Académique Documents
Professionnel Documents
Culture Documents
q=1
(2)
In general, among all the downstream nodes, the most
congested one is defined as the worst node and needs
more attention. On this basis we propose a control
scheme,
Rq (n) =+a xi(n Ti) xi
n
n
I=1
) + bjRq(j) +
T
1
j=1
c xk(nTk) xk (nTk1
N
1
k=1
) (3)
These coefficients are used to locate all the poles of the
closed-loop equations (2) and (3) within the unit circle
to ensure stability. The component xi is the target
queue length and is the maximum sending rate of
sources. To save computation time and CPU resources,
we present a simple SPI control scheme,
Rq (n) = +a xi(nTi) xi
N
I
I=1
) + bjRq(n
T
j
j=1
j) (4)
Similarly, these coefficients should make all the poles of
the closed-loop equations (2) and (4) within the unit
circle to ensure stability. In (3) and (4), if the buffer
occupancy of the switch node i is measured at the
instances n i, after the feedback delay i, the BCP
reaches the controller located at the source q (q =1, 2 .
. . m), and the source then takes out the buffer
occupancy of the destination nodes at time slot t =n. By
doing so, the proposed controller can be expected to
have flexibility to cope with the sharp oscillation in
buffer occupancy that could cause the network to lose
packets. In addition, the calculation in (3) and (4) is
completely independent of virtual connections traveling
through the multicast session. This means the scheme
has scalability.
SPID and SPI Algorithms:
SPID stands for Statistical Protocol Identification, it is
used in the application layer of the network for the
communication purpose, which reliably identifies the
application layer protocol by using statistical
measurements of flow data as well as application layer
data. The SPID algorithm utilizes Kullback-Leibler
divergence measurements to compare probability
vectors created from observed network traffic to
probability vectors of known protocols.
The SPID algorithm is designed to reliably identify
which protocol is being used in a network
communication session. Key requirements for the
algorithm are:
1. Small protocol database size
2. Low time complexity
3. Early identification of the protocol in a session
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 9 Sep 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 3024
4. Reliable and accurate protocol identification
The motivation for these requirements 1 and 2 are that it
shall be possible to run the SPID algorithm in real-time
on an embedded network controller, which has restricted
memory and converting capability. Motivation for
fulfilling 3 is that it shall be possible to use the results
from the SPID algorithm in a live traffic capturing
environment to automatically take precautions in order
to provide quality service to an active session, block
illicit traffic or store related traffic for off-line analysis1.
Haffner et al. (2005) also point out the need for
enterprises to degrade P2P services (via rate-limiting,
service differentiation and blocking) on their networks
in favour of the performance for business critical
applications. I have therefore required that the protocol
must be identifiable, with the SPID algorithm, based on
only the four first application data packets (i.e. not
counting the flow control signaling packets) for a
session. Note 4 have no further motivation than the
obvious in order to provide a high quality service.
SPID performs protocol identification by using
statistical fingerprints. The statistical fingerprints, of a
session or known protocol, are part of an object called
Protocol Model. Here application layer protocol in a
session is identified by comparing its protocol model to
protocol models of known protocols.
Fig 2: Process Identification Data Flow
The SPID algorithm present a reliable method for
identifying protocols based on both flow and
application. Strong identification feature is a result of
the innovative protocol attribute functions; this protocol
gives rich data format and the model comparison
algorithm.
The SPID algorithm does not require any manual
creation of protocol signatures; but it does require
training data that is pre-classified based on protocol, in
order to generate a protocol model DB. Protocol model
content format allows for the protocol models to be
updated as new training data made avail, without having
access to the last used training data.
SPI (Symbolic Probabilistic Inference) algorithm is used
for resolving general queries in Bayesian Networks.
Algorithm SPI implements the concept of dependency
directed backward search to probabilistic content, and is
incremented with respect to both the queries and
observations.
The SPI algorithm consists of several major processing
steps. The first step is to organize the nodes of a
Bayesian network into a tree structure for query
processing. We call these structures SPI trees. In the
second step, queries are directed to the root node of the
SPI tree. The query is decomposed into queries for the
node's sub trees. This recursive procedure continues
until a particular query can be answered at the node at
which it is directed. The answer is then computed and
returned to the next higher level in the SPI tree. Once a
node has responses from all of its sub trees it can
compute its own response to its predecessor node. This
process terminates when the root node processes all the
responses from its sub trees. An SPI tree is constructed
by organizing the nodes of a Bayesian network into a
tree structure. The only constraint on the construction
process is that if there is an arc between two nodes in the
original network, then one of the nodes must be a direct
or indirect predecessor of the other in the SPI tree. This
constraint allows many possible SPI trees. The first step
in building a SPI tree is to choose the root node. This
done by computing the maximum node to node distance
for each node. The node that has the smallest maximum
distance is chosen as the root node. This heuristic is
designed to produce a. "bushy" SPI tree which can take
advantage of distributed processing. The second step is
to use maximum cardinality search to build the tree from
the root node. This step constructs the tree based on the
connectivity principle and guarantees satisfaction of the
tree construction constraint.
Fig: An Example Network
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 9 Sep 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 3025
Fig: SPI Tree
The above two figures shows the implementation of SPI
algorithm the network which is not defined in a proper
way can be classified into a network which will have
complete sense and functionality for the traversal.
PERFORMANCE
To evaluate the performance of the studied multicast
congestion control, we work on the following two
simulation models, and are mostly interested in
analyzing the transient behaviors of networks. With the
performance analysis report, the duration of response
time, retrieval rate of consumer and stable state of buffer
occupancy are the main concerns. From the view of
control theory, a control scheme with short response
time has the following advantages: when the buffer of
receiver nodes is close to the threshold, one may tell the
sending node to reduce the sending rate and prevent the
loss of packets as soon as possible; while when the
available bandwidth increases, the sending node
increases the sending rate as soon as possible and
enhances the utilization rate of the bandwidth. In
simulations, we process the nodes that have small
differences of time delay and sending rate. Later we
unify the time delay and rate of sending; because the
situation of every node in each group (about 20
receivers) is almost same, we choose only one node
from each group as a main node. We then assume that
the link delay is dominantly compared with other delays
like processing delay and queuing delay. Simulations are
carried out over a wide range of patterns, and
propagation between two different nodes can lie in the
LAN case or the WAN case. Based on different models
in the network and network continuous changing
behavior, we present simulated experiments. In
simulation 1 (see Fig. 3), the multicast source S1 sends
data packets at 0 ms and the multicast source S2 starts to
send data packets at 1000 ms in the simulation time;
then the joining of S2 enhances the network dynamic
nature, and demonstrates the efficiency of the SPID and
SPI schemes. In simulation 2 there are more receivers
and longer delay than in first model, and then set
appropriate parameters to enable system stability. In
each model, we compare and evaluate the performance
of the SPID and SPI systems
Fig. 3. Multicast simulation model 1.
TABLE I
PARAMETERS IN THE SIMULATION MODEL 1
Parameters Reciver1 Reciver21 Reciver41
xi(Mb) 70 80 120
Li(Mbps) 2 3 4
TRji(msec) 4 8 16
A. Simulation 1
In this simulation, we focus on comparing the transient
behavior of the SPID and SPI network systems based on
Fig. 3. The relevant notations and assumptions are listed
in Table I, and =16 ms, N =60, and n1 =n2 =n3 =
20. According to the aforementioned simulation
parameters and system stability analysis in Section IV to
select the control gains, one computes the relevant
control parameters based on Functions 1820 (for SPID
scheme) and Functions 2829 (for SPI scheme). For
SPID scheme, we set to be 1/20, which is stable in the
system. Then, a = 1/600, c = 1/400, and b =[b1, b2,
b3. . .b16] ,
For SPI scheme, we also set to be 1/19, which is stable
in the system. Then, a =1/380 and b =[b1, b2, b3. . .
b16 ],
In this section, the simulation results of simulation 1 are
Shown in Figs. 48. The sending rates of sources S1 and
S2 are shown in Figs. 45, respectively. The initial
sending rate of multicast source S1 is 6 Mb/s. It can be
seen that, although the sending rate of the multicast
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 9 Sep 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 3026
source S1 has some fluctuation at first, as time goes on,
the sending rate is gradually adjusted and quickly at the
value of 2 Mb/s within 74.1 ms for SPID controller, and
325.4 ms for SPI controller. When the multicast source
S2 starts to send data packets at a rate of 1000 ms
Fig. 4 Sending rate of S1
Fig. 5 Sending rate of S2
Of source S1 has some fluctuation in response to the
multicast source S2 joins, and stabilizes at the latest
value of 1 megabytes per second within 287.5 ms for
SPID controller and 362.5 milliseconds for SPI
controller. Figs. 68 show the buffer transient responses
of the receiver node 1, node 21, and node 41,
respectively. In Fig. 6, these buffer occupancies of the
bottleneck receivers all have some fluctuation in the
beginning. Later gradually they become stable at the
value 27.5 megabytes for SPID controller and 230
megabytes for SPI controller. When the multicast source
S2 starts to send data packets at 1000 ms.
Fig. 6 Buffer occupancy of receiver 1
Fig. 7 Occupancy of receiver 21
The buffer occupancies of the receivers have some
fluctuation in response to the multicast joining with
source S2. For group1 which is receiving, they quickly
become stable at the value 240 Mb for SPID and 255
Mb for SPI. In Figs. 78, some packets get stored in the
buffer of receiver 21 and receiver 41 at the beginning.
As time goes on, the controller starts to adjust the
transmission rate of the source, and the remaining
packets in the buffer are cleared. Table II gives the
comparative analysis of the simulation model 1. The
second and fourth rows are about the stable sending
rates of sources S1 and S2 in time range [0, 2000] ms.
International Journal of Computer Trends and Technology (IJCTT) volume 4 Issue 9 Sep 2013
ISSN: 2231-2803 http://www.ijcttjournal.org Page 3027
the third and fifth rows are about the response times of
stable sending rates in the two different time ranges.
Fig. 8. Multicast system model 2
It is clear that our two schemes could quickly adjust the
buffer occupancy and the rates of sending node based on
the dynamic network environment. These simulation
results demonstrate our SPID and SPI controllers,
efficiency of system, faster response, less packet loss,
and more scalability. Based on these approaches, the
source adjusts the sending rate gradually to stabilize the
buffer occupancy and rate of sending node quickly.
Based on the earlier comparative analysis in Table II, we
find that SPID controllers can provide better
performance than SPI controllers in terms of the fast
response of the controlled sending rates and low buffer
occupancy of bottleneck receivers.
CONCLUSION
We have told that by implementing above
application called multi-input multi-output network
systems we can give assurance such as speed of
transmitting data, network traffic congestion, and
storage of data without reducing data in a packet format
by making each source or destination act as a
client/server. We have shown with the implementation
and moreover, we calculated time delay when data is
transferred from multiple inputs to multiple output
resources with this we can assure the data transferred
will reach destination with more efficiency and also
there will not be any loss of data in the transmission.
REFERENCES
N. X. Xiong, Y. He, L. T. Yang, and Y. Yang, A self-
tuning reliable dynamic scheme for multicast controlling
flow.
S. Dearing, Host extensions for IP multicast Stanford
University, Stanford, CA.
Shin and Zhang Analysis of feedback synchronization
signaling delay for multicasting flow control,
Siu Y, Suzuki H and Ren Performance of congestion
control algorithms in multicast ABR service
AUTHORS PROFILE
Satyavathi Gantala,
pursuing M.Tech(CSE) from
Vikas Group Of Institutions
(Formerly known as Mother
Theresa Educational Society
Group of Institutions),
Nunna, Vijayawada.
Affiliated to JNTU,
Kakinada, India
Boppudi Swanth,
Working as Assistant
Professor (CSE), Vikas Group
of Institutions (Formerly
Mother Teresa Educational
society Group of
Institutions), Nunna,
Vijayawada, Affiliated to
JNTU, Kakinada, India
Betam Suresh, is working
as an HOD, Department of
Computer science
Engineering at Vikas Group
of Institutions (Formerly
Mother Teresa Educational
society Group of
Institutions), Nunna,
Vijayawada, Affiliated to
JNTU, Kakinada, India