Decentralized Computing
The world would own and maintaining the open climate platform—as a concept this enormous decentralized network of participants and data sources poses an interesting problem in distributed computing.

Population Protocols

Population protocols, introduced by Angluin et al., are algorithms that simulate computation among devices with limited communication range and storage capacity. In order to model this type of system, agents are selected to interact in pairs chosen randomly.
This model is useful for analyzing ad hoc networks of low grade devices, like IoT devices, blockchain nodes, sensors, and more.
For a brief overview of this type of system, read this chapter on population protocols.

Literature Overview

Much of the literature in this area of research deals with optimizing time complexity under various space complexity constraints. There are numerous papers that develop algorithms for performing leader election, population counting, majority input, and more. This research can be used in this project to determine the optimal way to aggregate information from multiple sources, including both mobile and stationary low-grade devices.

Open Problems: CP3

Though this research area has a lot of overlap with this project, there are many problems that have not yet been solved in this area that make it difficult to apply existing research results directly.
  • There is no existing verification mechanism that enables nodes to validate information obtained from other nodes.
  • There is limited knowledge regarding privacy preserving algorithms in this model. This is especially important for private companies that would like to participate in the OpenClimate project and contribute their aggregate metrics while protecting their more sensitive data.
  • Most of the research in this area is based on "fairness assumptions" about the interaction rate of the population. For example, weak fairness says that every pair of agents in the system interacts infinitely often, whereas global fairness dictates that each pair of agents is selected to interact at each time step with uniform probability. Neither of these notions necessarily describes the devices used in the OpenClimate project, so it remains unclear what fairness assumption we can make and how they affect the latency of data aggregation in this system.
This prompt asks you to contemplate the open problems outlined above, and add your own in this page! For more info contact Talley.
Last modified 1yr ago