Telemetry report

[wilyam]

Subspace Telemetry Report

The monitoring of the Subspace Telemetry dashboard lasted 14 days from August 3 to August 17, 2025 to observe patterns in node behavior. Below are the key findings.

1. Node Count and Network Load

The number of active nodes on Autonomys Mainnet declined from roughly 837 to about 800-802 during the observation period. This confirms a slight reduction in participation but overall network stability.

Block production accelerated, with the best nodes now reaching heights above #4,029,900 at an average block time of ~6-6.3 s (faster than the earlier ~7-8 s).

Transaction queues were consistently near zero, indicating the network has ample capacity and is not experiencing congestion.

Anomalies such as high-altitude nodes or traveling nodes were not observed. All nodes appeared to have stable positions; none hopped across continents.

2. Client Implementation and Versions

During the period, the installed base appears split roughly evenly between Space Acres v0.2.x and CLI/subspace-node v0.1.x.

Naming conventions. GUI-launched setups often use automatically generated adjective-noun-number labels (e.g., curious-tub-3518), while CLI operators more often use human or sequential identifiers (e.g., king, Zhengwu-02, iron-19).

Operational nuance. Both stacks contribute similarly to network connectivity because they run the same underlying components (node + farmer). In practice, long-uptime/high-peer hubs are more frequently CLI-managed machines, whereas Space Acres deployments are more numerous but tend to keep fewer peer connections per node.

3. Peer Count Distribution and Super-Nodes

Peer counts vary widely. Most nodes have fewer than 100 peers, but several super-nodes act as hubs with hundreds of connections.

Earlier in the period, the node uneven-hot-6218 reached 702 peers, followed by shallow-wind-2968 with 680 peers (located in Helsinki), and black-sleep-7166 with 582 peers. These hubs support many farmers and act as connectivity anchors.

In the latest snapshot, no new 600+ peer supernodes appeared. The largest were:

grubby-can-2996 in Columbus, USA with nearly 400 peers

rich-brush-8292 in Lombardy, Italy with over 180 peers

This shows that while extreme supernodes are rarer, strong regional hubs still exist.

4. Node Uptime (Oldest vs Newest)

The column in Telemetry is representing node uptime (days online) revealed “veteran” nodes. modern-ball-9951 (Ashburn) and mysterious-engine-9096 (Hefei, CN) had uptime values around 570 days, while righteous-star-8957 (Ashburn, USA) and utter-popcorn-1118 (Hefei,CN) exceeded 220 days. These long-running nodes often reside in Ashburn (USA) or Hefei (China) and use subspace-node. Conversely, new nodes frequently have uptimes under ten days and minimal peers.

5. Transaction Queues

Sorting by the “Transactions in Queue” column showed that most nodes have zero or very few pending transactions. A handful of nodes, such as righteous-star-8957 and utter-popcorn-1118 (Hefei), briefly displayed queues exceeding 200 transactions. These spikes may indicate heavy local usage or network delays but were rare.

6. Naming Patterns and Possible Ownership Clusters

There are two distinct naming schemes:

Automated adjective-noun-number names. Many farming nodes running Space Acres are labelled with whimsical combinations like curious-tub-3518 or wanting-scent-5279. The geographic distribution shows clusters in China and Taiwan, suggesting centralised multi-farm deployments or a dense concentration of independent participants.

Human or sequential names. Examples include iron-18..20, Zhengwu-01..11, MW-01..1, node-111, node-115. Such series strongly imply coordinated groups of nodes under common ownership. Many sit in the same city (e.g., Zhengwu nodes in Zhangzhou), hinting at organized private clusters.

7. Geographic Distribution and the “Subspace Belt”

The map reveals a striking belt of nodes between 30° N and 50° N latitude. Dense clusters exist in China (Hangzhou, Hefei, Chongqing, Beijing, Guangzhou), Taiwan (Tainan and Taipei), Vietnam, Thailand, Russia and Eastern Europe (Moscow, Kyiv, Warsaw), and the eastern United States (Ashburn, New York). This “Subspace Belt” shows strong adoption in East Asia and Eastern Europe, forming two major “super-clusters” around China/Taiwan and Russia/Ukraine.

7.1 United States Nodes

In North America, nodes are mainly concentrated on the east coast, notably in Ashburn (Virginia)

and surrounding areas. Few nodes appear on the west coast. Ashburn’s popularity likely stems from major cloud providers (AWS, Azure, Google) hosting data centres there.

7.2 Asian Super-Cluster

The largest concentration of nodes is in China and Taiwan, forming what could be called the Asian Super-Cluster. Cities like Hefei, Beijing, Chongqing and Guangzhou host dozens of nodes each. Taiwan’s Tainan and Taipei also feature many nodes. This cluster likely reflects a combination of **accessible hardware**, enthusiastic farming communities and possibly co-located hosting facilities.

7.3 Eastern European and Russian Cluster

Moscow, Kyiv and Warsaw form the main hubs. Nodes here often show coordinated naming, suggesting organized operator groups.

Together with Asia, this creates a visual symmetry - two hemispheres of decentralization balancing across the globe.

7.4 Southern Hemisphere Outliers

Single points in Brazil, South Africa and New Zealand appear as lonely farmers.

7.5 New and Emerging Locations

Fresh observations revealed additional nodes in:

Lagos, Nigeria - a rare presence in Africa, expanding the network into the southern hemisphere (this node appeared on the last day of observation).

Dublin, Ireland and Paris, France - strengthening presence in Western Europe.

Verona, Italy and Menzingen, Germany - growing Western and Central European distribution.

Halifax, Canada and Columbus, USA - extending coverage in North America.

Wellington, New Zealand - notable as one of the first nodes in Oceania.

Columbus, USA - grubby-can-2996 (≈ nearly 400 peers).

Lombardy, Italy (region) - rich-brush-8292 (180+ peers).

Helsinki, Finland - shallow-wind-2968 (~680 peers).

Hobart, Tasmania – 1 island node.

These additions demonstrate a gradual globalization of the network, though the majority of nodes remain concentrated in the Northern Hemisphere.

8. Geopolitical factor:

In general the Subspace belt coincides with regions where US/EU control is weaker, but the community is strong (Eastern Europe, Asia). Affordable equipment is also an important factor.

9. Data centers and risks

Helsinki (Finland): A notable outlier is the set of super nodes in Helsinki with hundreds of peers (e.g., shallow-wind-2968, peaking at ~680 peers). Their location strongly suggests they are hosted in Hetzner’s data centers. Since Hetzner explicitly prohibits blockchain, mining, and farming activities, such nodes risk abrupt termination if detected. That makes them precarious “super-hubs”: they provide massive connectivity today but could vanish suddenly, impacting peer topology.

Ashburn (USA): There is also concentration there, but based on AWS/Azure/Google Cloud. There is no official ban, but there is a risk of high dependence on hyperscaler providers.

Ashburn = the largest traffic exchange hub in the US, with a high density of AWS/Azure/GCP.

Risks:

Centralization point: many “hubs” are located on the infrastructure of 2-3 providers. Single point of failure: regional failures/incidents in the cloud → part of the peer graph “collapses” immediately.

Provider policies: tightening of ToS/ports/anti-abuse can affect peering connectivity.

Network and price shocks: egress tariffs/route policies/peering changes at hyperscalers → increased cost/degradation of connectivity.

Jurisdiction: US regulatory requirements/DMCA abuse flow → possible shutdowns faster than with independent hosters.

Conclusion: high dependence on hyperscalers might impose a systemic risk (not the same as a direct “ban” like Hetzner, but vulnerability in fault tolerance and policy vector).

10. Traveling nodes

No anomalies such as high-altitude nodes or traveling nodes were observed. All nodes appeared to have stable positions; none hopped across continents.

Still there are cases looking like usage of a VPN.

------------------------------------------------------------------------------------------------------------

Additional Data:

Top 10 super hubs - nodes with >500 peers:

shallow‑wind‑2968 - Helsinki (≈670 peers).
uneven‑hot‑6218 - UnknownLocation (≈665).
black‑sleep‑7166 - Ashburn, USA (≈600).
shaggy‑history‑0290 - Ashburn, USA (≈585).
didactic‑argument‑7334 - Ashburn, USA (≈588).
mysterious‑engine‑9096 - Ashburn, USA (≈575).
spicy‑order‑8901 - Ashburn, USA (≈569).
bawdy‑cakes‑5345 - Beijing (≈566).
modern‑ball‑9951 - Ashburn, USA (≈562).
learned‑place‑0669 - Columbus, USA (≈398).

Nodes with 350-250 connections:

grubby-can-2996, conscious-meal-0867, tender-curtain-3820, hk65, mere-apples-9696, iron-18/19, Zhengwu-01/02, subspace-node04, AUT-12, overwrought-hose-2247, utter-popcorn-1118 (Hefei), domineering-picture-5478, righteous-star-8957 (Ashburn), Zhengwu-01, Zhengwu-02, etc. These nodes are located in the US (Ashburn, Columbus), China (Beijing, Zhangzhou), and Finland (Helsinki). Those most likely are DCs, again.

Nodes with the minimum number of peers (0-2) - at the bottom of the list:

massive‑snail‑9699 (Zaporozhzhya, UA) - 0 peers.
material‑industry‑9970 (Khon Kaen, TH) - 0 peers.
ripe‑work‑6690 - 0 peers.
permissible‑transport‑9567 (Columbus, USA) - 0 peers.
possible‑rainstorm‑1553 (Guiyang, CN) - 1 peer.
sczg01 (Los Angeles, USA) - 2 peers.

These nodes are practically isolated and do not participate in the peer network.

“Lonely farmers” and new geographies.

Individual nodes have been discovered in Lagos (Nigeria), Dublin (Ireland), Paris (France), Verona (Italy), Menzingen (Germany), Halifax (Canada), Wellington (New Zealand), as well as in Brazil and South Africa. They stand out on the map like islands in the southern hemisphere.
delicious-town-5274 (Curitiba, Brazil) - the only point in Brazil.
homeless-ant-3143 (Foshan, CN) - only one instance was spotted in Foshan.
fantastic-opinion-9398 (Kaohsiung, TW) - there is also only one node in Kaohsiung.
Hobart, Tasmania – 1 island node.

Old and New nodes (uptime).

The oldest nodes (uptime > 200 days) are modern-ball-9951 (Ashburn), mysterious-engine-9096 (Hefei), righteous-star-8957 (Ashburn), and utter-popcorn-1118 (Hefei). They have been operating for 570+ days (the first two) and 220+ days (the latter).

The newest nodes (uptime < 2 days) .
massive-snail-9699 (Zaporozhzhya, Ukraine), material-industry-9970 (Taipei, Taiwan), ripe-work-6690 (CN), permissible-transport-9567 (Warsaw, Poland), possible-rainstorm-1553. These are newcomers with 0-1 days of operation.

The newest nodes are those that have appeared in the last 24 hours (massive “invertebrate” names).

When a Space-Arces client coined a name linked to node behavior.

modern-ball-9951 Veteran, stable, one of the oldest nodes

mysterious-engine-9096 Veteran engine, runs steadily for 1.5+ years

righteous-star-8957 A “star” that shines with bursts of activity

massive-snail-9699 0 peers, completely lonely, slow/lonely vibe

Some more are: bawdy-cakes-5345, weird-breath-6698, overjoyed-underwear-8637, curious-tub-3518, fantastic-opinion-9398, nosy-pest-9311, bedwetta, adorable-grass-7777, affectionate-carrot-1234, etc. Funny names mostly belong to Space Acres farming client.

Old client versions.

Currently, almost all nodes use subspace-node v0.1.1 or space-acres v0.2.14. Older builds (subspace-node v0.1.0 or space-acres v0.2.0 - 0.2.12) are rare and mainly distributed across China (Zhangzhou, Hefei) and Russia/Ukraine. Old nodes potentially affect network performance. Old versions clients are rare and mostly concentrated in East Asia.

The region with the largest number of old nodes.

Among the nodes running old versions, Chinese cities (Zhangzhou, Hefei, Hangzhou) and Russian/Ukrainian cities (Rostov, Kyiv) predominate. These farms were deployed a long time ago and have not been updated.

Fast and slow network.

In the latency (ms) column, most nodes show 0-30 ms. Slow channels are isolated nodes in China: several nodes have a delay of ~3000 ms and above: short‑skate‑0499 (Xiamen) and moaning‑taste‑8170 (Shenzhen) up to 6000 ms; they operate on a weak channel.
The leaders in terms of minimum latency (0 ms) are most of the Ashburn and Helsinki nodes.

Transactions in queue.

Almost all nodes hold 0-1 transactions in queue. The leaders in terms of maximum queue remain:

righteous‑star‑8957 → Ashburn, USA - periodically accumulated up to ~220 transactions in queue.

utter-popcorn-1118 → Hefei, China - up to 200+.

These nodes consistently have long queues, which indicates high load or configuration issues; Is not possible to determine the exact source of the load using Telemetry.

They are also among the “oldest” - uptime 220+ days.

Possible VPN usage:

possible-rainstorm-1553 - in one section it appeared as Moscow, RU , in another - as Guiyang, CN.

This means that either the node was re-registered/migrated, or there was simply an IP geolocation error, or a VPN.

August 17, 2025.

[wilyam]