Re: I-D ACTION:draft-bvenkat-chips-on-avians-00.txt

[Balaji Venkat <bvenkat at cisco.com>]

Hi robert,

I really loved the critique on this draft. Can I use these comments in the 
draft for the second version ? Would you be willing to co-author this ?

Please let me know.

thanks,
balaji venkat
At 02:19 PM 12/09/1999 -0600, Robert G. Ferrell wrote:
>>The connection topology can be non point-to-point for each carrier
>>and as specified in RFC 1149 [1] can be used without significant
>>interference with one another, outside of early spring. 
>   
>Carrier species and circannual timing will be critical issues here.  
>For example, while members of the columbiformes traditionally employed in   
>RFC 1149-type ASCII message vectoring might experience little if any routing 
>variation throughout the year, an unfortunate selection of, say, _Sterna 
>paradisaea_  in either late winter or early fall might result in an
unacceptable 
>latency period due to suboptimal routing.  A standard avian carrier needs
>to be developed, perhaps by genetic engineering, which will have minimal 
>reactions to seasonal variations in the local diurnal cycle.
>   
>>The bio-medically engineered chip allows low frequency signals
>>to be transmitted by these specially equipped avians that helps
>>signals move around large objects such as skyscrapers 
>
>Extensive experience in the field as a graduate student led me to the
conclusion
>that this collison avoidance mechanism was highly unreliable during periods 
>of inclement weather, most notably thunderstorms.  Avian/fixed structure 
>contact, particularly when such structures exhibited large expanses of 
>transparent/translucent glass, were common and almost invariably resulted in 
>loss of the carrier and thus of the transported data.  Any probability of 
>success index should take into account the weather conditions at the time of 
>transmission and the architectural topology of the anticipated route.
>
>>In addition the chip also allows for high frequency signals to be
transmitted 
>>that are inaudible to the human ear.
>
>Care must be taken that the signals are not inadvertently made audible en
route, 
>lest the carrier attract extraneous and detrimental predator attention.
>
>>IP traffic funnelled through after such negotiation can be 
>>connection oriented as in TCP or unreliable transport as in UDP. 
>
>Connection oriented traffic is inadvisable, due to the latency of 
>transmission.  The carrier might wander away or be distracted during signal 
>exchange, with an attendant loss of data integrity.  
>
>>The issues to be discussed include addressing for each such 
>>avian prior to the negotiation, after the negotiation and for
>>each low altitude IP tower
>
>It should be noted that these towers must be constructed of a material
that is
>highly resistant to the corrosive effects of uric acid deposits, a
significant 
>by-product of carrier physiology.
>
>>The low delay is achieved by the high data content in the fast
>>moving tweets and chirps, the variations of which are unheard of
>>in the human hearable frequencies. Thus these tweets and chirps
>>may be unheard by the normal human ear except for the upper range
>>of lower frequency chirps that provide for high delay and low
>>throughput for traffic of the kind that requires delivery but not
>>instant delivery.  
>
>One factor to consider here is that a great many of the most suitable
carrier 
>species produce highly stylized but imperfectly predictable signals that
must be 
>filtered or suppressed in order to achieve an acceptable signal-to-noise
ratio, 
>given that message transmission is acoustically achieved.  One might obviate 
>this by use of carriers with limited signal production in the frequency
range in 
>question, such as some members of the pelecaniformes or struthioniformes,
but 
>substitution of these species introduces a entirely different set of
challenges 
>(in the latter case, for example, lack of aerial locomotion is problematic).
>
>>The layer 2 addressing is done by allocating a MAC address to
>>every chip that is set on board an avian's brain. Appropriate
>>surgical techniques may be used to implant the chip with 
>>connections to its auditory and vocal mechanisms. 
>
>Since research has shown that many of the more desirable messaging carrier
>species possess small quantities of magnetite in their cranial cavities, the 
>possible deleterious interactions of this material with the implant must be 
>explored more fully before the reliability of either data integrity or
signal 
>routing can reasonably be assured.
>
>>For this reason the avians are tagged to be released in areas exclusive
of the 
>>other's if they happen to have the same send/receive frequency. 
>
>A certain cross-channeling effect is inevitable, given that routing cannot
be 
>controlled once the carrier has left the station of origin.  Given this
fact, 
>one must be prepared to accept that collisions will occur (see also my
comments 
>on structure avoidance, above).
>
>>Such collisions would require drastic action such as 
>>shooting down the colliding avian that has contravened its 
>>avian arena boundaries. 
>
>In certain circumstances this would decrease the effective end-to-end
bandwidth 
>to the point that sneaker net might be a more effective means of
communication.
>
>>It is an intrinsic advantage of this design that the MAC 
>>address (the prefix at least) can be learnt from the 
>>frequency of the avian chip. The OUI portion of the MAC
>>address can be shorter than the standard 24 bits. 
>
>Of course, the potential for unique identifier increases geometrically if
one 
>considers that each avian carrier possesses a unique roughly 30 bit DNA 
>sequence.
> 
>>Avian arena changes can be negotiated through the mobility
>>of an avian into another avian's arena. Thus two avians on
>>the same frequency may arrange to swap one another or 
>>arrange to rearrange the distribution of same frequency 
>>avians through a protocol. This subject too is left for 
>>further enquiry.
>
>Uncontrolled avian-to-avian interactions of this type tend to be
sufficiently 
>traumatic to one or both of the carriers that data integrity would most
likely 
>be compromised in such situations.
>
>>A single chipped avian serves as a repeater.
> 
>I would suggest the Chipping Sparrow (_Spizella passerina_) as an excellent 
>candidate for this position.
>
>>With regard to the degradation of
>>its chirp and tweet beak and vocal cords, transmission of frames
>>may be found to be degrading thus leading the avian to be put to
>>its terminal end of service by removing the chip from its brain. 
>
>Packet, not to mention carrier, capture is problematic as the carrier
approaches 
>its TTL value.  A frequency of use and reliability of transmission
expectation 
>that varies inversely with carrier age is advisable.
>   
>>If a collision occurs then both avian carriers back 
>>off as per the CSMA/CD mechanism outlined in IEEE 802.3 standards. 
>
>Only if the carriers are being closely monitored and physically modulated
by a 
>human agent with collision avoidance and detection rules well in mind.  When 
>collisions occur among avian carriers, the general rule of thumb is to
expect at 
>a minimum a considerable increase in latency and in worse cases a complete
loss 
>of data and carrier.
>
>>Chip manufacturers provide appropriate interfaces to tap into a dead
avian or a 
>>live one to transfer data back and forth from an avian chip to the
>>said device which may be a router, that is tangibly visible as one
>>to humans.
>
>I might add that, while they do not exhibit favorable transmission 
>characteristics for any messaging other than campus-wide (and even then
usually 
>line-of-sight, with a strong throwing arm), dead avians are remarkably 
>predictable in their behavior and are less apt to be lost due to routing 
>anomalies.
>
>>The loss of a carrier in an arena can result in the stoppage of
>>traffic in that arena onto the adjacent one. This is taken care by
>>providing a backup avian carrier since avians usually travel in pairs.
>
>This is highly species-specific. Fault tolerance that relies on this
principle 
>narrows the field of prospective carrier species to those which form strong 
>pair-bonds, and further renders reliable signal transmission outside of the 
>breeding season an iffy proposition at best.
>
>>As discussed earlier security is not a problem except in the
>>cross avian arena border transition case, which might take place
>>if an avian finds a courtship to be undertaken with another
>>avian in a different avian domain. This is sought to be 
>>restricted by injecting suitable mitigating agents that 
>>suppress the enzymes responsible for such courtship in a given
>>avian carrier.
>
>Those same 'enzymes' (actually hormones) are also responsible for
vocalizations. 
>Suppressing them would render the payload inaccessible, at least until the 
>suppressive effect subsided.  This would introduce considerable latency, and 
>repeated or improperly conducted suppressions might reduce the TTL of the 
>carrier significantly. 
>
>RGF
>
>Robert G. Ferrell
>Internet Technologist (with otherwise useless MS in Avian Ecology and 
>Systematics)
>National Business Center, US DoI
>Robert_G_Ferrell at nbc.gov
>