Vous êtes sur la page 1sur 5

International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 7July 2013

ISSN: 2231-2803 http://www.ijcttjournal.org Page 2141



Aggregation Dynamics in Service Overlay Networks for File
Sharing
K.Deepika Reddy*1, E.Krishna*2, W.Vidyavardhan*3
Assistant Professor, Dept of Computer Applications, SNIST, Ghatkesar, Hyderabad, AP, India
M.C.A Student, Dept of Computer Applications, SNIST, Ghatkesar, Hyderabad, AP, India
M.C.A Student, Dept of Computer Applications, SNIST, Ghatkesar, Hyderabad, AP, India
ABSTRACT
This paper related to link monitoring.
Restricts the selfish nodes and creates the network
with only non selfish nodes. Non selfish nodes
network maintains good link connectivity without
any failure and provides stable and optimal solution.
This network shows good performance and scalable
solution. Previous network systems whenever create
uses the immediate neighbour selection technique.
This technique selects less distance of nodes for
creation of network without any verification.
Sometimes network failures are generated because of
insufficient energy levels of nodes. Automatically
packets are not delivering within time. In failure
location add the node with new energy levels. This is
called as a rewiring. Travelling the total number of
packets takes more amounts of time and energy
levels. To complete the packets transmission,
network expects dynamic changes. It is not stable and
dont provide any kind of reliable solution. We
introduce new design of network systems with the
help of selfish neighbour selection technique. This
technique uses the overlay routing for restrict the
selfish nodes. After removing the selfish nodes,
remaining non selfish nodes provides authentication
network. All authentication nodes have asymmetric
distance. Asymmetric distance nodes give the
interference problems. Using EGOIST server
implementation converts the asymmetric distance
nodes to symmetric distance nodes. It is possible
based on prototype design here. These symmetric
distance nodes distribute the data efficiently without
loss. These types of networks are called stable.
KEYWORDS: Wiring Strategies, Stable Network
Creation, Egoist Server.
I.INTRODUCTION
An overlay network is a layer of virtual
network topology on top of the physical network,
which directly interfaces to users. With the rapid
advancement of Internet and computing technology,
much more aggregate information and computing
resources are available from clients or peers than
from a limited number of centralized servers. A
foundational issue underlying many such overlay
network applications is that of connectivity
management. Connectivity management is called
upon when having to wire a newcomer into the
existing mesh of nodes (bootstrapping) or when
having to rewire the links between overlay nodes to
deal with churn and changing network conditions.
Connectivity management is particularly challenging
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 7July 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2142

for overlay networks because overlays often consist
of nodes that are distributed across multiple
administrative domains, in which auditing or
enforcing global behaviour can be difficult or
impossible. As such, these nodes may act selfishly
and deviate from the default protocol, by utilizing
knowledge they have about the network, to maximize
the benefit they receive from it. Selfish behaviour has
been reported in studies relating to selfish (source)
routing and free riding in P2P file-sharing networks.
Selfish behaviour also has many implications for
connectivity management. In particular, it creates
additional incentives for nodes to rewire, not only for
operational purposes (bootstrapping and substituting
nodes that went offline), but also for seizing
opportunities to incrementally maximize the local
connection quality to the overlay. While much
attention has been paid to the harmful downsides of
selfish behaviour in different settings , the impact of
adopting selfish connectivity management techniques
in real overlay networks has been an open problem .
In this paper, we formulate and answer such
questions using a combination of modelling, analysis,
and extensive simulations using synthetic and real
datasets. Our starting point is the definition of a
network creation game that is better suited for
settings of P2P and overlay routing applications
settings that necessitate the relaxation and/or
modification of some of the central modelling
assumptions of prior work. In that regard, the central
aspects of our model are bounded degree, directed
edges, non uniform preference vectors, and selfish
neighbour selection.
II. NETWORK MODEL
Previous work on overlay network creation
has focused on physical telecommunication networks
and primarily the Internet. Overlay networks are
substantially different, which prompts us to consider
the following overlay network model. We start by
relaxing and modifying some of the central modelling
assumptions of previous work. In that regard, the
central aspects of our model are the following.
1) Bounded Degree: Most protocols used for
implementing overlay routing or content sharing
impose hard constraints on the maximum number of
overlay neighbours. For example, in popular versions
of Bit Torrent, a client may select up to fifty nodes
from a neighbours list provided by the Tracker of a
particular torrent file. In overlay routing systems, the
number of immediate nodes has to be kept small so
as to reduce the monitoring and reporting overhead
imposed by the link-state routing protocol
implemented at the overlay layer. Hard constraints on
the number of first-hop neighbours are also imposed
in most P2P systems to address scalability issues,
uplink fragmentation, and CPU consumption due to
contention. Motivated by these systems, we explicitly
model such hard constraints on node degrees. Notice
that in the prior studies cited, node degrees were
implicitly bounded by virtue of the trade off between
the additional cost of setting up more links and the
decreased communication distance achieved through
the addition of new links. We also note that some of
these earlier network creation games were proposed
in the context of physical communication networks.
In such networks, the cost of acquiring a link is
instrumental to the design and operation of a critical
infrastructure. Such concerns do not apply in the case
of overlay networks such as those we consider in this
paper.
2) Directed Edges: Another important consideration
in the settings we envision for our work relates to
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 7July 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2143

link directionality. Prior models have generally
assumed bi-directional links. This is an acceptable
assumption that fits naturally with the unbounded
node degree assumption for models that target
physical telecommunication networks because actual
wire-line communication links are almost exclusively
bidirectional. In overlay settings we consider, this
assumption needs to be relaxed since the fact that
node v forwards traffic or requests to node u does not
mean that node u may also forward traffic or requests
to v.
3) Non- uniform Preference Vectors: In our model,
we supply each node with a vector that captures its
local preference for all other destinations. In overlay
routing, such preference may capture the percentage
of locally generated traffic that a node routes to each
destination, and then the aggregation of all preference
vectors would amount to an origin/destination traffic
matrix. In P2P overlays, such preference may amount
to speculations from the local node about the quality
of, or interest in, the content held by other nodes.
Other considerations may also include subjective
criteria such as the perceived capacity of the node, its
geographic location, or its availability profile.
4)Selfish Neighbour Selection: In a typical overlay
net-work, a node must select a fixed number (k) of
immediate overlay neighbours for routing traffic or
queries for files.1 Previous work has considered this
problem from two perspectives: (1) devising practical
heuristics for specific applications in real
deployments, such as boot-strapping by choosing the
k closest links, or by choosing k random links in a
P2P file-sharing system; and (2) providing
abstractions of the underlying fundamental neighbour
selection problem, which are amenable to theoretical
formulation and analysis as exemplified in the recent
work on Selfish Neighbour Selection (SNS). This
SNS formulation focused on characterizing the
emergent overlay topology when overlay nodes be-
have selfishly and employ Best-Response (BR)
neighbour selection strategies. Using BR a node
chooses the best k neighbours that optimize its
connection quality to the overlay, granted knowledge
of how other nodes have connected among
themselves.
III. DEFINITIONS
Best Response: Given a residual wiring S
i
, a best
response for node v
i
is a wiring s
i
S
i
such that C
i
(S
i

+{s
i
}) C
i
(S
i
+{s
i
}), s
i
s
i
.
Stable Wiring: A global wiring S is stable iffit is
composed of individual wirings that are best
responses. Therefore stable wirings are just pure
Nash equilibria of the SNS game, i.e., they have the
property that no node can re-wire unilaterally and
reduce its cost.
The SNS Game: The selfish neighbor selection
game is defined by the tuple ( V, {S
i
}, {C
i
}), where:
V is the set of n players, which in this case are the
nodes.
{S
i
} is the set of strategies available to the
individual players. Si is the set of strategies available
to v
i
. Strategie correspond to wirings and, thus, player
v
i
has (n1, k
i
)possible strategies s
i
S
i
.
{C
i
} is the set of cost functions for the individual
players. The cost of player vi under an outcome S,
which in this case is a global wiring, is C
i
(S). The
above definition amounts to a non-cooperative, non-
zero sum, n-player game. Let S
i
=S {s
i
} denote the
residual wiring obtained from S by taking away v
i
s
outgoing links.


International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 7July 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2144

IV. EGOIST OVERLAY SYSTEM
1. Basic Design of EGOIST:
Egoist is a distributed system that allows the
creation and maintenance of an overlay network
(evaluated on PlanetLab), in which every node
selects and continuously updates its k overlay
neighbours in a selfish mannernamely to minimize
its (weighted) sum of distances to all destinations
under shortest-path routing. For ease of presentation,
we will assume that delay is used to reflect the cost of
a path, noting that other metrics which we will
discuss later in the paper and which are incorporated
in Egoists implementation could well be used to
account for cost, including band-width and node
utilization.
In Egoist, a newcomer overlay node v
i
connects
to the system by querying a bootstrap node, from
which it receives a list of potential overlay
neighbours. The new comer connects to at least one
of these nodes, enabling it to participate in the link-
state routing protocol running at the overlay layer. As
a result, after some time, v
i
obtains the full residual
graph G
i
of the overlay. By running all-pairs shortest
path algorithm on G
i
, using Dijkstras algorithm, the
newcomer is able to obtain the pair-wise distance
(delay) function d
Gi
. In addition to this information,
the newcomer estimates d
ij
, the weight of a potential
direct overlay link from it-
self to node v
j
, for all v
j
V
i
. Using the values of d
ij
and d
Gi
, the newcomer connects to G
i
using one of
a number of wiring policies. In our implementation,
each node listens to all the control messages of the
link state protocol and propagates them only to its
immediate neighbours. In order to reduce systems
control traffic, each node propagates only unique
messages by dropping messages that have been
received more than once or have been superseded.
There are also two threads, one for estimating d
ij
,
and one responsible for estimating the new wiring
and propagating the wiring to the immediate
neighbours. In order to minimize the load in the
system, a node propagates its wiring to its immediate
neighbours only if this changes. Clearly, obtaining d
ij

for all n nodes requires O(n
2
) measurements.
However, we note that these O(n
2
) measurements do
not have to be announced or be continuously
monitored. In particular, each node needs to monitor
and send updates only for the k links that it chooses
to establish, with O(n) measurements to all nodes in
the overlay done much less frequently namely once
per wiring epoch, which is defined as the period T
between two successive evaluations by a node of its
set of candidate links and possible adoption of a new
wiring (i.e., re-wiring) based on such evaluation.
Since re-wiring is much less frequent than monitoring
of the established k links, the load imposed by the
link-state protocol is only O(nk) and not O(n
2
).
2. Neighbour Selection Policies in EGOIST:
Using BR, a node selects all its k neighbours
so as to minimize a local cost function, which could
be expressed in terms of some performance metric.
Since obtaining an exact BR is computational
expensive under both delay and throughput, we
employ fast approximate versions based on local
search to reduce computational costs and enhance
scalability. In addition to BR, we have also
implemented the following neighbour selection
policies in order to perform a comparative evaluation.
k-Random: Each node selects k neighbours
randomly. If the resulting graph is not connected, we
re-wire some links to enforce a cycle upon it.
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 7July 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 2145

k-Closest: Each node selects its k neighbours to be
the nodes with the minimum link cost. Again, if the
graph is not connected, we enforce a cycle.
k-Regular: In this case, all nodes follow the same
wiring pattern dictated by a common offset vector o
={O
1
, O
2
, . . . , O
k
}, used as follows: node i connects
to nodes i +O
j
mod n, j =1, . . . , k. In our system, we
set Oj =1+(j 1) (n1)/(k+1). One way to visualize
this is to
consider that all nodes are placed on a ring according
to their ids (as with a DHT). Thus, an offset vector
makes each node use its k links to connect to other
nodes so as to equally divide the periphery of the
ring.
V. CONCLUSION
This work started with a study of selfish
neighbour selection under strictly enforced neighbour
budgets and has come up with a series of findings
with substantial practical value for real overlay
networks. First, we have shown that a best-response
(i.e., selfish) selection of neighbours leads to the
construction of overlays with much better
performance than those constructed by simple
random and myopic heuristics. The reason is that by
being selfish, nodes embark on a distributed
optimization of the overlay that turns out to be
beneficial for all. Second, we have demonstrated
through the design, implementation, and deployment
of egoist, that it is indeed feasible to apply our best-
response wiring in practice and that the obtained
benefits are actually much larger under dynamic
environments where the simple heuristics lag even
more. Finally, we have used our egoist prototype for
achieving real-time requirements and carrying the
traffic generated by an online multiplayer P2P game
and has verified all the above observations.
REFERENCES
[1] S. Ratnasamy, M. Handley, R. Karp, and S.
Shenker.Topologically aware overlay construction
and server selection. In INFOCOM 02.
[2] V. Arya, N. Garg, R. Khandekar, K. Munagala,
and V. Pandit, Local search heuristic for k-median
and facility location problems, in Proc. of ACM
STOC 01, Hersonissos, Greece, 2001, pp. 2129.
[3] A. Bharambe, J. R. Douceur, J . R. Lorch, T.
Moscibroda, J. Pang, S. Seshan, and X. Zhuang,
Donnybrook: Enabling large-scale, high-speed,
peer-to-peer games, in Proc. ACM SIGCOMM,
2008, pp. 389400.
[4] L. Qiu, Y. R. Yang, Y. Zhang, and S. Shenker,
On selfish routing in Internet-like environments, in
Proc. ACM SIGCOMM, 2003, pp. 151162.
[5] M. Feldman, K. Lai, I. Stoica, and J . Chuang,
Robust incentive techniques for peer-to-peer
networks, in Proc. ACM EC, 2004, pp.102111.
[6] N. Magharei and A. H. Rasti, Prime: Peer-to-
peer receiver-driven mesh-based streaming, in Proc.
IEEE INFOCOM, 2007, pp. 14151423.

Vous aimerez peut-être aussi