NEAR Performance
Transaction performance of NEAR compared to other blockchains.
:city_sunset: About NEAR
- NEAR is a layer 1 blockchain that uses Nightshade technology to process transactions in parallel across multiple sharded chains, thereby improving the overall capacity of the blockchain.
- NEAR Protocol claims that it is able to achieve up to 100,000 transactions per second and achieve near-instant transaction finality with a one-second block cadence, while simultaneously keeping transaction fees at virtually zero.
Reference: \n
:telescope: Scope of Analysis
With all the claims about its superior performance, how is NEAR currently performing? How does it stock up to other L1s in terms of speed and performance?
Specific Questions
How fast is NEAR in comparison with other L1 blockchains (Ethereum, Avalanche, Osmosis and Solana) in terms of:
-
transactions per minute
-
block time
-
transaction success rate
\
:straight_ruler: Approach
- Chains compared in this analysis:
- Chains selected were Layer-1 (L1) chains with considerable TVL and user base
- Tables used:
fact_transactions, transactions
- Period: transactions performed in the last 14 days
Metrics Explored
- Transactions per minute: total number of transactions processed daily (failed and succeeded) divided by seconds per day
- Block time: time taken to form 1 block - seconds per day divided by blocks formed daily
- Transaction success rate: successful transactions divided by total transactions performed daily, multiplied by 100
TPS
-
Comparing the TPS and block time across chains, currently, the transaction speed is the highest on Solana, at 569 tx per second on average.
-
This is followed by Ethereum at 13.64 transactions and NEAR at 3.94. Note that the Avax transactions do not include subnet transactions.
-
The TPS observed on NEAR currently is only a fraction of what the chain claims it is capable of processing
From the deeper dive into Osmosis, we see a sudden surge in failed transactions from 1100 hours on 10th July.
The surge in TPS was mainly due to these failed transactions performed on Osmosis, which were likely from bots, given that the unique user count did not increase in the same volume.
Generally, we see a fair bit of fluctuation in TPS and block time across chains.
The most significant was on Osmosis, between 9th and 10th July. Here, we identify a sudden surge in TPS from 1.21 to 3.06 on 10th July. Prior to 10th July, TPS On Osmosis was between 1.4 to 1.8.
Corresponding to this surge in TPS, we see a sharp drop in the successful transaction rate on Osmosis, from ~80%+ to 35.7% (below). Despite the fluctuation in TPS and transaction success rate, the block time on Osmosis stayed around 6.5s through this period.
:mag_right: Key Findings
Overall, amongst the 5 L1s compared, Solana has the highest TPS (500+) and fastest block time (0.7s).
NEAR comes in second for block time (1.2s) and has a TPS (3.94 tx/s), which is rather poor compared to what it claims to be able to achieve.
Transaction success rate was highest on Ethereum (>95%) followed by Avalanche (~90%). NEAR performed relatively poorer (ranging between 74% to 86%) and was more varied in transaction success rate.
Block Time
-
In terms of block time, Solana has the fastest block time of < 1s, followed by NEAR at 1.24 seconds.
-
This was close to what NEAR claimed for its one-second block cadence.
Overall, NEAR comes in second for block time and has rather poor TPS compared to what it claims to be able to achieve.