I’m sure that those of you reading here are acquainted with Moore’s Law, and related observations: the cost per unit of computing power has been dropping rapidly for several decades. Just yesterday, I wrote about the retirement of the Roadrunner supercomputer system, the first system to break the petaflop performance barrier; that system has not, of course, started running slower; but its performance per unit of electricity consumed is no longer competitive.
Similar improvements are occurring in communications technology. (I have written before about the history of Ethernet technology, the nominal speed of which has increased from 10 Mbit/second in the early days to 100 Gbit/second,and up,today.) These increases are pretty impressive in their own right; if one thinks back to Internet access via modem dial-up, the mind boggles.
Thinking about these trends, one idea stands out: we have not, collectively, done an especially good job of anticipating them, or of making appropriate plans to adjust to these rapidly evolving technologies. Because of this, I am happy to join in, and promote, an effort, focused on communications technology, to be better prepared for future improvements.
The effective speed of data communications has been getting faster, not only because of increased network speeds (as I mentioned above for Ethernet), but because of better understanding of the underlying physical principles involved. This leads me to support a bold idea: it is time to pretpare for communication rates that exceed the speed of light. OK, maybe it won’t be on sale in time for Christmas of this year; but we have, collectively, been late so many times that being early might be a welcome change.
The key conceptual problem with faster-than-light communications is that, because of relativistic effects on time, the message may arrive before it is sent. (Relativity theory says that time slows down as one approaches the speed of light; it is plausible, and in accord with the equations, that time goes backwards once the speed of light is exceeded.) I’m glad to say that the society of Internet Protocol Cognoscienti (in which your humble servant plays a very minor part) has developed a draft standard [RFC 6921] for moving forward under this faster and more exciting regime. As stated in the abstract of the new standard:
We are approaching the time when we will be able to communicate faster than the speed of light. It is well known that as we approach the speed of light, time slows down. Logically, it is reasonable to assume that as we go faster than the speed of light, time will reverse. The major consequence of this for Internet protocols is that packets will arrive before they are sent. This will have a major impact on the way we design Internet protocols. This paper outlines some of the issues and suggests some directions for additional analysis of these issues.
It’s great to see this kind of proactive work by the standards bodies; in fact, I might suggest that you make note of the date.
Update, Tuesday, 2 April, 0:05 EDT
Please do take note of the date.