Exercise 16: Keys.

[camwyn]

Today's installment of the Adventures of Rocket-Boy takes place several years after the events of his phone conversation with Emile Jalbert. He's attending the college of his choice in Montreal. And he's being subjected to a writing exercise wherein the goal was to describe a set of keys, then describe what they opened.

This piece is exceptionally geeky, even for me, and may contain more than its fair share of errors of fact. I will happily accept corrections from those who read it who actually know what they are talking about.



Keys

"That is, beyond a doubt, the single worst cipher I have ever seen. What in the name of Hell were you thinking?"

Ira's shoulders hunched instinctively, despite the fact that Professor Rabesahala's invective wasn't aimed at him. There was something peculiarly penetrating about the man's voice, above and beyond the inflections stamped into every word by growing up in Madagascar. Rabesahala could reduce a careless student to tears with a single well-chosen sentence, and had done so on more than one occasion.

This was not one of those occasions. Gisèle, Ira's classmate, fixed the dark-skinned professor with a weary look. "I was thinking that it's adequate to the purposes you set us," she drawled. "You wanted a cipher that demonstrated an understanding of the basic principles of symmetric key cryptography. That's what you got."

"I have seen seven year olds who can code a better cipher algorithm than this, Miss Savard."

Gisèle sighed. "This isn't my only class, Professor Rabesahala. I do have other assignments. And you said it yourself, this is a basic assignment. It's not like it's a term paper or something."

"That is no excuse for shoddy work! Honestly, Miss Savard, I am surprised at you. . ."

Gisèle could not possibly have said anything worse. Rabesahala saw no difference between the first assignment of the year and the last, everyone knew that. Why the fair-haired girl thought she could get away with an excuse like that he'd never know. Ira shook his head and turned his attention back to his own terminal.

"And you, Mr. Dayan. What is this?" Rabesahala's voice lashed out in Ira's direction.

"Excuse me?" Ira said mildly, lifting his head but not bothering to move his hands from the keyboard.

"I asked you for a basic cipher algorithm, if you will remember. Not for a randomly shifting time-sensitive subkey generator dependent on an external server for production of multiple keys during a single communications session." Rabesahala waved a steel grey data solid at Ira. "Particularly not one that appears to be the bastard child of Merkle-Hellman and IDEA."

Ira grinned. "Sorry, Professor," he said. "Standard key generators are boring."

"Yes, but they are simple! This is not! This is - this is demented. Merkle-Hellman was broken forty years ago. It is not even meant for symmetric key generation."

"I know, but the mathematics are the most elegant I've ever seen-"

"It is a public key algorithm, Mr. Dayan. Public key. The amount of contortion your code has to go through to produce a lone usable key-" Rabesahala shook his head. "I admit, the constant shifting of subkeys is an ingenious touch-"

"Thank you."

Rabesahala ignored him. "But it requires the use of a separate subkey transmission system, without which the rest of your code is useless. This is a hybrid system that requires an entirely separate piece of encryption in order to work. Secure it may be, but fast it is not. . . What did you intend to use this for, anyway?"

"Academics, mostly." Ira indicated the terminal next to him. "Say Professor Mansuy, over in Bio, is looking to do real-time work on left-hand xenoproteins on the analytical servers in the Institut Curie. She's had her work stolen before, she's after a Nobel, she wants something that'll keep the data under wraps and give the raiders a headache. Both the Ecole and the Institut have the kind of computational power you'd need to run this kind of encryption, and the kind of budget you'd need to handle the extra machinery. It's not really suitable for government work - I mean, it's not like they couldn't afford it, it's just it doesn't live up to the AES standard."

"No. It does not." Rabesahala wrinkled his nose and set the data solid in front of Ira with a click. "And it is not supposed to. Before you can run, Mr. Dayan, you are supposed to walk. While you clearly have an idea of secret key principles, this work of yours is almost entirely flash and dazzle. Please, rework it to reflect what I have assigned, and bring that to me in a week's time. You admire Merkle-Hellman's mathematical simplicity?"

Ira nodded.

"Then structure your math so that I can see that. What you have given me - I am surprised that your computer did not throw up from the migraine you must have given it. This is a mess."

"Yeah, but it works!"

"Yes, but it is a mess," Rabesahala returned. "I did not ask for a mess! Bring me what I asked for and then we will talk."

With a grimace of resignation, Ira nodded. "How long do you want the key to be?" he asked.

"Short," said the professor crisply. "And one per session please, I do not want you overreaching yourself so early."

"All right."

Rabesahala nodded and went on to the next student. Ira sighed, twiddling the steel grey cylinder in his fingers. He didn't like the class much, but he'd needed an information science elective, and the remote operated vehicle systems the Astro-Engineers were working with all relied on secure data transmission principles. You couldn't do better than Rabesahala when it came to that kind of thing. It was just that Ira's theoretical maths were - well - horrible. Applied mathematics he could handle - physics equations, for example. There was an effect in the real world to hang the numbers on, and as long as he kept the effects in mind, he had no problems whatsoever making those numbers dance. Pure maths were another story, and code generation was dangerously close to pure maths for his liking. Without a tangible end result to keep in mind, he moved through complex equations like a bull through wet swampland. That was why he'd started from the knapsack problem. Merkle-Hellman was simple.

He set the data solid down and glowered at nothing in particular. Dammit, he'd have to start over from scratch.

Comments

[dormouse-in-tea.livejournal.com]

*giggles* A bull through wet swampland!

I like the piece, and I like the way you approached the writing exercise.

[camwyn]

Hey, a key's a key. I was thinking cryptographic keys almost from the instant I saw the exercise (the alternative was a batch of antiquated maps, or something like the Rosetta Stone).