To Slow or Not? Challenges in Subsecond Networks. Johnson, N. F. 355(6327):801–802.
To Slow or Not? Challenges in Subsecond Networks [link]Paper  doi  abstract   bibtex   
[Excerpt] [...] today's electronic exchanges are an all-machine playing field with extreme subsecond operating times that lie far beyond the ∼1-s real-time response and intervention of any human. High-speed algorithms now receive, process, and respond to information on the scale of microseconds, and the only guaranteed future speed barrier is the speed of light. Hundreds of orders are executed across multiple exchange nodes within 1 ms (millisecond). [...] The need to develop a systems-level understanding concerning regulation in subsecond networks, is beginning to extend beyond the financial world. Two recent examples concern the navigational networks in driverless cars [...] and unmanned aerial vehicles (drones) [...]. Given that the navigational processing in their underlying networks of sensors and software operates much faster than human response times, what regulatory principles should be hardwired or encoded in vehicles in order to respond effectively to hardware or software defects? The core scientific challenge is to be able to predict the types of extreme behaviors that a given policy (e.g., an intentional time delay) will generate in a decentralized, subsecond network of decision-making machinery, in which each component receives continually updated information. [...]
@article{johnsonSlowNotChallenges2017,
  title = {To Slow or Not? {{Challenges}} in Subsecond Networks},
  author = {Johnson, Neil F.},
  date = {2017-02},
  journaltitle = {Science},
  volume = {355},
  pages = {801--802},
  issn = {0036-8075},
  doi = {10.1126/science.aai8618},
  url = {https://doi.org/10.1126/science.aai8618},
  abstract = {[Excerpt] [...] today's electronic exchanges are an all-machine playing field with extreme subsecond operating times that lie far beyond the ∼1-s real-time response and intervention of any human. High-speed algorithms now receive, process, and respond to information on the scale of microseconds, and the only guaranteed future speed barrier is the speed of light. Hundreds of orders are executed across multiple exchange nodes within 1 ms (millisecond). [...] The need to develop a systems-level understanding concerning regulation in subsecond networks, is beginning to extend beyond the financial world. Two recent examples concern the navigational networks in driverless cars [...] and unmanned aerial vehicles (drones) [...]. Given that the navigational processing in their underlying networks of sensors and software operates much faster than human response times, what regulatory principles should be hardwired or encoded in vehicles in order to respond effectively to hardware or software defects? The core scientific challenge is to be able to predict the types of extreme behaviors that a given policy (e.g., an intentional time delay) will generate in a decentralized, subsecond network of decision-making machinery, in which each component receives continually updated information. [...]},
  keywords = {*imported-from-citeulike-INRMM,~INRMM-MiD:c-14285349,digital-society,economics,economy-bias,ethics,feedback,high-frequency-feedback,technocracy,technology-mediated-communication,web-and-information-technologies},
  number = {6327}
}
Downloads: 0