Every letter, despite being easily modifiable electronic, must be unalterable honest trustworthy undeniable non-spyable perfect forward secrecy, and dated signatured such that can not be denied. To get all the properties at once, despite quantum computers, in always detected proven violations, efficiently which means compared to competing encryption has been done and is the basis of my company.
Unalterable?
Unalterable is done by appending a digest that can be proven to generate by any modifier. For this purpose, a reader can attach any small figure of choosing to any inward mail and is attached to any mail that fails the expected constraints. Because the image is user chosen. There is no reasonable method to forge the picture no matter how good the criminal is! This I call perfect presentation of testing of the mail.
Honest?
NIST digests exist that change at random, half the bits in check bits, on random bit change in the message. Such digests can have embedded check bits that certify some part. The Sender will be able to change these sub-signatures only with probability allowed by the sub-signatures. This is crucial to detection of perfidy of the sender! There is no additional reason to doubt information contained in the letter signed separately. No further communication is needed beyond the public check functions of the signer and are done silently, denying the entire letter with even one failure. Whether a letter will fail in the reader can be accurately tested by the sender! Any unsigned claim in the sender should not be believed by the reader! To make it more transparent, parts of the letter may be enclosed in < and > and each such inclusion expects a signature. Exact syntax allowed in the angles depend on the signer. For example, I might say <M. Tech, computer science, IIT Kanpur, 77> and <department topper, IIT Kanpur 75> and both these can be signed once by IIT Kanpur and signatures saved with me for use in any letter or certificate. This is a disruptive change to usual email, improving it to certificate level! My email changes all with reliable job applications to sold services! No extra communication is needed!
Unspyable?
AES-384 is used. Its decryption can only be speeded by square-root using Quantum computers.
Trustworthy?
To ensure correctness of dates, what is done is the intersection of two half open time intervals of after that yield only one common date. One interval is some random number published by me (which could not be picked before) with me, or some story in the newspaper. Another is registration of the letter with me (Not the letter, My-SHA-512 on email is enough). Two, My-SHA-512 and My random are enough for date.
Undeniable?
How else was the signature of the entire letter computed?
Local trust?
Instead of digest, what is appended is the AES-384 encryption of it in the reverse encryption of the user. Encryption of the signature yields the digest and then the signature is verified, the process need not be exact inverse needs to generate the code of digest on encryption and be as hard to invert. It can be done using a different system!
Perfect forward secrecy?
Perfect forward secrecy means even if every thing about the system becomes known, every past message is still undecidable. It looks like a tall order but is very easy. Essentially every message is sent by a new system, forgotten every round. That is done by a separate system that generates random integers sent with every message to be used by the receiver for reading, sender for writing for the next message! The random number can be read off a quantum system! No way can any breach expose the past, not saved or creatable.
Freely develop able but still controlled?
If one understands Jules Verne and how his characters reached the moon, once the basic trick is mastered (anti-gravity), all his magic invention become scientific! Alas, there is no anti-gravity, even in anti-matter! I say because a lot of my developments will be controlled by my company even if the developments are done by other people. That is because square root encryption using huge integers is an unsolved problem for humans but doable by me, and I am no longer an academic. As I imagine, other developers will be forced to use the api of my company. Like all good NP-complete problems, finding the root is hard, but checking is trivial - simply square any answer.
No comments:
Post a Comment