Extended Abstract for Server Sky

• proposed for ISDC 2013

Sorry I did not have this ready earlier - I presented Server Sky as a poster at the Space Horizons Conference at Brown University on February 21, and again today (Mon Feb 25) to some friends on the faculty at MIT, and I wanted to include feedback from both.

Here's the poster covering most of what I will say, about server sky at one session and about space debris location in a separate talk. Please refer to that poster, and imagine that sections 9 and 10 will form a separate talk on space debris location, with much expanded information on converting derelict upper stages into valuable ballast mass. My screen talks are almost entirely visual, heavy on animations, with very little boring text on screen.

I have since done more calculations on Belleville spring deployment. Since glass is very stiff and fractures easily, I will not be able to store much energy in warped thinsats; just enough to move the thinsats at a few centimeters per second. However, by slowly tumbling the deployment stage (perhaps 2 rpm), this will dispense thinsats with enough radial velocity to separate them into an expanding cloud. I will have animations ready by the time of the conference.

At the Brown conference, I had some good conversations with Mike Hecht of MIT's Haystack Obervatory, whose 37 meter dish is used by Lincoln Labs as part of the space debris tracking network. I am in the MIT library right now researching phased array radar, and will update my presentation with what I learn tonight.

Some feedback from my faculty friends:

A useful application of launchloop will be providing bandwidth for eGovernment in rural India. India wants to provide 2MBps internet to hundreds of thousands of rural villages - this will be difficult to do with traditional comsats or with land lines and microwave links, but relatively easy with server sky. Server sky can also perform translations between text and speech to provide "audio internet" to the cell phones of not-yet-literate rural farmers. I want to talk about applications like this that can pay for distributed high bandwidth space systems, even if only pennies per day per user are available.

I also talked with a professor about efficient dataflow hardware for finding correlations between radar returns and templated patterns. His work is about leaving data in RAM tightly bound to processors in array, then shifting groups of instructions past those processors. For small-code, big data applications this is more efficient. Server Sky can do correlations in a similar way, creating large arrays of expected correlations and shifting returns past them, triggering further analysis for the few returns that exceed a threshold.

Over all, the idea of server sky is that space can be made into an electronics/digital problem, where solutions will expand at a Moore's Law rate.

After I get back to Oregon on Wednesday, I will expand and separate these abstracts.