1. Route Reply Packet

Note that “a route reply packet” is different from “a reply route discovery packet”. The former is an ACK with respect to a data packet, while the latter is a reply packet with respect to a route discovery packet. In the thesis, we use the term “a reply route discovery packet” interchangeably with the term “a route discovery reply packet,” both referring to a reply packet to a route discovery packet.

2. Receipt Packet

A hop-by-hop receipt packet at the network layer may not be needed if the underlying link-layer subsystem can inform the sender node that a packet cannot be delivered to the next hop because the link between the sender node and the next hop breaks. The thesis does not consider this option because for the case we evaluate the system with a network topology simulator (to be described later), the simulator does not have link-layer mechanisms, and for the case without the use of a network topology simulator we do not have access to a link-layer API for IEEE 802.11 to allow the link layer subsystem to inform a sender node that a packet cannot be delivered to the next hop.

3. On Calling Packet

While in real situations, a user can chat with multiple users simultaneously, our wireless ad hoc messenger only allows one node to chat with another node due to lack of screen space in a Pocket PC.

4. Sequence Number for Route Discovery Packets

A negative sequence number is used for a route discovery packet to distinguish from the sequence number of a data packet; -4 used as the initial sequence number for route discovery packets is an arbitrary choice.

5. Synchronization of Time Clock in Each Pocket PC

Note that the synchronization process here refers to all nodes starting an experiment simultaneously. There may be clock drifts that exist in different mobile devices, which we do not consider in the thesis. Also we do not move mobile devices in an experiment, but sit them all in the same room with each of them reading in the same configuration file and following the mobility pattern defined in the configuration file to emulate node movements in an experiment. Therefore, the actual physical distance separating two mobile nodes is mush shorter than the virtual distance separating the same two nodes defined in the configuration file.