Information Technology Reference
In-Depth Information
overlay is formed by the entirety of all groups.
Typically, the overlay hierarchies are two-tier
or three-tier. Their key difference refers on the
number of groups each tier has, the structure of
the overlay each group forms and whether super-
peers do exist. The existence of such super-peers
offers increased computing resources, stability,
and effectiveness in routing.
In an attempt to solve the deficiencies of DHT,
the nonDHT P2Ps avoid hashing. Hashing ignores
data locality and cannot support range queries,
two serious disadvantages. SkipNet, SkipGraph
and TerraDir are three of the most representative
such systems (Li & Wu, 2004). SkipNet stores data
close to users, SkipGraph offers support to range
queries, while TerraDir is targeted for hierarchi-
cal name searches. Searching in these systems
is achieved based on the specified neighboring
relationships between nodes.
In loosely structured P2Ps, the overlay struc-
ture is not strictly specified. It is either formed
based on hints or formed probabilistically. Some
systems develop their structure based on hints or
preferences, while in others the overlay is con-
structed probabilistically. Searching in loosely
structured P2P systems depends on the overlay
structure and how the data is stored. Approaches
do exist, where searching is application dependent,
based on hints or done by reducing the numerical
distance from the querying source to the node
that stores the desired data.
process to locate the destination. It broadcasts a
route request to its neighbors, which then forward
the request to their neighbors and so on, until the
destination or an intermediate node with a route
to the destination is located. Nodes are identified
by their IP address and maintain a broadcast ID,
which is incremented after every route request
they initiate. The broadcast ID together with
the node's IP address, uniquely identify a route
request. In the same manner, the transmitted data
requests can be identified.
Routing algorithms for MANETs are radi-
cally different from the traditional routing (e.g.,
Open Shortest Path First) and information search
protocols (e.g., Distributed Hash Table) used in
hardwired networks, due to the absence of “fixed”
infrastructure (servers, access points, routers, and
cables) in a MANET as well as the mobility of
the nodes. For wireless ad-hoc networks, there
have been proposed various routing/discovery
protocols, which roughly fall into the following
categories (Agrawal & Zeng, 2003): (a) table-
driven routing protocols, (b) source-initiated on-
demand routing protocols, and (c) hybrid routing
protocols. Apart from the former, which require
consistent, up-to-date routing information from
each node to every other node in the network and
thus are practically unfeasible for large-scale and
dynamic MANETs, the remaining two families
of information (node) discovery protocols rely
on some form of
broadcasting
; broadcasting is
best suited in cases where information packets
are transmitted to multiple hosts in the network.
Flooding
is the simplest broadcasting approach,
where every node in the network forwards the
packet exactly once; flooding ensures full cov-
erage of the MANET provided that there are no
network partitions. Flooding, though, generates
too many redundant transmissions, causing the
broadcast storm problem
(Ni, Tseng, Chen, &
Sheu, 1999).
Numerous algorithms have been proposed to
address this problem (Lou & Wu, 2004). They
can be classified as follows: (a)
probabilistic
Information discovery/provision In
Wireless mobile ad-hoc networks
A Mobile Ad-hoc NETwork (MANET) is a col-
lection of wireless Mobile Hosts (MH) forming a
temporary network without the aid of any central-
ized administration or standard support services
regularly available on the wide area network to
which the hosts may normally be connected. When
a source node desires to send a message to some
destination node and does not already have a valid
route to that node, it initiates a path discovery
Search WWH ::
Custom Search