DOC PREVIEW
Berkeley COMPSCI 161 - CS 161 Midterm 1 Review

This preview shows page 1-2-3-4-5-6 out of 17 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 17 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Shamir Secret SharingZero-Knowledge ProtocolsAuthenticationFirewallsCS 161: Computer SecurityMidterm 1 ReviewPart 2October 4, 2006Marco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 1/17Shamir Secret SharingMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 2/17Sharing a secret◮Let’s say we want to hide a secret with t people such thatany q of them can reconstruct the secret but any q − 1cannot◮We can use a polynomial with random coefficients:f(x) = xq+ aq−1xq−1+ · · · + a1x + a0(modp)◮The secret is f(0) = a0, and the shares we distribute aref(1), f(2), ..., f(t) (the other coefficients are also keptsecret)◮Now any q people can solve the polynomial for a0but q − 1have no information◮Note: Could do something similar with real numbers, butintegers are easier so we use modular arithmeticMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 3/17Secret sharing: simple example◮Let q = 3 and p = 11 (don’t confuse with the p, q of RSA):f(x) = x3+ a2x2+ a1x + a0(mod11)◮Let’s say the secret we want to hide is “5”, and werandomly choose coefficients a2= 3 and a1= 9◮For t = 6, we compute f(1) = 7, f(2) = 10, f(3) = 9,f(4) = 10, f(5) = 8, and f(6) = 9.◮Any three can now solve for coefficients:f(2) = 10 = 8 + 4a2+ 2a1+ a0(mod11)f(3) = 9 = 5 + 9a2+ 3a1+ a0(mod11)f(6) = 9 = 7 + 3a2+ 6a1+ a0(mod11)Marco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 4/17Zero-Knowledge ProtocolsMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 5/17Zero-knowledge proof of identity◮Goal 1: Alice knows that the person she’s talking to is Bob.◮Goal 2: Bob reveals no additional information to Alice.◮Assumptions:◮Alice knows Bob’s public key◮Taking square roots modulo n is "hard"◮Protocol (everything mod pq):◮Bob picks secret b, publishes b2as public key (persistent)◮Alice wants to check Bob’s identity, asks Bob to begin◮Bob picks random r (new each run), sends "commitment"r2to Alice◮Alice flips coin (or chooses); heads means reveal r (Aliceverifies r2), tails means reveal rb (Alice verifies r2b2)◮If Bob passes, Alice is 50% convinced of his identity◮Repeat arbitrarily many timesMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 6/17How can Mallory pretend to be Bob?◮Doesn’t know b (because he can’t factor), but can gamethe system◮To know the "heads" answer, picks r like Bob would◮When Alice sends "heads", he sends r◮If Alice sends "tails", what can he do? nothing!◮To know the "tails" answer, picks t and computesr2= t2/b2to send to Alice◮When Alice sends "tails", he sends t◮Alice thinks he sent rb and checks by squaring it, but(rb)2= r2b2= (t2/b2)b2= t2so fake Bob has fooled Aliceon tails◮If Alice sends "heads", he can’t do anything because hedidn’t pick r and then square it, he just picked somethingthat he called r2, but he can’t take the square rootMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 7/17Why doesn’t the protocol leak information to Alice?◮When she flips "heads", Bob just reveals a randomnumber, which Alice could have picked by herself◮When she flips "tails", Bob sends her rb, but this is alsorandom because multiplication induces a permutation, so arandom number times anything lands on a random valueMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 8/17AuthenticationMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 9/17Authentication◮Authentication is verifying an identity, or verifying theoriginator of a message◮Many types of authentication◮Person → person◮Person → local computer◮Remote computer → person◮etc.◮Difficult to get right and easy to screw up◮Most real attacks today are authentication attacks:phishing, “pretexting”, spyware password pop-ups, etc.Marco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 10/17Needham-Schroeder◮Symmetric encryption with trusted server◮Each user shares symmetric key with server (A shares keya, B shares key b, etc.)◮A → S: {B}a◮S → A: {t, {A, t}b}a◮A → B: {A, t}b◮A ↔ B: { messages...}tMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 11/17Problem and fix◮Replay attack:◮M → B: {A, t}b◮M → B: { something that shouldn’t be repeated }t◮Solution: nonces (unique value, such as a random numberor timestamp)◮Revised N-S: every message has timestamp, so attackercan’t replay◮Problems remain: requires real-time trusted third partyMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 12/17FirewallsMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 13/17Firewall overview◮Motivation: Every network service is a potential hole◮Block services in the network before they reach machines◮Enforces security policy: policy on which services shouldbe visible, which should be blocked, and how wedistinguish insiders from outsiders◮Default allow vs. default deny◮Default allow is easier on users and bothers them less◮Default deny is more secure in several ways: fails safe,catches unknown attacks, hedges against commonmistakesMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 14/17Packet filters◮Checks each packet against series of rules◮Rules test IP, protocol, port, etc. to decide drop or allow◮The first matching rule decides the action◮Syntax: hactioni hprotoi haddri → haddri◮Each haddri is of the form hipi:hporti[/{in,out}]◮The * wildcard matches any value◮(in/out difference is which interface packet received on)Marco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 15/17Firewall example◮allow tcp*:*/out → 1.2.3.4:25/in◮allow tcp*:*/in →*:*/out◮allow tcp*:*/out →*:*/in (if ACK is set)◮drop* *:*→*:*Marco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2 16/17More on firewalls◮Reference monitor◮Mediates all access to network◮Three requirements: Always invoked, tamper resistant,verifiable◮Application-level firewall can do more than packet filter◮Inspect and enforce application protocols◮Do more nuanced filtering◮Stateful firewall can do more◮Assemble TCP connections (not just look for ACK)◮Limit number of open requests, etc.◮VPNs extend perimeter over secure channel to remotemachine◮Good for working from home, bad if home computer getsvirusMarco Barreno CS 161, Fall 2006 Midterm 1 Review, Part 2


View Full Document

Berkeley COMPSCI 161 - CS 161 Midterm 1 Review

Documents in this Course
Rootkits

Rootkits

11 pages

Load more
Download CS 161 Midterm 1 Review
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view CS 161 Midterm 1 Review and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view CS 161 Midterm 1 Review 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?