Contributed Post from the Community:
Hot topic · builder-side · non-market · drawn from X and related sources
Multicast paths before the node catches up
The practical edge since the last check is network-side measurement, not only node release notes.
LightBSV has been stress-testing pure transaction data delivery over multicast — no Teranode or Aerospike in the path yet — to see what the wire can carry before the full stack is in the loop. At 100 Gbit the framing is on the order of 60 million+ TPS per link, with sharding and stacking of links in the design, and an aim to push toward 100 million TPS-class network capacity ahead of Teranode’s own path. Separate notes put a working 1M TPS service estimate around 15–20 modern servers, including Aerospike nodes with NVMe. Full Teranode/Arcade/Broadcaster overlay testing is described as the next step, with conforming transaction, subtree, and block formats on the wire as the interoperability bar.
Teranode’s public thread has stayed educational: gRPC as binary, streaming service communication for high message rates; Merkle trees as the summary structure behind subtree roots; UTXOs as the ownership unit Aerospike is meant to track at large scale. The longer roadmap framing remains Chronicle restoring original protocol behaviour and Teranode scaling the base layer.
Around wallets and indexing, the same scale argument continues: peer-to-peer exchange and proof delivery versus architectures that need a top-down view of the whole chain. Broadcaster is reiterated as Arc API compatible; BRC-100 remains framed as an optional wallet-to-app interface, not a veto over other clients.
The current pulse is capacity planning on the network path — measured multicast delivery first, full node integration next — while the stack education and anti-scan-everything case keep running in parallel.
Source Links
Leave a Reply