Date post: | 24-Jun-2015 |
Category: |
Technology |
Upload: | iaeme |
View: | 112 times |
Download: | 0 times |
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
105
SEAMLESS VIDEO STREAMING USING IMPROVED HANDOVER
PREDICTION AND SESSION HANDOVER IN MOBILE NETWORKS
Vidhate Amarsinh1, Devane Satish
2
1(Department of Computer Engg, RAIT, Nerul, Navi Mumbai)
2(Department of Information Technology, DMCE, Airoli, Navi Mumbai)
ABSTRACT
The application like video streaming on mobile devices has fetched a lot of attention in the
last decade. The significant problems like handover latency and lack of buffering are the real culprits
in the seamless continuity of video streaming applications targeted on mobile networks. This
paper presents a novel framework which considers an efficient handover prediction and
IntraDomain/InterDomain session handover as tools to take a charge of video continuity
under variable mobility conditions. The results of the simulation study shows that the proposed
framework can improve streaming continuity due to accurate handover prediction, proper
IntraDomain/InterDomain session handover with the support of session rate prediction.
Keywords: Video Streaming, Handover Decision, IntraDomain Handover, InterDomain Handover,
Session Rate.
1. INTRODUCTION
With the advancements of video streaming applications, an integration of mobility and
information services has become an urgent issue in the modern world. The video streaming
applications are span from traditional telecom services such as Voice over IP (VoIP) and video
conference to entertainment and Video on Demand (VoD) [1].
One of the most challenging issues in supporting mobility is the avoidance of flow
interruptions when clients roam from one wireless locality to another. The after effect triggers a
condition called video freeze. In order to provide uninterrupted services and maximum user-
perceived quality, a successful video streaming solution needs to adapt to mobile handover scenarios.
The handover decision typically considers only connectivity signal strength from various Access
Points (AP) in the proximity, which is not adequate to take handover decision for video streaming
applications.
INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING &
TECHNOLOGY (IJCET)
ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)
Volume 5, Issue 4, April (2014), pp. 105-118
© IAEME: www.iaeme.com/ijcet.asp
Journal Impact Factor (2014): 8.5328 (Calculated by GISI)
www.jifactor.com
IJCET
© I A E M E
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
106
The second major issue is Inter domain handover. An intelligent differentiation between Inter
Domain and Intra Domain session handover plays a vital role and reduces the probability of video
freeze. The Point of Attachments (PoA) like AP and Access Routers (AR) are equipped with an
inbuilt buffer. An Intelligent use of those buffers leads to save workload on the video servers.
There is also a need to predict the session rate before the handover takes place, as mobility
speed and connectivity strength varies. A weak framework without the consideration of above issues
typically hampers the video continuity [4][5].
Our work is motivated by 03 objectives. 1] Let the best AP must be selected based on the
various handover decision parameters other than just RSSI 2] A successful differentiation between
Intra Domain and Inter Domain to achieve an optimized workload saving. 3] A proactive session rate
prediction before handover, so as to calculate the exact frame after handover.
In this paper, we design and investigate a framework based on video streaming application to
take care of the above objectives. There is an extensive research has taken place in wireless network
handover decision and execution [2][3][6][7][11]. However the above issues in the context of video
streaming has not been addressed yet. In an attempt to give full scope to the topic, we have
intentionally suppressed several other issues related to handover like Layer 2 and Layer 3 handover
latencies, protocols etc.
2. LITERATURE SURVEY
Based on the motivation given above, we present an extensive literature survey to put
forward the core issues for video streaming continuity in mobile environment.
Franc Kozamernik [1] have reviewed the basic concepts of media streaming over the mobile
internet particularly those associated with Internet Protocols (IP), server technologies and delivery
aspects of video.
Khatib Noaman Ashraf, Vidhate Amarsinh and Satish Devane [2] have presented the state of
the art analysis for mobility management protocols along with a comparative study of signaling delay
and handover latency.
Vidhate Amarsinh & Devane Satish [3] have stated that the growth of audio, video and
multimedia applications are hampered because of limitation in MIPv6 during handover, as MIPv6 is
not designed for continuous streaming. They state the limitations to support QoS parameters like
variable jitter, delays in addition to loss of packets for streaming video during handover. They have
tried to improve the latency in handover by modifying the signals related to handover, which has
resulted in reducing the signaling cost and latency.
Xiaohuan Yan et al. [4] have presented a comprehensive survey of the handover algorithms
designed to satisfy various requirements based on parameters. To offer a systematic comparison,
they have categorized the algorithms into four groups based on the main handover decision criterion
used. Also, they have evaluated tradeoffs between their complexity of implementation and efficiency
for various proposals.
Pollini et. al. [5] has suggested various approaches to take Handover (HO) decision as RSS
with threshold, RSS with threshold and hysteresis and future prediction of RSS. Inclusion of
Threshold and Hysteresis Margin reduces Unnecessary HOs, but still a wrong decision for HO may
drop the call due to increase in HO delay. Especially hysteresis margin avoids ping-ponging effect.
Prediction of the future RSS helps in reducing unnecessary HO as compared to threshold and
hysteresis methods. But still RSS alone is not sufficient to take decision.
Ali Safa Sadiq et al. [6] states that traditional (based on one metric Received Signal Strength
Indicator) predictions of handover decisions do not perform well. It is a pressing need to develop an
intelligent approach to predict the handover decision process, thus yielding seamless handovers.
They have proposed a Mobility and Signal Strength-Aware Handover Decision (MSSHD) approach
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
107
to predict the handover decision in wireless networks. The Received Signal Strength Indicator and
the direction of Mobile Node parameters are considered as inputs to the fuzzy inference system to
predict the handover decision, and hence switching to the best preferable access point, resulting in
reduced handover latency as well as the wireless access media delay.
Ravindra Agarwal and Amarsinh Vidhate [7] have stated that, many times the handover
decision is taken not only based on Received signal strength (RSS) but also other factors like
available bandwidth, total required bandwidth, expected delay, packet jitter, packet loss and cost per
byte. They have analyzed various handover decision techniques to understand the handover
initialization reasons and proper handover trigger.
P. Bellavista [8] et al. have considered that, with signal strength, other factors like handover
awareness, QoS awareness and location awareness are also some of the crucial factors to be
considered for handover decision. But more parameters introduce more delay, which may not be very
suitable for applications like video streaming.
Sanjay Dhar Roy et al. [9] have proposed received signal strength (RSS) based strategy for
handover in heterogeneous networks which considers RSS and bandwidth. Further these strategies
have been modified by considering averaging of RSS. For comparison purposes, the performance of
the VHO algorithm also considers hysteresis and dwell timer.
Wireless Bandwidth estimation tool (WBest) [10] was designed for fast, non-intrusive and
accurate estimation of available bandwidth in IEEE 802.11 networks.
Anagha Raich and Vidhate Amarsinh [11] have presented various parameters for the
selection of the best path based on signal strength, RTT and packet losses.
Amarsinh Vidhate and Satish Devane [12] have stated that the applications like video and
audio streaming does not sustain continuous data flow due to handover as it disconnects the flow
during handover over the mobile IPv6 networks. They have introduced a novel methodology on
session handover by using session rate prediction to enable video session continuity without video
freeze for mobile wireless networks. The results are presented to differentiate latency and workload
between IntraDomain and InterDomain session handover to facilitate seamless streaming over the
mobile networks.
Hua, K.A. et al. [13] have proposed a Dynamic Stream Merging (DSM) technique for
efficient video-on-demand services to mobile users on wireless mesh networks at the edges. DSM is
a new communication paradigm, in which multicast topologies are created incrementally through
dynamic merging of server streams at the mesh nodes. This is accomplished without the knowledge
of the server.
Sundaram, V. et al. [14] have proposed a light weight, fast and distributed scheme that uses a
session rate prediction technique and a Dynamic Relationship Tree (DRT) for wireless sensor
networks. The proposed session handover scheme promotes workload sharing and supports mobile
clients moving at speeds as high as 250 miles per hour.
Gunjgur, P.N. and Vidhate A.V. [15] have stated that the continuity of multimedia
applications may hamper due to improper session rates during transmission. In this study they have
surveyed various papers for session rate prediction of streaming media using network traffic
prediction methods. Their novel method as bandwidth estimation is carried out for the wireless
network which plays a significant role in predicting the session rates. Their proposed session rate
expression helps to understand the significance of predicting the session rate for streaming media in
mobile wireless network and studied various proposals in this regard alongwith the state-of-the-art
analytical analysis.
From the literature review, we list major issues like 1] There is an urgent need of a novel HO
prediction and decision algorithm 2] An appropriate differentiation between IntraDomain and
InterDomain for session continuity 3] A well prediction of a session rate. We present the framework
based on the above problems.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
108
3. VIDEO STREAMING FRAMEWORK
There is a prominent goal of better Quality of Experience (QoE) for the mobile user when the
video is being watched. A small freeze during the play keeps the user frustrated for whatever the
number of reasons. As an end user, he/she expects that the video should be played without interrupt
and with better quality of service. From the designer point of view, we divide the framework into 03
phases as
3.1. Proactive Handover prediction
3.2. IntraDomain / InterDomain Session Handover
3.3. Session Rate Prediction
3.1. Proactive Handover Prediction A handover decision plays a vital role for the continuance of the data flow during mobility
period. An intelligent selection of network parameters add to its right selection of PoA which leads
towards a valuable contribution for the applications like video streaming and its video continuity
during handover period. There are various network parameters identified which are listed below.
• Received Signal Strength Indicator (RSSI), Round Trip Time (RTT), Available Bandwidth,
Packet loss ratio, Peak Signal to Noise Ratio (PSNR), Network Connection Time (Lifetime),
Power Consumption, Monetary Cost, User preferences, Location and velocity of MN, Quality
of Service (QoS) and QoE etc.
• Other parameters: More coverage, lesser latency, lesser CIR (Carrier-to-Interferences Ratio),
lesser SIR (Signal-to-Interferences Ratio), lesser BER (Bit Error Rate), etc.), more security
level, proper QoS class based on the applications , are some of the parameters to be considered
for fulfilling the best selection of PoA.
Figure 1: Various parameters for handover decision
3.1.1. Handover Prediction In order to prevent the service interruptions while serving the video stream to MNs, the
prediction of handover is imperative. It allows performing the required service management
operations in advance with respect to the actual communication-level client handover. The main aim
of handover prediction depends on proactively moving the client to the exact predicted next wireless
location. This is done with the required buffered data for enabling the service continuity. In this
phase, the Handover prediction is performed proactively based on data link monitoring using the
Received Signal Strength Indication (RSSI) of clients. The handover decision is triggered when the
roaming client encounters a change in its link state.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
109
3.1.2. Proactive handover decision (Traditional) The proactive method of triggering the handover takes place in prior to the absolute loss of
original clients signal. i.e. if the RSSI of the new client crosses the threshold (Preset) value, then the
handover gets triggered [5].
We show the above proactive handover trigger using the following conditions.
Let RAP be the RSSI value of visible access point (AP)
Let RC be the RSSI value of current AP.
Let Thh be the threshold value of hysteresis handover.
If RAP > (RC + Thh)
Then
Handover is triggered
End if
The threshold value Thh helps in preventing heavy bouncing effects also called ping-pong
effect. If more predictions are simultaneously enabled, then the proactive handover considers AP
with strongest RSSI value, is considered.
3.1.3. (Improved) Proactive Handover Decision Algorithm
The sufficient indication of RSSI from the nearby APs, above the threshold gives a trigger to
handover. But it is not adequate to handle applications like video streaming due to the variable nature
of mobile networks in terms of bandwidth, delay, jitter and interference.
We consider four parameters for the AP selection function, viz RSSI, RTT, Packet Loss rate,
available bandwidth and PSNR.
In case of RSSI, it is to the fact that signal strength does not fade in a linear manner, but
inversely as the square of the distance. This means that if you are at a particular distance from an
access point and you measure the signal level, and then move twice as far away, the signal will
decrease by a factor of four. You move by 2x and the signal decreases by 1/4x; hence, the “inverse
square law”. The IEEE 802.11 standard defines a mechanism by which RF energy is to be measured
by the circuitry on a wireless NIC. This numeric value is an integer with an allowable range of 0-255
(a 1-byte value) called the Receive Signal Strength Indicator (RSSI).
RTT is the length of time it takes for a signal to be sent plus the length of time it takes for an
acknowledgment of that signal to be received. This time delay therefore consists of the propagation
times between the sender and the receiver.
PSNR is the ratio between the maximum possible power of a signal and the power of
corrupting noise that affects the fidelity of its representation. Because many signals have a very wide
dynamic range, PSNR is usually expressed in terms of the logarithmic decibel scale.
Packet Loss Ratio is the ratio between number of packets sent Vs number of packets received
and checked at different time intervals. The lesser the ratio, better is the performance. We restrict our
discussion to these four parameters only and consider for the decision making.
Available bandwidth is one of the important parameters as the target application is video
streaming, which is bandwidth consuming. Few statistics are presented here for the verifications. The
bandwidth for a video streamed at 300 kpbs watched by 50 clients for 01 hr. would be calculated as
300/8(conversion factor) =37.5, 37.5 X 60 (Seconds) = 2250, 2250 X 60 (Minutes) = 135000,
135000/ 1000 kb=135 mb/hr [16].
An Improved proactive method of triggering the handover takes place in prior to the absolute
loss of original clients signal. i.e. if the RSSI of the new client crosses the original value, and having
minimal RTT value , minimal packet loss and sufficient bandwidth available , then the handover gets
triggered. These parameters make sure about not only the next point of attachment but also the best
point of attachment for the range applications like audio/video streaming.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
110
We demonstrate the above proactive handover trigger function using the following
conditions. Let RAP be the RSSI value of visible Access Point (AP), RC be the RSSI value of current
AP, Thh be the threshold value of hysteresis handover, RTTmin is the minimum RTT value returned by
the RTTmin function, λmin is the minimum packet loss value returned by the λmin function.
The AP selection function is defined as
( , )ni F i
i
f W N= ∑ (1)
Initially the client requests the network and bandwidth availability information along its
projected route or probable route. The route is recognized in prior. The lookup service then replies
with the available network’s position, RTTs from various APs in the proximity, Received signal
strength of various APs and a sequence of bandwidth samples from the lookup server. This is
performed using crowd-sourcing technique [12]. In this method, the MN gathers the measurements
made at various geographical locations of the commutation route. By firing a query to the lookup
database server, the MN can able to get the information. This information is returned to the lookup
service server and stored in the look up data base for future use. Every AP is going to do so that the
information spreads to other APs.
New AP Selection algorithm
If RAP > (RC + Thh) & & RTTmin && λmin Then
Handover is triggered
min min max max min min_ * * ( * ) *available availablenew AP b b RTT RTT RSSI RSSI hhf W N W N W N T W Nλ λ= + + + +
(2)
Select the Best suitable AP/BS and start HO association/ Reassociation
End if
Else
Handover is triggered
_new AP c hhf R T= +
(3)
Select the suitable AP/BS and start HO association/ Reassociation
End
End.
Figure 2: New AP Selection Function
The normalized values are given as
m a x
'
m a x
kR S S I
R S S I R S S IN
R S S I
−=
m i n
m a x m i na v a i l a b l eb
b bN
b b
−=
− (4)
m i n
m i n
m a x m i n
R T T
R T T R T TN
R T T R T T
−=
− m i n
m i n
m a x m i n
N λ
λ λ
λ λ
−=
−
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
111
Where the sum all weights is 1.
min max min1
avaulableb RTT RSSIW W W Wλ+ + + =
(5)
The common weights are considered as,
availablebW = 10 %,
minRTTW =10%,
maxRSSIW =70% & min
Wλ =10%
The parameters in table 1 include network ID, time of taking values, global positioning
system (GPS) coordinates, bandwidth, round-trip time and packet loss rate. This information is
adequate, if client mobility is random. This technique helps in selecting the available AP with the
most capacity for handover. These parameters are required to calculate strict session rate during
handover so that the quality of the video should not be reduced.
Table 1: Lookup Database
3.2. Intra Domain and Inter Domain Session Handover with session rate prediction.
Figure 3 demonstrates the intra and inter-domain handover. N1 and N2 represent the networks.
Mobile Node (MN) after identifying the foreign node performs inter or intra-domain handover based
on the node availability within or outside the network.
As discussed, a hierarchical tree is used to represent parent child relationship between various
hierarchies. Based on the topology, the buffering location will move from child to parent and vice
versa.
Figure 3: Intra-domain and Inter-domain Session Handover
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
112
The advantages are multiple like, it will be able to multicast streaming tree based on parent-
child relationship, whenever there is a new request for a video stream. It will detect redundant
streams and merge using intelligent buffering, before sending them downstream. This way it can
save more bandwidth and relax the stress at the gateway in terms of signaling cost.
We use hierarchical tree concept for the parent child relationship between the nodes. The
objective behind this concept is that, to promote the workload sharing among the nodes and relax the
stress on the gateway further. This automatically reduces the signaling cost and promotes the
localized buffering.
There are various issues and challenges to provide seamless session handover when the
networks are wireless. The issues are as bandwidth and bandwidth estimation is variable, high packet
loss, dynamic adaptation, variable jitter and jitter delay, interference and error rate is high. Due to the
above issues, the QoS, QoE degrades and seamless session handover becomes a challenge.
There is a need of data localization at the nearby point of attachment so that session must be
continued without repeating the frames and reduce the load on the gateway or the server further. We
present the topology for video streaming in the below figure. We are targeting video applications like
video on demand where one video can be demanded by many clients simultaneously on the mobile
device.
Figure 4: Inter Domain and Intra Domain Handover
R1, R2 are the routers, AR1, AR2 and AR3 are the Access Routers (AR),BS stands for Base
stations and MN is a mobile node.
Video streaming requires many sessions and session comprises of many frames. To
seamlessly transfer video sessions, a typical session rate prediction is required. It’s a proactive
session rate prediction which calculates typical sessions post handover to enable session continuity.
Our attempt is to save sessions and frames transfer from gateway or video server every time, when
the demanded sessions are same and simultaneous. To enable a proper session rate prediction, a
bandwidth estimation ,frame size estimation and frame rate estimation are the vital parameters. In
wireless enviroments, these parameters are variable.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
113
3.2.1. IntraDomain Session Handover algorithm
Let FNi and HNi be foreign and home node respectively.
The steps involved in the intra domain handover are as follows
1) Each MN analyzes the FNi based on its Function fnew_AP. When it identifies a suitable FNi with the
help of fnew_AP, it initiates the handover process.
MN → REQHO _ HNi
The client sends a request message (HO_REQ) to HNi requesting handover with FNi.
HO_REQ includes the clients existing video frame sequence ID.
2) HNi upon receiving the request verifies its node cache which results in the following two
solutions.
a. If HNi contains the respective FNi in its cache, then it forwards HO_REQ to FNi. then FNi
acknowledges HNi by sending a reply message (HO_REP)
HNi → REQHO _ FNi
HNi ← REPHO _ FNi
b. If HNi does not contain FNi in its cache, then it forwards the request to parent node. This
continued until FNi receives HO_REQ
HNi → REQHO _ parent node
3) In prior to performing step 2 a), HNi computes the most probable video frame sequence using
session rate prediction technique (explained in subsection d) which has to be delivered subsequent
to handover process.
4) Now, FNi initiates early buffering using the computed sequence and upholds client activities after
connection.
Basically the above concept promotes two types of session handover, 1] InterDomain HO &
2] Intra Domain HO. In the figure 4, when the MN is moving from BS1 to BS2, it is termed as
IntraDomain HO and it promotes the buffering at the AR1 till it connects to BS2. But in case when
MN is connected to BS2 and now connecting to BS3 then it would be inter domain handover and the
contents has to be buffered at R2, instead of AR1 and AR2. We would like to shift the location of
buffer from child to parent, based on type of Handover.
In Intra Domain case, BS1 is the HN and BS2 is the FN, as they are from same domain, MN
sends HO_REQ to HN and as FN is in its cache, HN forwards the request to FN. Subsequently FN
sends HO-REP. Due to session rate prediction; the typical session after handover is predicted. And
the handover takes place. In case of redundant flows, the contents during handover are buffered at
AR1.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
114
3.2.2. Inter-Domain Session Handover Algorithm
The steps involved in inter-domain handover are as follows
1) Each MN analyzes the FNi based on its Function fnew_AP. When it identifies a suitable FNi with
maximum RSSI when compared to HNi, it initiates the handover process.
Client → REQHO _ HNi
The client sends a request message (HO_REQ) to HNi requesting handover with FNi.
HO_REQ includes the clients existing video frame sequence ID.
2) HNi upon receiving the HO_REQ broadcasts the message after parent node confirmation. i.e.
In order to prevent HNi from unnecessary broadcasts, the parent of HNi following the reception
of HO_REQ message wait until an update is received from its child for time t. Then it re-
broadcasts the message after time expiry.
3) Step 2 allows any FNi in different clients control also to detect request message.
4) FNi upon receiving the request responds with the reply message to HNi. Thus the handover is
similar to intra-domain handover.
In Inter-Domain Handover, MN is trying to connect from BS2 to BS3 due to its mobility.
Here BS2 is HN and BS3 is FN. Now MN sends HO-REQ to HN but it doesn’t have FN in its cache
as it lies to different domain. So child HN sends the request to parent AR1 and if it is not there also,
it forwards to its parent. Due to the child parent and its decision to place buffer for redundant
streams, it would save lot of workload on the server, as many contents are coming from the regional
buffers.
3.3. Session Rate Prediction
Handover is a typical challenge when the environment is variable and the application is like
video streaming. Sessions are formed by many frames and we need to estimate session rate, frame
rate and frame size to estimate session rate. It must happen proactively. We see the issues & the
challenges faced by mobile wireless network. Session Handover seem to be the significant
challenging problem over here. We need to estimate the fluctuating bandwidth [6].
In this phase, the home node computes the most probable sequence of video stream that needs
to be delivered for the MN during handover.
To enable a proper session rate prediction, a bandwidth estimation ,frame size estimation and
frame rate estimation are the vital parameters. In wireless enviroments, these parameters are variable.
Let Rv be the rate at which client receives the video streams.
Let RTTo be the optimal round trip time
Let Rseq be the sequence number of video stream segment from which the foreign node has to deliver
the MN to ensure seamless video streaming.
Rv = )(
)(
ic
ic tt
αα −
− (6)
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
115
RTTo = ∑RTT*η (7)
Rseq = c
v
o
R
RTTα+ (8)
where ∑RTT = sum of the RTTs computed from each path along the route of request from the
home node to the foreign node
tc = current time , ti = the time when the client initiates handover
cα = current video sequence number at tc , iα = current video sequence number at ti
Eq (8) reveals that Rseq is dependent on RTTo. The selection of exact RTTo value is based on the
metricη .
/tT R T Tη ≈ ∑ (9)
Where Tt is the total time, the home node supports the client until re-association.
Tt in Eq (9) depends on transmission range ( txα ) and speed of the client (ν ).
i.e. If (ν = slow) || ( txα = large) Then
Tt increases
End if
Thus, η varies withν , txα and node count in the network.
4. SIMULATION RESULTS AND DISCUSSION
We have used widely known network simulator NS2 [17] for our simulation. For the
simulation purpose, we have used UDP Continuous Bit Rate (CBR) traffic.
Figure 5: crowd sourcing based lookup outcome (1st try and 2nd Try)
The screen shot in figure 5 show the result of crowd sourcing where data rate is 400 kpbs,
various RTTs from different APs spread across the cellular proximity, total bandwidth available and
the total packets lost are considered. Due to intelligent AP selection function, there is a tremendous
difference between the packet losses as shown in figure 3 and 4. There is a difference between the
bandwidth also. Due to an intelligent PoA selection, the packet losses are reduced from 1191 bytes to
101 bytes.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
116
Figure 6: Various RTT values Figure 7: RSSI and PSNR without and
with proposal
In figure 6, we want to show that minimal RTT as a parameter also plays important role in
best AP selection. In figure 7, our proposal allows the packets lost recovery, reducing the frame loss’
rate after the handover, so MN receives the frames that were not received during the connection
discontinuation. This increases the average PSNR of the video forwarded, monitored in the
transmission each 5 seconds.
We also present an empirical example to proof the work. The value of tc = 300 ms,
ti = 200ms, αc = 100, αi = 80, η = 0.2, ∑ RTT = 61.1, Rv = 5, RTTo = 12.22, Rseq = 102
The example shows that, post session handover continues at sequence packet number 102.
Few more values are shown in table 2.
Table 2 : Various results of sequence frames
Figure 8: Total time of association is variable
tc ti cα iα ∑RTT Rv RTTo Rseq
300 200 100 80 61.1 5 12.22 102
312 208 110 89 41.1 5 8.22 112
320 217 120 91 78.7 4 15.74 124
780 560 210 170 56.6 6 11.32 221
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
117
Figure 8 is a typical representation of total time of association which is based on various
RTTs, which ultimately is based on velocity and transmission range of MN.
Figure 9: Handover Latency InterDomain / Figure 10: Saving of workload on
IntraDomain Session Handover Gateway/ Server
Figure 9 shows a lot of handover latency reduction when it comes to IntraDomain session
handover as well as it reduces lot of workload on the server/gateway, by detecting redundant frames
and buffering it for avoiding retransmission from the server. Figure 10 is a graph between latency in
seconds Vs no. of sessions handover. It is clearly observed that there is almost 80% saving of
handover latency between IntraDomain and InterDomain session handover, if it is detected
proactively. It also increases with number of session handovers if the video sessions are redundant.
This scheme is more suitable for the similar video which is being demanded by many users with a
small time difference.
5. CONCLUSION AND FUTURE RESEARCH
Handover plays a vital role for data forwarding with minor packet loss. After the experiments
being conducted in NS-2, we concluded that, the use of traditional mobility protocols face difficulties
during video session continuity due to longer handover latency and non buffering support during
connection discontinuity. Our novel framework not only selects the best PoA with a proactive
handover prediction, but also reduces latency and packet losses during connection discontinuity,
without hampering the video quality. An efficient session handover during mobility alongwith
IntraDomain / InterDomain differentiation enables video continuity and minimize video freeze also.
We have specifically suggested Intra Domain Session HO and Inter Domain Session HO, along with
session rate prediction which proves the better improvement and saving of workload on the server
that suit to the scalable video applications.
REFERENCES
1. Franc Kozamernik, “Media Streaming over the Internet - an overview of delivery
technologies,” EBU Technical Review,Oct-2002.
2. Khatib Noaman Ashraf, Vidhate Amarsinh, and Devane Satish, “Survey and Analysis of
Mobility Management Protocols for Handover in Wireless Network,” 3rd IEEE International
Advance Computing Conference (IACC), pp. 413-420, February 2013.
3. Vidhate Amarsinh and Satish Devane, “Improved Latency Handover in Fast MIPv6 for
Streaming Video,” Springer-Verlag Berlin Heidelberg 2013, ICAC3 2013, CCIS 361, pp. 381–
392, 2013.
International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 5, Issue 4, April (2014), pp. 105-118 © IAEME
118
4. Xiaohuan Yan, Y. Ahmet ekerciolu, Sathya Narayanan, “A survey of vertical Handover
decision algorithms in Fourth Generation heterogeneous wireless networks,” computer
Networks, Volume 54, Issue 11, Pages 1848-1863, 2 August 2010.
5. Pollini,G.P., “Trends in handover design,” IEEE Communications Magazine,34(3)82-90,Mar
1996.
6. Ali Safa Sadiq, Kamalrulnizam Abu Bakar, Kayhan Zrar Ghafoor, and Alberto J. Gonzalez,
“Mobility and Signal Strength- Aware Handover Decision in Mobile IPv6 based Wireless
LAN,” Proceedings of the international multiconference of Engineers and Computer Scientists,
IMECS 2011, March 16-18, 2011, Hong Kong.
7. Ravindra R.Agrawal, Amarsinh Vidhate, “Optimized Heterogeneous Wireless Network with
Scoring Methods,” International Conference & Workshop on Recent Trends in Technology,
(TCET) 2012 Proceedings published in International Journal of Computer Applications
(IJCA), Vol. icwet, number 8, pages-28-32, March 2012, Published by Foundation of Computer
Science, New York, USA.
8. P. Bellavista, M. Cinque, D. Cotroneo, and L. Foschini, “Self-Adaptive Handover
Management for Mobile Streaming Continuity,” IEEE transactions on network and service
management, Vol. 6, No. 2, pp. 80-94 ,JUNE 2009.
9. Roy, S.D.; Anup, S., “Received signal strength based vertical handoff algorithm in 3G cellular
network, Signal Processing,” IEEE International Conference on Communication and
Computing (ICSPCC), 2012,vol., no., pp.326,330, 12-15 Aug. 2012.
10. Mingzhe Li, Claypool, Mark Kinicki, Robert, “WBest: A bandwidth estimation tool for IEEE
802.11 wireless networks,” Local Computer Networks, 2008. LCN 2008. 33rd IEEE
Conference on, vol., no., pp.374-381, 14-17 Oct. 2008.
11. Anagha R. Raich, Amarsinh Vidhate, “Best Path Selection Using Location Aware Modified
AODV,” International Journal of Advanced Computer Research (ISSN (print): 2249-7277,
ISSN (online): 2277-7970) Volume-3 Number-3 Issue-11 September-2013.
12. Vidhate Amarsinh and Satish Devane, “Improved session handover in mobile wireless
networks for video streaming,” IEEE International Advance Computing Conference (IACC),
2014, pp 411-415, Feb., 2014.
13. Hua, K.A. and Fei Xie, “A Dynamic Stream Merging Technique for Video-on-Demand
Services over Wireless Mesh Access Networks,” 7th Annual IEEE Communications Society
Conference on Sensor Mesh and Ad Hoc Communications and Networks (SECON), pp.1-9,
June 2010.
14. Sundaram, V. and Hua, K.A, “Seamless Video Streaming: A Light Weight Session Handoff
Scheme for Dynamic Stream Merging Based Wireless Mesh Networks,” IEEE International
Conference on Multimedia and Expo Workshops (ICMEW), pp.107-112, July 2012.
15. Gunjgur, P.N. and Vidhate, A.V., “Session rate prediction for multimedia streaming,” IEEE
International Advance Computing Conference (IACC), 2014,pp.348-353, Feb., 2014
16. Streaming Bandwidth Calculator, http://www.netromedia.com/pricing/calculators.aspx.
17. Kevin Fall, Kannan Varadhan, “The ns Manual,” The VINT Project, A Collaboration between
researchers at UC Berkeley, LBL, USC/ISI, and Xerox PARC, May 2010.
18. Dheyaa Jasim Kadhim and Sanaa Shaker Abed, “Performance and Handoff Evaluation of
Heterogeneous Wireless Networks (Hwns) using OPNET Simulator”, International Journal of
Electronics and Communication Engineering &Technology (IJECET), Volume 4, Issue 2,
2013, pp. 477 - 496, ISSN Print: 0976- 6464, ISSN Online: 0976 –6472.
19. S. Sri Gowri, K.Venkata Satya Anvesh and K. Sri Pavan Kumar, “Performance Evaluation of
Handoff Parameters in Mobile Systems”, International Journal of Electronics and
Communication Engineering &Technology (IJECET), Volume 3, Issue 2, 2012, pp. 164 - 170,
ISSN Print: 0976- 6464, ISSN Online: 0976 –6472.