Stop and Go The proxy will wait until it receives k packets, then send the k packets out in a random order. Pooling The proxy will keep a pool of...

• Stop and Go The proxy will wait until it receives k packets, then send the k packets out in a random order. • Pooling The proxy will keep a pool of between k and 2k packets at all times. When the pool gets to size 2k, it will randomly select k packets to send from its pool, then wait for k more packets to arrive. a. What are the performance implications of the two designs? How does their throughput compare? How does their latency compare, specifically how much longer do we expect a packet to wait in a Pooling proxy compared to a Stop and Go proxy? b. How large is the anonymity set for each packet sent out from a Stop and Go proxy? How large is it for a Pooling proxy? c. The absolute size of the anonymity set is not really a "fair" comparison between the two designs because it's exponentially unlikely that a packet waits for a very long time. Instead, it has been proposed to compute the entropy or the min-entropy of the distribution of most probable senders for a packet. How do the two designs compare on these metrics?