Defend Your Network With A Bucket Full Of Jell-O

Screen Shot 2014-01-10 at 10.09.57 AM

Craig Shue and his colleagues at  Indiana University have tested and seriously propose the use of  just such a device for detecting denial of service attacks.

This bucket of Jell-o is a prototype 3D  analog computer and you might really consider using one to defend your network against cyber attacks. In the late 1980s and early 1990s I worked on the design of an analog and hybrid digital/analog neural computer based around the Intel ETANN 80170NX chip.  And I’ve had an interest in analog computing and unconventional computers ever since.

Well, they didn’t exactly use a bucket of Jell-o, but rather a “computational foam”. Mills et al, describe the device in more detail:

Conductive sheets and solids have been mentioned, and can be implemented easily. Any conducting or semiconducting material that can be formed into a sheet, layers of sheets, or a solid, and to which connections can be attached as ohmic point, line or area contacts will work. Silicon VLSI (“empty space”), conductive plastic foam, and gelatin “doped” with table salt have all been used in our machines.

Screen Shot 2014-01-10 at 10.26.31 AM

Foamputer 1995

Screen Shot 2014-01-10 at 10.10.43 AM

Bench Test EAC using Conductive Sheet

The latest versions of the device have progressed to a USB connectable device with the obligatory cool blinky things on top.

Screen Shot 2014-01-10 at 10.10.09 AMScreen Shot 2014-01-10 at 10.55.20 AM

But it is what’s inside that is perhaps a bit more interesting. The device consists of a conductive material and a digital interface that allows control of voltages applied to it and read out of the resulting voltages. Shown here is an implementation of Rubel’s extended analog computer or EAC designed by Bryce Himebaugh at Indiana University in 2005 connected to a digital computer host, either a PC or a Macintosh, which runs the EAC operating system (jEAC) designed by Ryan Varick also at Indiana University in 2006.

Screen Shot 2014-01-10 at 10.10.20 AM

The EAC is a machine that can “see” and “learn”.

The extended analog computer operates by analogy. Two systems are brought into congruence with each other. One system is the problem to be solved or, more generally, an application—a set of similar problems to be solved on a broad range of data sets—and the second is the EAC. This agreement between systems is first devised by visualizing it, then sending instructions to the EAC to configure it. 

The result is an entirely novel approach to use of computation such that students using the EAC prototypes are expected to extract characters from the so called “butterfly alphabet” as their initial project with the device. Consider writing the Java code to do this as a sort of “Hello World” application your first day using the language.

The first application of the EAC was studying butterfly wing pattern morphogenesis, and it was ” designed by translating a visual model of the butterfly wing to the conductive sheet”.

Screen Shot 2014-01-10 at 10.10.55 AM

Screen Shot 2014-01-10 at 11.01.56 AM

What emerges out of this idea is a type of machine where the programmer disappears and a user presents a problem without explicitly knowing how to solve it in advance. The machine sees and learns. This paradigm can be applied to other types of unconventional computers such as physarum or slime mold computing.

Screen Shot 2014-01-10 at 11.15.29 AM

And about that Jell-o, Mills et al report:

Sodium chloride (table salt) was added to the mixture before use. The resulting conductive gel had a 3×3×3 grid of electrodes on non-conductive plastic rods molded into it. Its operation was more accurate than the 2D foam. Several experiments were run with heterogeneous gels, adding plastic chips or short lengths of stripped wire. Their ability to model systems such as extraction of oil from shale is promising. Conductive plastics and semiconductors can be injection molded or formed in layers with electrodes embedded or sandwiched between them. Further research into 3D EAC design is supported by the success of this prototype.

Screen Shot 2014-01-10 at 11.16.49 AM

Screen Shot 2014-01-10 at 11.27.25 AM

 

 

References

Mills, Jonathan W., et al. “Empty space computes: The evolution of an unconventional supercomputer.” Proceedings of the 3rd conference on Computing frontiers. ACM, 2006.

Shue, Craig, Brian Kopecky, and Chris Weilemann. Denial of Service Attack Detection using Extended Analog Computers. IUCS Technical Report TR624, 2006.

Mills, Jonathan W., et al. “Extended analog computers: A unifying paradigm for VLSI, plastic and colloidal computing systems.” Workshop on Unique Chips and Systems (UCAS-1). Held in conjunction with IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS05), Austin, Texas. 2005.

L. A. Rubel, The Extended Analog Computer, Advances in Applied Mathematics,
14 pp. 39-50, (1993).

Kirchhoff-Lukasiewicz Machines, http://www.cs.indiana.edu/~jwmills/ANALOG.NOTEBOOK/klm/klm.html

Bournez, Olivier, and Manuel L. Campagnolo. “A survey on continuous time computations.” New Computational Paradigms. Springer New York, 2008. 383-423.

B. Himebaugh, Design of EAC R002, www.cs.indiana.edu/~bhimebau (2005).

R. Varick, Building tools for the analog researcher: an interaction design challenge, Senior Research Project Report, Indiana University (2006).

1 Comment

Leave a Reply