1MB limit was too small and could easily be filled with garbage transactions. They hoped to increase the block size to raise the cost of filling blocks with garbage transactions. The opposition to block expansion argued that allowing garbage transactions to be quickly added to the blockchain would not deter attackers, but rather enable them to succeed.

First Round of Attack

On June 20, 2015, a Bitcoin wallet service provider named CoinWallet.eu announced that it would conduct a "Bitcoin stress test". They claimed this was to demonstrate the necessity of increasing the block size. The attack was scheduled for June 22, with the goal of creating a backlog of 241MB of transactions.

However, the first round of attacks was unsuccessful. The attackers stated that their server crashed after the memory pool reached about 12MB. They spent approximately 2 Bitcoins as transaction fees in this failed attack.

Second Round of Attack

On June 24, CoinWallet.eu announced that it would carry out a second round of attacks on June 29. This attack seems to be more effective, with some users complaining that transactions are taking a long time to confirm. However, some mining pools have successfully filtered out junk transactions, mitigating the impact of the attack.

This has sparked a debate about whether junk transactions should be filtered. Some believe it would harm the interchangeability of Bitcoin, while others argue that it is necessary to do so to protect the network.

Round Three Attack

On July 7th, the third round of attacks occurred, on a larger scale. It is reported that the attackers spent over $8,000, using various strategies to generate a large number of junk transactions. For example, sending dust transactions to public wallets or sending small amounts of Bitcoin to addresses with known private keys.

During the most intense attacks, some developers believed that increasing the block size was the best defense measure. Some mining pools helped clean up these transactions by creating large transactions to consolidate junk outputs.

Fourth Round of Attack

In September 2015, CoinWallet conducted the final round of "stress testing." This time they took a different approach, announcing a giveaway of 200 Bitcoin by directly posting the private key on the forum. This resulted in over 90,000 transactions being generated, but since many were conflicting transactions, the impact was not as severe as in the third round.

Impact of Attacks

These attacks had a significant impact on Bitcoin, not only changing the relay strategy of transactions at a technical level but also shaping people's perceptions of junk transactions on the blockchain. The network subsequently made some changes:

• Miners increased the block size limit strategy to 1MB
• The minimum relay fee has increased by 5 times.
• Introduced memory pool limits
• Intensified the debate about block size

An academic study found that during peak attack periods, 23.41% of transactions were junk transactions. These attacks caused the average fees for non-junk transactions to increase by 51%, and processing delays increased by 7 times.

Conclusion

The garbage transaction attacks in 2015 demonstrated that even a relatively small expenditure could have a significant impact on the Bitcoin network. These events deepened the understanding of garbage transactions and prompted the network to implement some countermeasures. To this day, discussions on how to define and handle garbage transactions continue. This history reminds us that garbage transaction attacks are not a new phenomenon, and the network needs to continuously adapt and evolve to meet various challenges.