We learned about the primary mode of operation of TCP Reno, where cwnd increases by 1 for each RTT. Equivalently, cwnd increases by 1/cwnd for each ACK received. This operational mode is called congestion avoidance. However, this results in linear increase of cwnd over time. At the beginning of a TCP connection, we would like to quickly ramp up cwnd before transitioning to congestion avoidance mode. Most TCP protocols have a (somewhat confusingly named) slow start phase that accomplishes exactly this. In this mode, cwnd increases by 1 for each ACK received. (a) If we plot cwnd versus time, instead of having a linear increase as in congestion avoidance, what is the rate of increase in slow start? (b) Draw a space-time diagram for the slow start phase, for 4 RTTs. Assume cwnd starts at 1