ns2 project in Tennessee
ns2 project in Tennessee the success of this and other protocols and the proper
functionality of the agents rely on having a suitable naming
(identification) scheme for the agents. Many ns2 project in Tennessee mechanisms
can be used; however, for the system to be scalable, the
naming scheme needs to be scalable also. One suitable
scheme is the hierarchical naming used for the Internet.
Here, the leaders on the top level of the ns2 project in Tennessee hierarchy take a
common root name followed by each machine/cluster
name. The next levels use their leader’s name as a prefix
to their names. For example, assume that the ns2 project in Tennessee structure
shown in Fig. 2 belongs to UNL and, then, the top-level
leaders can be UNL.L1, UNL.L2, and UNL.L3. Leader 4 is
then called UNL.L3.L4 and the agents under leader 2, for example, are called UNL.L2.A1, UNL.L2.A2, etc. Such a
scheme, while potentially complicated for ns2 project in Tennessee a small system,
allows the system to systematically grow without any need
to change previously assigned names or the ns2 project in Tennessee naming scheme
itself. This also allows agents to use the machines’ actual
Internet URLs as their names, thus allowing ns2 project in Tennessee easy access
through the Internet. Adopting this scheme, however,
requires some form of neighbor discovery ns2 project in Tennessee mechanism as
in IPv6 [14] to ensure the use of unique names for the
participating agents. In general, the overhead incurred in
constructing a hierarchical structure is relatively high, thus
it may not benefit a system with a small number of nodes.
However, it is essential in two environments:
1. The system is composed of multiple ns2 project in Tennessee smaller
systems (clusters, NOW, multiprocessor machines,
etc.) that do not have full connectivity to all their
nodes. Thus, the head node in each subsystem is
assigned a leader that is responsible of connecting
it to other subsystems. 2. The system includes ns2 project in Tennessee very large clusters comprising
tens/hundreds of nodes, thus accessing all nodes in
a linear fashion is very time consuming. Here, the
threshold needs to be selected to optimize the
utilization of the suitable structure. Analytical
models or experimental evaluations can be used to
select that value. This protocol is used in case a ns2 project in Tennessee leader fails to respond to an
AM message sent by another leader. If a leader Lx at one
level times out before receiving an AMA response from
another leader, say Ly, the following steps are taken by Lx
to try to recover from the problem. 1. Lx broadcasts the problem to all other leaders at the
same level using the LNRM and informs them that it
will try to solve the problem. 2. Lx pings the node/machine where Ly resides to see
if it is connected and up. 3. If the node is still up, then a. Lx initiates a remote agent activation command
to reactivate the agent using the AAM and b. when the new agent is up, Lx activates it as a
leader and sends it all relevant leader information. The new leader, Ly, uses the startup
protocol to restore its information. 4. If the agent does not ns2 project in Tennessee reinitialize (e.g., has been
deleted from the node) or the node does not respond if a connection exists to another node in the
cluster, Lx activates that node’s agent as a
leader. The new leader then assumes ns2 project in Tennessee its new
role and updates its routing and resource
information using the startup protocol,
b. if no connection exists, Lx reports the problem to
the administrator and excludes all routing
information to the cluster led by Ly from the
routing tables.