Now that Zilliqa adoption is on the upswing, new dapps are being launched and Scilla is about to reach the milestone version 0.10 release, it’s a good time for some future-looking. Nothing is as seemingly inevitable as the rising demand for green solutions, and critical for us — the possibility of being rejected as ‘dirty’ on the basis of carbon emitted. Now I have to disclose that I am both a die-hard environmentalist and a Zilliqa hodler, and I don’t think those two things are a contradiction. The rest of this article will contextualize the use of Proof of Work in the chain and set us up to estimate both the 2021 total and 2021 per-transaction CO2 that Zilliqans should worry about. But first, we have to understand exactly where Zilliqa’s use of PoW comes into the picture today, then you can look forward to a future series and perhaps governance portal vote on this issue.
Proof of Work and it’s Critics
Proof of Work is essential to a functioning public blockchain and it is likely that every Proof of Stake or Proof of Authority solution I’ve evaluated as of today is doomed to fail. There, now you know I am a heretic in crypto. I would love to indulge in a two-hour rant about why protocol-level Proof of Stake is a fool’s errand, and the people beating that gong with gusto have ulterior motives behind their [mostly] fake climate concern, but that will have to wait. For now we are going to look at what Proof of Work does for Zilliqa specifically, and what that means for the composition of the atmosphere. It’s impossible to do this analysis from scratch in a blog post, so we will piggy-back on the existing work already done to estimate Bitcoin and Ethereum’s impact. There isn’t a standard method to estimate the impact of these coins, but there are two methodologies I like: the first estimates the top PoW networks by how much collateral carbon is emitted based on reported electricity costs, usage statistics and some rational estimates, which Digiconomist also adapts from Bitcoin to Ethereum. We will use the Ethereum data to inform our estimate of the Zilliqa impact numbers in this post.
Much of the accessible academic literature on the subject looks at the carbon emitted per coin or per new dollar of wealth created, which I think muddles the water a bit and certainly swings with coinbase rewards and token prices. Although all are valid approaches, none can be said to be strictly correct, so I selected Digiconimist’s model, who is a critic of the energy consumption in blockchains but I believe uses more grounded real-world data to get to their benchmark than most of the academic estimates that often use strange national electricity values and hard to justify ballpark estimates to produce estimates that vary often by more than an order of magnitude.
To do this derivation from the Ethereum estimates, we must make an assumption that is easy to justify. The Ethereum network and Zilliqa network share the same PoW algorithm — Ethhash — which comes with ASIC resistance and several other factors that make it justifiable to approximate the characteristics of the mining pools for Zilliqa and Ethereum are linearly proportional — in other words, Zilliqa’s hashpower comes from roughly the same types of machines, running with similar mixes of electric power sources to the Ethereum mining network. This is easier to justify given that some of Zilliqa’s largest mining pools are dual-mining Ethereum when the Zilliqa DS block is over for the Epoch and their hardware would otherwise be idle.
Where Zilliqa is Thrifty
Zilliqa uses PoW, but not necessarily for the same things that you expect. Proof of Work is part of the block minting process on Ethereum and Bitcoin, but Zilliqa only uses Proof of Work to mint the DS block and organize the shards during the beginning of an Epoch. This means that running over the same interval of 60 minutes, Ethereum and Bitcoin will be using 100% of their hashrate to solve puzzles in PoW, and Zilliqa depending on the interval might not use PoW at all as it only needs PoW for one 60 second cycle at the start of a new Epoch for leader election and to resort all nodes into new shards. After that, there is no need to solve puzzles to process Zilliqa transactions, only an online machine with minimal additional resources is required to validate transactions for the network and earn $ZIL in portion to your contribution.
Because the TX and DS blocks are not entirely predictable, we will not get a perfect number for the conversion factor, but due to the designed rate of block production we can use it as an estimate with a new DS block happening every 1–2 hours. We will use a sanity check value from viewblock shows that the last Epoch with DS block 10864 lasted for 85 minutes).
This means that the entire Zilliqa network needed to run PoW cycles for 1/85 or 1.17% of the time Ethereum was running them just to achieve the same effective network hashrate during that time. We see an even greater reduction when we consider that as of today Ethereum’s hashrate is at least 540 TH/s and Zilliqa DS block hashrate is somewhere around 4 TH/s , so we can estimate that Ethereum’s hashrate demand is roughly 135 times Zilliqa’s. Already we can see that Zilliqa’s carbon footprint could probably fit into the margins of error around Ethereum’s, even without controlling for the difference in transactions processed at that hashrate. So to recap, due to Zilliqa’s sparing use of PoW and smaller miner network we can already say it’s PoW ‘useless puzzle’ is causing roughly 0.009% of the carbon to enter the atmosphere as Ethereum for a given span of time. Let’s use the digiconomist Ethereum stats to get a ballpark estimate for Zilliqa’s impact in 2021:
This year Ethereum is expected to emit roughly 14.65 Mt CO2 due to its mining activities (comparable to Tanzania) , meanwhile Zilliqa according to this estimate has to atone for 0.009% — a measly 1277 tons CO2. In context this is the carbon footprint of about 85 Americans in 2016.
Carbon emissions from Zilliqa mining is already a very manageable issue. To offset these values we would need to buy offsets from the EUA ICE future exchange for 42 euros per ton, for a total of €53,634 to declare Zilliqa’s PoW activities for this year carbon-neutral. Offsetra would sell us the credits at £7 for a total of about £9K.
In context, Ethereum is on the hook for €615 million this year. Bitcoin has to offset about 2.6 times as much carbon as that, for about €1.6 billion this year.
Carbon Footprint of a Single Transaction
Now that we have a useful ratio, we can reasonably assume that it holds for individual transactions too, with some small adjustments. We don’t know what a saturated Zilliqa network looks like at any hashrate as it’s never been congested for long enough to get a good measure. We cannot say with certainty what the footprint per transaction is because most blocks mined so far on Zilliqa had room for more transactions. Fortunately, someone spammed the network on Feb 4th, 2021, so we have at least a lower-bound on how many transactions the network can churn out at that hashrate:
So then for this estimate, we will assume that the Feb 4th day is tightly bounded, and use the total number of DS blocks along with the total number of transactions as a proxy value to derive the ratio of how many transactions Zilliqa gets per the hashrate it consumes compared to Ethereum. Since 100 TX epochs occur per DS epoch, we can just use the day’s blocks as an estimate for how many PoW cycles were run:
On Febuary 4th Zilliqa made roughly 15 DS blocks at a hashrate of 2.88 TH/s and processed 1,047,000 transactions. Now we can work with this data to derive an estimate for the Carbon footprint of an individual Zilliqa transaction compared to Ethereum’s on that day. Averaged over that strange day this translates into Zilliqa rolling at roughly 12 TPS, so we know it isn’t saturated and the actual carbon per transaction of a fully-utilized Zilliqa network is likely much, much lower. However better data is not available yet, and we can take the hit of a high estimate — we have the wiggle room in our carbon accounting to be generous.
Thus, If we use Etherscan’s estimates for that day, we can get a picture for the proportional guilt I should feel for staking my $ZIL.
Carbon Footprint Per-Transaction Breakdown
Ethereum: 31,610 gCO2 per Tx in 2021
Because Zilliqa only used PoW for roughly 15 minutes of that day (1.04%) for the DS blocks, we will reduce the per-transaction cost by that factor first. 31.61 kgCO2 * 0.0104 = ~0.329 meaning that Zilliqa, without factoring in sharding’s hashrate savings would only emit about 329 grams of CO2 per transaction with the Ethereum network’s equivalent mining power.
Let’s further reduce this by the factoring in the difference of mining power consumed:
2.88 TH/s /368 TH/s = 0.00782608695 (hashrate ratio)
329 gCO2 * 0.00782608695 = 2.57gCO2 just accounting for the reduction in required hashpower from sharding.
Now given that Ethereum processed more transactions in that period, the last step is to scale this number by that factor:
1,310,000 ETH TX/ 1,050,000 ZIL TX = 1.248 (Tx ratio)
2.57gCO2 * 1.248 = 3.21 gCO2 per Tx on Feb 04.
Therefore, on February 4th if the Ethereum and Zilliqa network was experiencing a typical 2021 day, an Ethereum transaction emitted nearly 1,000x a Zilliqa transaction, and this reduction is only going to go up as the Zilliqa network utilization in TPS goes up.
Context is Everything
Given these small values, it is hard to compare them to the explosive claims about Bitcoin and Ethereum’s nation-scale energy needs. In fact, only one country on Earth emits less carbon than Zilliqa, and it’s Tuvalu with its 0 emissions and 12,000 inhabitants.
So if we cannot compare Zilliqa’s awesome wastage to a country, what should we compare it to?
Well, using estimates for brisk walking this is about 2 minutes of a typical human walking for them to exhale a Zilliqa transaction, and 1.4 days for an Ethereum transaction.
Or we can use this calculation to say that a Zilliqa transaction is equivalent to about ~87 breaths from a 75kg healthy human , while Ethereum is churning out CO2 equivalent to ~858,000 breaths per transaction.
Or if you’d prefer — A Zilliqa transaction is about 16 Google searches, or 160 retweets of this sweet blog post.
Therefore if the Zilliqa network and Ethereum network performance on Feb 4th is a rough portionality benchmark, and we extrapolate that to the rest of the year, Zilliqa can claim at least an expected thousandfold reduction in per-transaction CO2 compared to ETH in 2021.
Into the Future
My ambition for the Zilliqa network and my [hopeful] reward for my devotion to the Zilliqa project is to design the mechanism that Zilliqa will incorporate to ensure the offsetting its carbon as an autonomous community. I don’t believe Zilliqa Research should do one-off offset purchases on our behalf in a unilateral manner — instead every token holder with the interest to be involved in governance should have a say in how we handle this issue. I have noted that the circular economy notion has begun to take root in the community, even informing the design of our non-protocol staking mechanism.
I hope to attract more of the circular economy enthusiasts to Zilliqa, to join my quest to build out Zilliqa into the Ecosia of blockchains. I want to be able to confidently say that Zilliqa isn’t just net carbon neutral, but that using this protocol will draw carbon out of the atmosphere due to the added value of your activities.
For my last words on our PoW cousins — I’ve always found it bizarre that BTC, ETH and other blockchain’s energy consumption is contextualized as either mineral production or compared to the vacuum of space. In reality it should be compared to its competitors — the global swarm of fatcats zipping around daily in private jets, their dutiful army of armored trucks, their horde of electrified vaults, skyline of air-conditioned sky scrapers, and herd of commuting office drones. When we can say that our alternative requires none of that and that using it draws down carbon, then which is truly wasteful?
I look forward to a future day where we can confidently say, “No, Zilliqa isn’t dirty money. In fact, our protocol prints Mangroves!”