Florian Weimer wrote:
You mean something like remote attestation?  I find it hard to believe
that this capability is available today in a relatively open
environment, on a platform supporting multiple applications developed
by different applications.

re:
http://www.garlic.com/~lynn/aadsm23.htm#49 Status of SRP
http://www.garlic.com/~lynn/aadsm23.htm#50 Status of SRP

i got involved in tracking down a virus/trojan like problem in the 70s on the internal network
http://www.garlic.com/~lynn/subnetwork.html#internalnet

basically if you are going to allow loading of stuff that can do its own execution w/o many safeguards ... you are going to be extremely vulnerable to numerous kinds of attacks.

either you have to very tightly control what applications are loaded .... or possibly do a fixed function deployment that can support multiple different applications ... possibly based on some form of data driven architecture (i.e. the data specification possibly adapts the functional operation to different applications w/o requiring loading of executable code).

we had done the AADS chip strawman was done this way ... basically single function operation w/o any ability to load executable code ... that was adaptable to a large number of different applications
http://www.garlic.com/~lynn/x959.html#aads

another possible solution is very strong partitioning of any loadable executable content that is allowed extremely limited/controlled capability.

in the 60s as an undergraduate, i had done a lot with extremely controlled partitioning ... which i learned much later got used in various environments that had extremely high integrity requirements ... random drift
http://www.nsa.gov/selinux/list-archive/0409/8362.cfm

i had this discussion with the general manager of the business unit that included java and java virtual machine (when it was in its very early infancy) ... turns out that I had done some work with the person (general manager) nearly 20 years earlier in a different life.

many of the modern generation of POS terminals are trying to cope with this problem ... getting all sorts of frequent application downloads of various kinds ... and still attempting to operate within constraints of their trusted security module implementation.

basically if finread
http://www.garlic.com/~lynn/subpubkey.html#finread

is countermeasure to widely acceptable PC vulnerabilities (many that arise because of the ease and common practice of loading executable content) ... then if you deploy such a finread terminal that is operated using similar conventions ... then it will acquire similar vulnerability characteristics (as the environment that it is suppose to be a countermeasure for).

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to [EMAIL PROTECTED]

Reply via email to