Ad Hoc Networks : Technologies and Protocols - University of ...
Ad Hoc Networks : Technologies and Protocols - University of ...
Ad Hoc Networks : Technologies and Protocols - University of ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
134 Transport Layer <strong>Protocols</strong> in <strong>Ad</strong> <strong>Hoc</strong> <strong>Networks</strong><br />
will contend with the data stream on the forward path <strong>and</strong> reduce the rate enjoyed<br />
by the data stream. (ii) If the forward <strong>and</strong> reverse paths are not the same,<br />
the progress <strong>of</strong> the TCP connection will be dependent on both the forward path<br />
<strong>and</strong> reverse path reliability. Thus, the chances <strong>of</strong> a connection stalling increase<br />
when different paths are used. Note that even if the forward <strong>and</strong> reverse paths<br />
are different, due to the shared channel in the vicinity <strong>of</strong> the sender <strong>and</strong> the<br />
vicinity <strong>of</strong> the receiver, the data <strong>and</strong> ACK streams will still contend with each<br />
other. Figure 5.4(a) shows the number <strong>of</strong> times the data stream <strong>and</strong> the ACK<br />
stream experience independent path failures for the 1 flow scenario. It can<br />
be observed that the forward <strong>and</strong> reverse paths experience the same order <strong>of</strong><br />
magnitude <strong>of</strong> failures.<br />
TCP relies on retransmission timeouts as a backup loss detection mechanism.<br />
As described in Section 5.2.2, the RTO value for a TCP connection can be<br />
considerably inflated <strong>and</strong> vastly different from the optimal value. Figure 5.2(c)<br />
presents the average <strong>of</strong> the maximum RTO values set by connections during<br />
their lifetimes. It can be observed that for higher rates <strong>of</strong> mobility, the maximum<br />
RTO values scale up to few tens <strong>of</strong> seconds. This is true even in the case <strong>of</strong><br />
heavy loads. This can result in significant time delays in loss recovery, <strong>and</strong><br />
hence result in gross under-utilization <strong>of</strong> the available b<strong>and</strong>width.<br />
5.2.7 Absolute Impact <strong>of</strong> Losses<br />
In the discussions thus far in the section, we have touched upon the negative<br />
impact <strong>of</strong> mobility related losses on TCP’s performance. Losses, in addition<br />
to being inaccurate indicators <strong>of</strong> congestion for TCP, also have an absolute<br />
impact on the throughput performance <strong>of</strong> the TCP connection. In this section,<br />
we pr<strong>of</strong>ile some <strong>of</strong> the directly contributing causes for such losses.<br />
Impact <strong>of</strong> MAC Layer. The MAC layer is responsible for detecting the<br />
failure <strong>of</strong> a link due to congestion or mobility. Since the MAC layer (IEEE<br />
802.11 DCF) has to go through the cycle <strong>of</strong> multiple retransmissions before<br />
concluding link failure, there is a distinct component associated with the time<br />
taken to actually detect link failure since the occurrence <strong>of</strong> the failure. Importantly,<br />
the detection time increases with increasing load in the network. A<br />
high MAC detection time will result in a higher likelihood <strong>of</strong> the TCP source<br />
pumping in more packets (upto a window’s worth) into the broken path, with<br />
all the packets being lost <strong>and</strong> the source eventually experiencing a timeout.<br />
When a link failure is detected by the MAC layer, the link failure indication<br />
(in DSR) is sent only to the source <strong>of</strong> the packet that triggered the detection.<br />
If another source is using the same link in the path to its destination, the node<br />
upstream <strong>of</strong> the link failure will wait till its MAC layer receives a packet from the<br />
other source. Then the MAC layer will go through its cycle <strong>of</strong> retransmissions<br />
to detect the link failure <strong>and</strong> only then would that source be informed <strong>of</strong> the