CPS 214 Written Homework 3 - Fall 1997
Due: Monday, Oct. 20, 5:30 PM
10 points
- Consider a sliding window protocol across a channel in which the
one-way propagation delay is 270 ms. If the channel has a transmission
rate of 1Mbps, is there any performance benefit in using a send
window larger than 1024 frames in size? Assume frames are 10000 bits in
size.
- Assume that there are never any transmission errors over a
particular physical channel. Which protocol would you expect to have
better performance across the channel: selective repeat or Go Back N?
- When running protocols within the x-kernel, execution begins at
the label ``main'', which is supplied by the x-kernel. How then do the
application routines (e.g., the asp test routines) get invoked at run
time? Hint: Use gdb (or some other debugger) to single-step
through the execution. Place a breakpoint at asp_init, and
then follow the execution.
- Consider building a CSMA/CD network running at 1Gbps over a 200
meter cable. The signal speed in the cable is 200,000 km/sec. What is
the minimum frame size?
- Even though both token ring and CSMA/CD technologies are popular
in LANs, WANs never use the CSMA/CD design. Why?
- A 16-Mbps has a token holding time of 10 msec. What is the
longest frame that can be sent on the ring?
- Consider an unslotted FDDI-like ring network that has a
circumference of 100km long and runs at 100Mbps. After sending a
frame, a station drains the frame form the ring before regenerating
the token. The signal propagation speed in the fiber is 200,000 km/sec
and the maximum frame size is 4000 bytes. What is the maximum
efficiency of the ring? That is, what percentage of the available
bandwidth can be used for sending data? You may ignore such sources of
overhead as header fields.
- Peterson (chapter 4): 2.
- Consider the new Arpanet routing algorithm. If there are 50
switches in the network, each switch connects (on average) to 3 other
switches, and new link metric updates are sent every 1/2 second, how
many packets per second will cross each link?
- Tanenbaum (chapter 5): 14.
- Tanenbaum (chapter 5): 20.
- Some audio enthusiasts complain that audio CDs don't sound as
good is the analog recordings recorded onto vinyl LPs. However, we
know from Nyquist that CDs carry frequencies above 20kHz (CDs are
sampled at 41kHz), which is beyond the range of human hearing. Are
such audio enthusiasts imagining things, or is there a rational
explanation for their complaints? Hint: samples are stored at 16-bit
values.