Attack-resistant decentralized time and location services via Nakamoto chain consensus. Time is one of the most precisely measurable units, but affordable time keeping today depends on trusting central time services. These central time services are able to perform attacks by manipulating the apparent time to their advantage. This text describes a completely decentralized solution to this problem. Today the most accurate affordable time solution is a GPS disciplined oscillator. A GPSDO combines a stable crystal oscillator (typical an Oven-Controlled Crystal Oscillator) and a GPS receiver. An OXCO provides excellent short term stability (usually better than 1e-12 on the one second time scale), substantially better than atomic clocks (esp inexpensive rubidium clocks), but because they are relative frequency sources rather than absolute they do not have good long term stability. In addition to being less expensive initially than atomic clocks crystal oscillators have also have much longer working lives. GPS is a centrally controlled time and location service operated by the US government. GPS itself receives absolute timing from large atomic clock ensembles operated by (government controlled) standards bodies which are relayed to the orbiting system via ground control stations. By holding location constant and solving for time the GPS recievers are able to deliver precise time measurements. However, orbital uncertainty, limited stability of the orbiting atomic clocks, location and time specific differences RF index of refraction of the ionosphere and atmosphere, and intentional degradation of the system by its operator limit the short term accuracy of timing provided by GPS. A GPSDO operates a control loop where the observed phase error between the GPS time and the local time is used to drive the local oscillator towards the correct frequency (typically by adjusting the oven temperature in an OXCO). When properly designed a GPSDO system provides equal or more accurate timekeeping to a single atomic clock at all time scales in a small, reliable, system which only costs a few hundred dollars. Unfortunately, this solution is completely at the mercy of the operators of the GPS system— they can lie about the time and existing systems will believe the lies. Moreover, the US Government explicitly states that they will degrade or deactivate the GPS system if doing so is required to achieve military objectives, and the system has features built in for this purpose and these attacks can be geographically targeted. Current GPSDOs will gladly believe even the most outrageous lies from the GPS system— if, after months of reporting the year as 2012, the GPS system claims that the date is 2009 the GPSDOs will happily report it as such. Although the devices could be improved by sanity checking the received signal, this is not really a solution since a rejected attack signal would no longer provide a long term reference and the clocks would drift out of sync. (and it would potentially open new attacks— e.g. a short term jamming signal claims the gps time is 2014 and then cynical GPSDOs refuse to go backwards to the correct time) However, we can use Nakamoto chain consensus to decentralized these systems. A participating set of nodes maintain accurate adjustable local frequency references. They measure the offset between their local time and a reference oscillator available to all nodes. After making a measurement they write a statement about this offset and attempt to solve a POW to extend a Nakamoto chain. New blocks in this chain are only accepted by a node if their offsets are plausible. Nodes then perform a median operation of recent offsets to produce a consensus offset between against the reference, and then use that offset to drive the frequency of their local clocks into agreement with the consensus. In effect, we have a GPSDO-like system but with a consensus correction being applied. In the case where the reference _is_ the GPS system the above is still highly vulnerable to denial of service by central authorities (though not vulnerable to simple offsets). However, we can instead us another globally available reference signal which is strongly attack resistant: The sun. Participating nodes would sample RF noise on some agreed band(s) being emitted by the Sun and continually record it, with their sampling clock being driven by their stable local oscillator. Nodes would then publish timestamped recent fragments of this signal. Peers would then perform a cross-correlation between the fragments and their own timestamped recordings in order to find the offset. Only nodes which can concurrently observe the sun can actively participate in the agreement, but because the local oscillators are relatively stable this should not prevent good long term stability. (Effectively your clock would be corrected in the daytime and free run at night). Such a system could also timestamp multiple reference clocks concurrently so that other reference signals could be used when available. (crazy stuff starts here) The above system could be extended to also provide location service. Assuming the participating (mining) nodes know their own locations, and that their receivers are sensitive enough to receive both RF noise from the sun as well as the reflections of this noise off multiple astronomical objects in the solar system then a node could estimate the distance between it and these various bodies using the cross-correlation with the direct signal from the sun, essentially the same mechanism used by passive radar. Using these measurements, the node can solve an orbital dynamics model for the current model-time then publish the model-time (referenced to solar signal fragments) in the chain and their blocks will only be accepted by nodes which find the claim plausible. A node which does not know its location but which can make the same radio signal observations and can observe the hash chain can accept the chain consensus time and determine its position. (this part is crazy, sadly because the reflected signals would almost certainly be too weak, and the measured distances too inaccurate to make the technique useful)