This assignment is due no later than 6:20 p.m. on Tuesday, 3 December.
This homework assignment has six questions; answer all of them. Answers
should no longer than half a page, or around 100 to 150 words. Please site
references you use to answer your questions, including the textbooks you’re using.
Your homework may turned-in on paper or via rclayton@monmouth.edu
by
email. If you mail your assignment, please submit a printable document
— a PostScript .ps
or PDF .pdf
document, for
example — and not a source document — a
Word .docx
or Latex .tex
document, for example. Make sure you mail a
printable document by the deadline.
-
Your network consulting firm has been contracted to build an unstructured
peer-to-peer network with a topology following a power-law relation. The
network consists of two nodes classes: expensive, capable nodes and less
expensive, less capable nodes; capability is measured in throughput and storage
capacity. Explain how you would use the different node classes to build a
capable network as cheaply as possible, where cost involves only the cost of
the nodes. A capable network is one in which every node is able to handle the
offered load most of the time (that is, “use the cheapest nodes
everywhere” is not a good answer).
-
You want to store an object O in a Gnutella network of 1000 nodes such
that 50% of the searches for O have an expected hop count of at most 2.
What is the smallest number of copies of O you should store? Assume
O is a shareable resource, so there's no advantage to storing more than
one copy of O at a node.
-
Networks A and B are both unstructured peer-to-peer networks. One of
the networks A and B has a topology that satisfies the small-world
relation, the other doesn't. Explain how you would distinguish the networks by
topology.
-
A colleague of yours believes that objects located at low-addressed nodes in a
Chord network will be found more quickly in searches than objects located a
high-addressed nodes. What is your analysis of your colleague's belief?
-
Many structured peer-to-peer networks are based on a distributed hash table
(DHT). One of the tricky aspects of normal (undistributed) hash-table data
structures is making them dynamic so they can grow and shrink in capacity as
needed. Pick a DHT-based structured peer-to-peer network and describe how it
handles increases and decreases in capacity.
-
Assume an Edutella-like hybrid peer-to-peer network in which the structured
part of the network is an N-dimensional hypercube and the hosts in the
unstructured part of the network are directly attached to a hypercube node.
What is the maximum hop count between two nodes in the unstructured part of the
network?
This page last modified on 2013 April 21.