The index function problem is defined as follows. Alice has a string x distributed uniformly over {0, 1}n. Bob has an index i distributed uniformly over [n]. The goal for Bob is to guess xi when only one round of communication is allowed, i.e., Alice can send only one message to Bob. The naive protocol for this problem is that Alice sends all her bits to Bob. Then the message length is n. Our g...

As with the tail bounds we saw in the last lecture, Chernoff’s formula is used to bound the probability that a random variable deviates from its mean. For this bound, however, the random variable under consideration will have a very specific structure. Let X1, . . . , Xn be independent {0, 1}-valued random variables, and define pi = Pr[Xi = 1] = E[Xi]. Then, X = ∑ iXi will be the random variabl...

Last lecture, we discussed Janus, which uses system call interposition to achieve software sandboxing isolation. We talked about a few problems that Janus had, particularly race conditions that can potentially allow malicious code to break software isolation property. Since then, a few more sandboxing projects have been implemented to address such issues. Today, we are going to look at a few re...

Ursula A. Kelly. The Foresters’ Scribe: Remembering the Newfoundland Forestry Companies through First World War Letters of Regimental Quartermaster Sergeant John Barrett. St. John’s: ISER Books, 2020. ISBN 978-1-89472-573-6. An article from journal and Labrador Studies (Volume 36, Number 2, 2021, pp. 171-344), on Érudit.

The main objective of this lecture is to prove a lower bound on the number of registers (variables) required for any distributed algorithm that solves the Mutual Exclusion problem. We will work in an asynchronous shared memory model. Recall that Peterson’s Algorithm on n processors solves Mutual Exclusion by using two arrays, each of size n. So 2n is an upper bound on the number of registers ne...