Please correct this article if you feel it is incorrectly representing the reality or the actual representation of what could have been called a merely sketch of one possibilities.Since I am french-native, one could possibly flames me all night for faulty english grammars, but please accept my apologies.
I was thinking holistic tracing kind-of biometric device. One could design such program whose job is to capture input from the biometric-recognition system(maybe). Such device would then implement holistic tracing; as in "it is possible to infer the existence of an attack if a set of observations (that are perhaps superficially unrelated) can be approximately matched to a structure that represents knowledge of the methods that attacks employ at a high(er) level" - sasha description of holism - because we could *do* system_calls tracing as the core holistic mecanism. Such tracing is the 'key' or "knowledge of the methods" to holistic approaches, in my humble opinion. Now I dont want to pretend how we could (or not) make use of biometrical technologies as you folks says, but let's pretend that tracing under biometrical circumstances (foo) might requisite some kind of transport/proxy whatever-else supporting layer. This transport model would then feature adaptative comodities, as requested. Comments ? It is unsure if tracing is the right thing to use for that application. Yet, tracing might not even be the best way to do it. See, systems_call tracing may be one way(*) of doing behind-the-borders (holistic) detection.I was thinking of posix compliances, too.
in a nut shell,...here's a early-moffo-sketch of what it's like:
[exterior]---data---->[biometric processing][custom micro kernel/mach service]...then we have:
- Adapted system_calls for the biometric device. - Intrusion detection system builtin (near realtime would be ideal but not realistic... imho.) - Holistic approache (see Phrack #57) seem good to me, seem good to me. - Could *it* have a use for first-aid/emergency response(??)(not to over-do-it, btw...) - Hmm, could this thing fly? Finally, this kind of 'holistic concept' applied to biometrics is rather unclear to my point of view.I suggest we push stronger the holistic debate, to the extend of knowing particulary what appliances these kind of technologies might become benefit for.. Is it really suited to exterior(read realtime) condition ? What operating system you think is best suited for this job ? Anyways, i was foreseeing this as a real challenge for the hurd. To some extend, it would be easy to setup some translators for biometrical-recognition use. Perhaps that the device would have to transport his "utter packet" to some daemon/switches/whatever ? Thinking of features, I was thinking of the common SSL Leay functions in order to 'secure' the designed protocols under some Hurd/settrans translators. (for info: http://www.8ung.at/shell/guide.html) That said, this is just my two cents and I dont expect comments or emails. I wish you all a good day and keep the good job, as in keep the sources open.
incidah
references/greets:
1. sasha,HOLISTIC APPROACHES TO ATTACK DETECTION Phrack # 57 www.phrack.org Volume 0x0b, Issue 0x39, Phile #0x0b of 0x12 http://www.phrack-dont-give-a-shit-about-dmc a.org/ show.php?p=57&a=11 2. The Hurd Hacking guide @ http://www.8ung.at/shell/guide.html 3. THC @ www.thehackerchoices.com (read the almost done very, very nice guide to system_calls hacking)
Slashdot Mirror
Please correct this article if you feel it is incorrectly representing the reality or the actual representation of what could have been called a merely sketch of one possibilities.Since I am french-native, one could possibly flames me all night for faulty english grammars, but please accept my apologies.
...here's a early-moffo-sketch of what it's like:
...then we have:
c a.org/ show.php?p=57&a=11
I was thinking holistic tracing kind-of biometric device. One could design such program whose job is
to capture input from the biometric-recognition system(maybe). Such device would then implement holistic tracing; as in "it is
possible to infer the existence of an attack if a set of observations (that
are perhaps superficially unrelated) can be approximately matched to a
structure that represents knowledge of the methods that attacks employ at a
high(er) level" - sasha description of holism - because we could *do* system_calls tracing as the core holistic mecanism. Such tracing is the 'key' or "knowledge of the methods" to holistic approaches, in my humble opinion. Now I dont want to pretend how we could (or not) make use of biometrical technologies as you folks says, but let's pretend that tracing under biometrical circumstances (foo) might requisite some kind of transport/proxy whatever-else supporting layer. This transport model would then feature adaptative comodities, as requested. Comments ? It is unsure if tracing is the right thing to use for that application. Yet, tracing might not even be the best way to do it. See, systems_call tracing may be one way(*) of doing behind-the-borders (holistic) detection.I was thinking of posix compliances, too.
in a nut shell,
[exterior]---data---->[biometric processing][custom micro kernel/mach service]
- Adapted system_calls for the biometric device.
- Intrusion detection system builtin (near realtime would be ideal but not realistic... imho.)
- Holistic approache (see Phrack #57) seem good to me, seem good to me.
- Could *it* have a use for first-aid/emergency response(??)(not to over-do-it, btw...)
- Hmm, could this thing fly?
Finally, this kind of 'holistic concept' applied to biometrics is rather unclear to my point of view.I suggest we push stronger the holistic debate, to the extend of knowing particulary what appliances these kind of technologies might become benefit for.. Is it really suited to exterior(read realtime) condition ? What operating system you think is best suited for this job ? Anyways, i was foreseeing this as a real challenge for the hurd. To some extend, it would be easy to setup some translators for biometrical-recognition use. Perhaps that the device would have to transport his "utter packet" to some daemon/switches/whatever ? Thinking of features, I was thinking of the common SSL Leay functions in order to 'secure' the designed protocols under some Hurd/settrans translators. (for info: http://www.8ung.at/shell/guide.html)
That said, this is just my two cents and I dont expect comments or emails.
I wish you all a good day and keep the good job, as in keep the sources open.
incidah
references/greets:
1. sasha,HOLISTIC APPROACHES TO ATTACK DETECTION Phrack # 57 www.phrack.org Volume 0x0b, Issue 0x39, Phile #0x0b of 0x12
http://www.phrack-dont-give-a-shit-about-dm
2. The Hurd Hacking guide @ http://www.8ung.at/shell/guide.html
3. THC @ www.thehackerchoices.com (read the almost done very, very nice guide to system_calls hacking)