Information Technology Reference
In-Depth Information
the next backup. During any component switchovers, the respirocyte
computer must control the real-time distribution of power that is
transmitted hydraulically to local station subsystems and also along a
dozen independent interstation trunk lines that allow stations to pass
hydraulic power among themselves as required, permitting load shifting
and balancing. Redundancy management also applies to onboard com-
puters, which are themselves multiply redundant. Further analysis is
needed to determine the best techniques for safely switching among them
without sacrificing system reliability.
11. Flawless Compact Software. Another fundamental limit to onboard
computing is the minimum practical volume that may be occupied by
registers or other physical media capable of storing or executing onboard
data or instructions. For example, a 1 micron
3
storage volume could
10
10
bits (
contain
1GB assuming 8-bit words) on a hydrofluorocar-
bon memory tape with a maximum
B
B
10
7
bits
B
10 sec access time, or
B
(
1MB) in a tightly packed block of mechanical registers comprised of
nanoscale diamond rods with a maximum
B
1millisec access time [1af].
Given the presumed requirements for tenfold redundancy and the need
for non-memory computer components, maximum allowable onboard
data storage may often be an order of magnitude smaller than the above
figures, or less. Given such limited availability of onboard nanorobot
computer memory, it seems inevitable that efficient ultracompact soft-
ware and space-efficient algorithms will again come into vogue. Such
software was commonplace in the early days of PC development when
total RAM memories were often limited to 48KB or less—a meager
allocation that was nonetheless sufficient to hold an entire word proces-
sing program. High software reliability must also be a major design
criterion [63] because medical nanorobots must be extremely reliable.
People have already been killed or almost killed [64] by software bugs in
conventional medical devices. Policies for medical and product liability
insurance will almost certainly incorporate strict requirements for pro-
vable software reliability and system noncrashability as measured by
formalized industry metrics because of the potential significant risks to
human health if a nanorobot control malfunction due to software error
occurs during therapeutic use—a stringent requirement that should, at
long last, make the formalized production of ''error-free'' code economic-
ally viable. The related supervisory software must also be designed with a
simple user interface to minimize the possibility of error even when
operated by weary, distracted, and fallible human medical personnel [65].
B
15.4. NANOROBOT CONTROL PROTOCOLS
A nanorobot control protocol is a specialized sequence of nanorobot actions that
may be executed in carefully specified circumstances in order to ensure that the
Search WWH ::
Custom Search