Sharing weather data autonomously using Fetch.ai — Summary

• Agents interact via a peer-to-peer network, with the agents representing data advertising their services to agents seeking that data.
• Agents then negotiate a micropayment that will be paid in exchange for the data.
• Trust ratings share information about agents’ track records and highlight those which fail to fulfill the obligations of a transaction. This ensures agents are incentivized to act honestly.

In this series of articles we have examined how Fetch.ai agents negotiate and transact with each other. In order to provide as simple an explanation as possible, we used an example of a girl (Charlotte) in Cambridge who wanted to know what the weather was like in London, ahead of her journey to meet friends in the capital.

Using today’s technology, Charlotte could use a mobile app to learn about the weather, before using several other apps to find out information such as the quickest way to travel into London and the restaurants she could visit that are located close to nearby train stations. However, by using Fetch.ai’s network, Charlotte doesn’t need to rely on numerous apps on her phone. Instead, her agent can provide the answers to all of her questions by finding the information autonomously.

To discover what the weather is like in London, Charlotte’s agent searches for another agent that could provide this information. This search is made possible by Fetch.ai’s Open Economic Framework (OEF). This acts as a decentralized search and discovery mechanism. You can learn more about this core component of the Fetch.ai network by reading part one of this series of articles. Agents representing data are continuously advertising their services on the OEF. This enables Charlotte’s agent to discover and connect instantly with an agent representing data collected by a weather station in London. The two agents would then begin to negotiate the price the weather station agent would receive from Charlotte’s agent in return for the data. If the weather station sets the price too high, Charlotte’s agent will end the negotiation and find another agent that offers information about the weather in London at a better price.

When Charlotte’s agent has successfully negotiated a price she is willing to pay (i.e. a fraction of a cent) for information about the weather in London, the transaction process begins. The trade occurs using the Fetch.ai smart ledger. If this ledger didn’t exist, the agents would need to act in good faith, blindly trusting that other agents would fulfill their own obligations. They would also need to rely on mechanisms outside of the Fetch.ai network in order to settle transactions (e.g. by bank transfer). This would be impractical for a number of reasons.

Instead, by using the smart ledger, the agent representing the weather station does not need to trust Charlotte’s agent. This is because, as the selling agent, it would receive a receipt from the purchasing agent (Charlotte’s agent), allowing it to verify that the transaction is complete. The weather data would only be sent to Charlotte’s agent after the transaction was verified. However, although this method overcomes the issue of trust for the weather station, it still requires Charlotte’s agent to trust that the weather station agent will send the data it has promised. If it doesn’t, Charlotte’s agent will have been left shortchanged.

The solution is for agents to have a trust rating. This enables other agents to learn whether an agent has a good track record of honouring its obligations. Agents with a poor, or blank, rating will find it more difficult to conduct economic activity on the network. This is because agents, just like humans, will always seek to trade with parties they can trust. Agents that have a bad trust rating are, over time, identified and are expelled from the Fetch.ai network.

In this way, Charlotte’s agent can autonomously negotiate and trade with confidence with an agent representing a London weather station. It will make the trade, knowing the transaction will almost certainly be completed honestly and successfully to the benefit of both parties.

Get Involved

• Find out more about Fetch.ai’s Autonomous Economic Agents by visiting our developer website. Alternatively, read our documentation, which includes some innovative demos.
• Join our Developer Slack Channel and explore our growing GitHub repository.
• Watch videos outlining our technology and follow our tutorials on YouTube.
• If you’re not a developer, you can stay up to date with our progress by joining us on Telegram or by following us on Twitter.