Stability and Scale
This page presents operational signals from the live BDS mainnet market. All figures are derived from on-chain events, copied epoch tallies, or direct measurement of the production network.
Network at a Glance
The BDS mainnet market has been running on Ethereum mainnet for 58 days (as of this writing). The measurement window below covers the first 30 days of production operation.
Figure 1: Daily batch submission completion rate over the first 30 days of production. The network maintained 97.70% completion across 217,295 total epochs.
Key Production Metrics
| Metric | Value | Source |
|---|---|---|
| Total epochs assigned | 217,295 | On-chain priority assignments (Day 1-30) |
| Epochs with submissions | 212,297 | Snapshot batch submitted events |
| Completion rate | 97.70% | Derived from above |
| Missed epochs | 4,998 | Gaps in submission coverage |
| Median end-to-end latency | 89 seconds | Priority assignment to on-chain finalization |
| Mean latency | 88 seconds | Same measurement window |
| P95 latency | 138 seconds | Tail latency measurement |
| Full steady-state days | 26 of 30 | Days at 7,200 epoch cadence |
The 2.3% gap in coverage clustered around brief restart windows rather than being evenly distributed. Once a validator commits to submitting a batch, it completes 99.01% of the time.
Validator Performance
Six validator nodes maintained 22 or more consecutive days of operation during the measurement window.
Figure 2: Validator node constellation showing completion rates and active days. Node size indicates total submissions; color indicates reliability tier.
Stable Validators (22+ days active)
| Node | Active Days | Completion Rate | Total Submitted |
|---|---|---|---|
| Node 1 | 30/30 (100%) | 99.87% | 29,670 |
| Node 5 | 30/30 (100%) | 99.90% | 49,717 |
| Node 4 | 29/30 (96.7%) | 99.20% | 27,252 |
| Node 3 | 28/30 (93.3%) | 97.42% | 33,827 |
| Node 10 | 26/30 (86.7%) | 98.01% | 27,690 |
| Node 11 | 22/30 (73.3%) | 98.79% | 25,319 |
Three additional transient nodes participated for shorter periods during operator testing and rotation.
Across the six stable long-running nodes, the aggregate completion rate was 98.95% over 193,475 total submissions.
Submission Latency
Latency is measured from priority assignment (validator selected for epoch) to on-chain finalization (batch submission confirmed).
Figure 3: Latency distribution by hour of day. The density plot shows 212,297 submissions with median 89s, mean 88s, and P95 138s.
The heatmap reveals consistent latency patterns across all hours, with no significant degradation during any specific time window. This indicates stable validator infrastructure and reliable RPC connectivity throughout the measurement period.
Epoch Coverage Pattern
The coverage matrix shows submission activity across all epochs in the 30-day window.
Figure 4: Day-by-epoch-block coverage matrix. Bright green indicates high submission counts; dark cells indicate gaps. The matrix shows sustained coverage with isolated gaps clustering around operational transitions.
The matrix demonstrates that gaps were not random failures but clustered around specific operational events (node restarts, configuration updates). The network recovered full coverage within hours of any interruption.
Snapshotter Participation
A separate measurement of 1,811 epochs (April 17–24) shows the snapshotter layer participation:
| Metric | Value |
|---|---|
| Unique eligible nodes per day | 2,970 – 3,047 |
| Median eligible nodes per epoch | 305 |
| Active validators per epoch | 3–5 (median 4) |
| Minimum validators observed | 3 (every measured epoch) |
Figure 5: Radial distribution of slot activity from sampled epochs. The visualization shows 3,078 total slots with activity levels categorized by submission count.
The slot distribution shows broad participation across the slot ID space (0–8191), with no concentration that would indicate centralization or collusion patterns.
What These Figures Show
Continuous cadence under pressure
BDS mainnet processes one-block epochs on Ethereum mainnet. This creates continuous pressure on transport, aggregation, and finalization rather than allowing activity to concentrate into infrequent batch windows. The 97.70% completion rate across 217K epochs demonstrates sustained operation under this pressure.
Observable mesh health
The published local-collector setup includes explicit health states:
healthywhen both discovery and submission topics have at least 2 peersdegradedwhen one side falls below thresholdprunedwhen one side has 0 peers
These definitions make mesh health observable rather than assumed. The live measurements show 5 healthy submission-topic peers and 162 total connected peers in sampled collector logs.
Measurable validator participation
The network shows:
- 3 to 5 validators present per tallied epoch
- ~3,000 eligible nodes receiving rewards daily
- 98.95% aggregate completion across stable validators
- 99.01% completion rate once submission commitment is made
This provides concrete evidence of multi-validator participation at production scale.
Competitive Position
Many data products can show a dashboard. The DSV network can show:
- Published epoch workflow with timing guarantees
- Market-scoped on-chain finalization
- Public verification calls for exact finalized CIDs
- Operator-visible mesh health thresholds
- Validator-priority-controlled settlement
- 58 days of sustained production history
That combination is what makes DSV interesting as a production system rather than a theoretical architecture.
Related Pages
Methodology and Reproducibility
All figures are derived from:
- On-chain events on Powerloom mainnet (chain 7869)
- Copied epoch tallies from
snapshot-activity-tracker-updater(1,811 epochs) - Validator activity analysis from
dsv-validator-activityscripts (30-day window)
Contract addresses and full methodology: github.com/powerloom/curated-datamarkets
Last updated: 2026-04-30