Introduction

The CSM Testnet performance evaluation concluded on September 30, 2024, shortly followed by the publication of the CSM Mainnet Early Adoption (EA) list. One pathway to EA list inclusion was through participation as a node operator in the testnet, maintaining both good performance and sufficient participation time.

• Performance was measured as the percentage of successfully completed duties out of total assigned duties. Good performance was defined as exceeding 86.60%.

• Sufficient participation time was set at exceeding 6750 epochs, i.e. 28 days.

Performance data was published by Lido, alongside detailed CSM performance criteria.

This post offers a brief descriptive analysis of node operators' performance in CSM's Testnet based on the aforementioned performance data. You can find all the source charts referenced in this post in this dashboard.

Operator and Validator Counts

A total of 2,649 node operators participated, registering 8,086 validators. This averages to about 3 validators per operator. However, as we'll demonstrate later, the distribution is highly skewed.

Let's first examine how many operators achieved both good performance and sufficient participation time.

Of the 2,649 node operators:
• 2,440 (92.11%) achieved both good performance and sufficient participation time
• 123 (4.64%) did not achieve sufficient participation time
• 86 (3.24%) did not achieve good enough performance

When we disaggregate the data by validator count, the proportions differ significantly.

Of the 8,086 validators:
• 4,056 (50.16%) achieved both good performance and sufficient participation time
• 3,237 (40.03%) did not achieve sufficient participation time
• 793 (9.80%) did not achieve good enough performance

This shift in distribution is primarily due to two outlier node operators who managed 3,000 validators but didn't participate long enough to achieve sufficient participation time.

The following figures illustrate the significant skew in the validator-per-node-operator distribution:

Participation Time

Now, let's examine the participation time data. Recall that the sufficient participation time was set at 6,750 epochs.

The figure shows that most node operators participated for 8,400 to 11,400 epochs, thus achieving sufficient participation time.

Performance

Finally, let's analyze the performance numbers. Remember, the performance threshold was set at 86.60% of successful duties out of all assigned duties.

In this figure, we observe a heavily skewed distribution towards 100% performance. Notably, there are 33 operators with zero participation.

Zooming into the 90–100% performance range, we see the distribution remains heavily skewed towards excellent performance—99% and above.

Further focusing on node operators with 99–100% performance, we find that most achieved between 99.9% and 99.98%. Remarkably, 16 node operators attained perfect 100% performance.

Summary

• 2,649 node operators participated in the CSM Testnet, managing 8,086 validators.

• 92.11% of node operators achieved both good performance (>86.60%) and sufficient participation time (>6,750 epochs).

• The distribution of validators per node operator was highly skewed, with two outliers managing a combined 3,000 validators.

• Most node operators participated for 4,410 to 5,985 epochs, exceeding the sufficient participation time threshold.

• Performance was heavily skewed towards excellence, with many operators achieving 99% and above.

• 16 node operators achieved perfect 100% performance.

• The high level of competence demonstrated by node operators suggests a promising outlook for the CSM mainnet launch.

You can look at all the aggregated data in our Dune dashboard.

In conclusion, the CSM Testnet analysis reveals a highly successful and promising foundation for the upcoming Mainnet launch. The overwhelming majority of node operators demonstrated exceptional performance and commitment, with over 92% meeting or exceeding the stringent criteria set for good performance and sufficient participation time. This level of excellence bodes well for the future of the Lido ecosystem, suggesting a robust, reliable, and efficient network that can support the growing demands of decentralized finance.

The high performance across a diverse range of operators, indicates a healthy and resilient network. The fact that many operators achieved near-perfect or perfect performance scores is particularly encouraging, as it suggests a strong dedication to maintaining optimal system functionality.

As the CSM moves towards its Mainnet launch, these results provide confidence in the system's ability to handle real-world challenges and maintain high standards of operation. The Lido community can look forward to a stable, secure, and efficient staking solution that leverages the collective strength of its well-prepared node operators.

About Firstset

We commit our intellectual, social and computational capital to help bootstrap the cryptoeconomic networks of tomorrow. We are an independent team of crypto-native node operators and builders with a mission to support emerging chains and other kinds of decentralized networks from day one.