What should distributed systems look like?

Dr Wright began the session by discussing BTC and noting that in this system nodes do not engage in direct peer-to-peer transactions.

The Bitcoin Masterclass (Slovenia) Day 2

Dr Craig S. Wright, Chief Scientist at nChain, held the second edition of his Bitcoin Masterclass series at an exclusive venue in Slovenia in February. The Masterclass course forms part of a monthly series aimed at helping attendees understand the fundamentals of Bitcoin and the technology behind it.

The Bitcoin Masterclass series will give you a comprehensive overview of Bitcoin’s history, theory and design. Wright discussed the future of Bitcoin and the unbounded scalability of nChain’s blockchain solutions, with potential use cases across several industries.

In the third session on the second day of his Slovenian Masterclass, Dr Wright discussed multicast and IP2IP. He also discussed the structure of distributed systems and addressed how we might link, index, and search for information on them.

Dr Wright began the session by discussing BTC and noting that in this system, nodes do not engage in direct peer-to-peer transactions, but instead transmit information to Internet Service Providers (ISPs) for relaying.

He noted that in this system there is the possibility of losing connectivity if one link fails and suggests the importance of establishing multiple links. He explains that while increasing the number of links incurs higher costs, it significantly enhances overall resilience.

Dr Wright emphasises that just as Bitcoin itself is decentralised, individual nodes can also be distributed. He refers back to a previous lecture where he discussed the resilience of applications located in different areas, such as data centres and cities.

The physical importance of nodes

Dr Wright also discussed the significance of the physical locations of nodes, the underlying infrastructure, and their interconnectivity. In a three-node configuration, multicast can be employed to ensure that the information broadcasted by the main node reaches the other two nodes, he said.

With a User Datagram Protocol (UDP), if information fails to reach other nodes, it simply does not reach them. Any detection of missed information would need to be handled at the application level. By comparison, the Reliable Data Protocol (RDP) facilitates the retransmission of packets (including through anycast) until the other nodes successfully receive them.

All these considerations are essential when designing systems that function within a corporate network or as part of applications, he said.

Other topics discussed during this session include:

  • Building resilient systems;
  • Key value databases;
  • How the music industry could benefit;
  • Payments in Bitcoin;
  • Hive-based applications;
  • Creating redundant applications;
  • Automated document processing;
  • The Diffie-Hellman process.

You can watch the full Masterclass session below: