One of the main indicators that tell you how heavy your load is is the uncle rate for gas usage in a transaction on the Ethereum Blockchain. All Satoshian blocks published on proof-of work blockchains could be at risk of being “obsolete”, i.e. not being included in the mainchain if another miner has previously posted a block before the new block reaches him. This creates a situation in which there is an a. “race” for two blocks so only one block will remain.
A seminal study in the area was done by Decker and Wattenhofer (2013). The average propagation time for a block was 2.08 seconds. This includes 0.08 seconds per kilobyte within a block. (e.g. A 1 MB block would take 82seconds. A recent Bitcoin Unlimited study showed that due to improvements in transaction propagation tech, blocks take less time to reach their destination. The 600 second block time should mean that a 1 second increase in propagation will result in a lower probability of the block becoming stale.
In Ethereum, we have very solid information to analyze due to the “uncle” mechanic. Obsolete blocks can be put back on the chain with “uncles” receiving up to 75% on the initial block reward. This mechanic was initially introduced to reduce centralization pressures. It reduced the advantage that well-connected miners had over less connected miners. However, there are many secondary benefits, one of which is the ability to quickly search for stale blocks. The blockchain itself is the database. We can dump data from blocks 1 to 2283415. September 2016 attacks) is the source of data to analyze.
Using this script http://github.com/ethereum/research/tree/master/uncle_regressions/block_datadump_generator.py, we can generate some source information. The source data is: http://github.com/ethereum/research/tree/master/uncle_regressions/block_datadump.csv. The columns are listed alphabetically and represent the block number. They can also indicate the number of uncles within each block. The total bounty of uncles refers to the amount of gas consumed.
We can use this script http://github.com/ethereum/research/tree/master/uncle_regressions/base_regression.py to interpret it. The results show that the uncle rate is usually between 0.06 to 0.08 with an average gas consumption per block around 100,000 to 300,000. By using the gas consumption per block, we can calculate how much one unit costs. The probability of a block containing a uncle increases when gas is added. These are the coefficients: BLOCK 0 TO 200K: 3.81984698029E/08 BLOCK 200K- 400K: 5.35265798406E/08 BLOCK 500K- 600K: 2.33638832951E/08 BLOCK 40K- 600K: 2.33638832951E/08 BLOCK 4000K- 600K: 2.33638832951E/08 BLOCK 400K- 600K: 2.33638832951E/08 BLOCK 600K- 800K: 2.1244524216E/08 Block 1200K at 1340K – 3.244893833E 08 Block 1400K at 1600K- 3.1258208662E-8 Block 1600K to 1800K – 3.18276549008E. -08 BLOCK 1800K-2500K: 2.4110734444455 – 2200K-2285K: 1.866
So the probability of a block becoming a uncle by adding a 1,000,000-dollar transaction to it with gas increases by 1. Frontier was closer to 35%. The “base” (i.e. uncle rate for a 0 gas block) is 6.7%. We will accept the results as they are and will not draw any conclusions. However, there is a second problem that I will discuss later.
Transaction Prices and Gasoline
The price of gasoline includes transaction spreads and uncle fees. Discussions on Bitcoin development often argue that block size limits are unnecessary as miners have an incentive to limit their block size. This means that every kilobyte more increases the stalerate and threatens your block reward. Given the 8.2 seconds of impedance per second Megabyte as discovered by Bitcoin Unlimited, this could indicate that there might be a transaction fee. Breakeven is 0.000167 BTC per Megabyte, assuming no block sizes limits.
The long-term economics of a capless program are uncertain. There will be no block reward and miners won’t lose any transactions by adding more transactions. These are all good reasons to question their viability. Other transactions may be seen as a transaction fee. To reduce the coefficient, there are some modifications that can be made to this protocol.
In the current Ethereum environment, the block rewards for 5 ETH have been set and will stay that way until the algorithm is changed. Accepting 1,000,000 gallons gasoline is 1.86% less likely to make an uncle. Fortunately, unintended side effects can occur when a mechanism is used. The average uncle reward for 3.2ETH is 3.2ETH. You have 1.86% chance to place at risk if you buy 1,000,000 gallons gasoline. This is to be seen as a transaction fee.
Gas prices are currently high due to the uncle mechanism, which results in a loss of 0.093 Ethereum per block. However, if the uncle system was improved, it could bring the equilibrium gas price down to 11 shannon. Data analysis has shown that there is a correlation between transaction volume and block size in leftover bytes. This is supported by the theory that uncles can form from blocks with high propagation times.
The analysis also found that the gas coefficients are higher than the previous analysis, with an average of 0.04 uncle rates per million. To optimize gas costs, it is recommended to increase the cost per byte by 50% and the base gas cost by 5500 per transaction. It is also important to focus on security measures to protect execution as well as data.
The miner has the authority to decide the price of gas. The social uncle rate is 0.04 per million, and the optimal gas limit is 5.5 million. Long-term, a dynamic gas limit formula that targets block processing time would be best. Future improvements to virtual machines are expected to reduce the uncle rate, but networking improvements will eventually be necessary. Casper is also of great help, as the gas consumption of an uncle is almost zero.
