Agent-based Distance Vector Routing: A Resource Efficient and Scalable approach to Routing in Large Communication Networks Page: 17
The following text was automatically extracted from the image on this page using optical character recognition software:
shows the aggressive nature of DVR inherent in its rapid path-cost conver-
gence when compared to the moderate yet comparable convergence behav-
ior of ADVR. It can be shown, that in a static network, a single agent can
achieve the correct convergence of routing tables at all nodes in the network,
provided that it uses an appropriate migration strategy, which allows for com-
plete traversal of the network. Nevertheless, a single agent is insufficient to
complete this task in a time that is comparable to that of concurrent messag-
ing i.e. DVR. Hence, a population of agents will have to be deployed. These
agents implicitly cooperate, thereby accelerating the process of route discovery
and path cost convergence.
Route discovery plays an important role in the performance of communication
networks. It is crucial to evaluate any routing algorithm with respect to the
speed at which every node in the network obtains a route for every other node
in the network. Even if these routes are sub-optimal, they provide a benchmark
to measure the availability of the network to be used by other applications.
Figure 9(b) depicts the number of nodes that acquire complete connectivity
to all other nodes in the network over time. It is observed that the aggressive
parallelism in DVR facilitates quick assimilation of network connectivity for
DVR. On the other hand, a small population of constituent agents, restrained
in their concurrency are insufficient to discover routes as rapidly as DVR.
Route discovery in ADVR can be improved to outperform DVR by escalat-
ing the agent population, thereby increasing the degree of concurrency. Even
though increasing the number of agents in the network increases the resource
consumption by agents, it is extremely low when compared to DVR. It is im-
perative to note that the performance of ADVR in terms of route discovery is
greatly affected by the migration strategy adopted by the agents. A detailed
comparison of the migration strategy is presented in (Amin et al. 2001).
4.2.3 Analysis of Agent Population in ADVR
As mentioned earlier, agents are the carriers of information in ADVR. Hence,
the agent population in the network determines the resource overhead. A static
agent population represents an upper bound on the degree of message concur-
rency, and the hence resource overhead. All the above experiments assume a
fixed agent population, however Figure 10(a) shows the the effects of dynamic
agent population control mechanisms using Node Pheromones. As explained
in Section 3, values of I, Q, and A have to be manipulated manually in order
to exercise effective control on agent population. It was observed that irre-
spective of the initial population, the system converges to a stable number
of agents in the system. Networks initialized with a small number of agents
escalate the agent population to a certain value thereby improving the path-
cost convergence of the network. On the other hand, networks initialized with
a large number of agents realize the per-agent overhead and continuously re-
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Amin, Kaizar A. & Mikler, Armin R. Agent-based Distance Vector Routing: A Resource Efficient and Scalable approach to Routing in Large Communication Networks, article, March 25, 2002; [New York, New York]. (digital.library.unt.edu/ark:/67531/metadc111275/m1/17/: accessed March 17, 2018), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT College of Engineering.