1. Fuzzy logic in Satellite
communication Systems
Samad Razaghzadeh Shabestari
Telecommunication Research Lab.
3/31/2010
2. Outline
Satellite Communication Systems
SLAs (System Level Agreements)
QoS (Quality of Service)
Problem Space
Fuzzy Logic Controller
Motivation
RED-FL Algorithm
Control problems
Conclusion
3. Introduction
Satellite networks play a significant and major
role in numerous fields.
Four major growth in telecommunications
marketing areas are:
1. messaging and navigation services
(wireless and satellite),
2. mobility services (wireless and satellite),
3. video delivery services (cable and satellite),
4. interactive multimedia services (fibre/cable,
satellite).
Vast marketing opportunity with coupling
satellite characteristics and terrestrial networks
1. Wide area coverage (extensive geographic
reach)
2. Multicast capabilities
4. The Architecture of Satellite-Based IP
Network
Satellite relays signals among ground
terminals and between the hub
gateway(HG) and ground
terminals(GT).
All GTs communicate only with the
HG, and vice versa.
HG dynamically allocates Upload and
Download link Bws to GTs according
to SLAs and conditions such as the
systems resources availability and
weather conditions.
HG executes QoS control link
congestion avoidance and system
operation, administration,
maintenance and provisioning.
GTs allocate BW to users connected to
it within the available link capacity
assigned to it by HG.
LANs, servers and workstations may
connect to the systems through HG
and GTs.
6. Major drawbacks of satellite communications
High propagation delay due to their altitude:
Impact on the Quality of Service (QoS) to the end users
SNR(Signal to Noise Ratio):
Affect the data transmission done through the satellite
Bandwidths associated with satellite systems :
Be a scare resource compared to terrestrial systems
Dynamic weather conditions: may affect severely overall performance of
satellite communications and result channel fading or degradation.
Channel fading or degradation most often leads to network congestion.
Congestion control : significant issue in satellite networks due to the
inherent large channel latency (delay) or high bandwidth-delay product.
the network congestion control :critical in broadband satellite-IP based
networks, for committing to the Service-Level-Agreement (SLA) with
regard to Quality-Of-Service (QoS).
7. Theoretical Background-SLAs
SLAs stands for Service Level Agreements
To provide predictable and reliable services, service providers of
communication services negotiate Service Level Agreement with the
customers.
The purpose of SLAs is to identify the shared objectives between the service
providers and the customers
Example:
- Service Provider – “We agree to provide you this particular level of service
based on agreed-to set of guidelines”
- Customer – “We agree to abide by your guidelines in anticipation that you
provide us this agreed level of service”
The most important agreements between service provider and customer
(User) :
1. Availability of the promised level of service to the user
2. Performance measures of various components of the user‟s workload
3. Bounds of guaranteed performance and availability
4. Cost of the service provided to the customer
More specific parts of an SLA are service level specification (SLS), traffic
conditioning agreement (TCA), and traffic conditioning specification (TCS).
8. Theoretical Background-QoS
Service providers define the “Quality of Service” (QoS)
a customer can expect in the Service Level
Agreements.
Quality of Service is defined by IETF (Internet
Engineering Task Force ) as “ A set of service
requirements to be met by the network while
transporting a flow.
Quality of Service is one of the most important
performance measures of any communication service.
9. QoS (Continued)
Qos in packet switched networks is expressed by the following set
of parameters to meet the requirements defined in SLAs:
1. Packet loss rate:
Occur when there is congestion in the transmission channel and these
channels are forced to drop the packets since it is almost impossible
to transfer them any forward to their destination nodes.
2. Jitter (Variation of latency):
Jitter variations in the IP packet transfer delay.
3. Queuing delay:
Experienced by packets while passing through the network.
Latency or Queuing delay is very low whenever the networks are
free of congestion. However; when congestion starts to build up in
the channel, the packets are forced to wait so latency can be very
high.
10. QoS (Continued)
Three major categories of approaches to
ensure end-to-end QoS:
1. Integrated Services(IntServ)
Suitable for small scale work and not very popular.
2. Differentiated Services(DiffServ)
difficult to predict end-to-end behaviour.
impossible to sell different classes of end-to-end connectivity to end users, as one
provider's highest-priority packet may be another„s low priority.
only works if the boundary hosts honour the policy agreed upon.
3. Active Queue Management (AQM)
The first two are based on flows
distinction but AQM can fairly regulate
network traffic w/o flows traffic.
11. QoS-AQM
A packet dropping/making mechanism for
router queue management.
Targets to reduce the average queue length
and thus decrease the packet delay.
Common in best effort networks where all
streams have the same network access right.
A technique of preventing congestion in packet-
switched networks.
12. RED(Random Early Detection)
One of the most well-known AQMs
An efficient queue management and congestion
avoidance algorithm by controlling the average queue
length in a reasonable range and dropping packets
based on statistical probabilities.
Proven to be stable [4] but its performance is sensitive
to tune the parameters dynamically
Sets the min.Threshold(qmin ) And max.Threshold(qmax )
for queues and handle newly arrived packets according
the following rules:
13. RED (continued)
IF qmin ≤ queue length ≤ qmax , THEN RED
drops newly arrived packets will be
dicarded with a drop probability
IF queue length>qmax , THEN RED drops
all newly arrived packets.
IF queue length<qmin , THEN no packet
dropping.
14. Problem Space
Designing rules to maintain the SLAs for sat. Based Networks are
more complex than non-sat. networks since predictability of
service in sat networks and performance of QoS are effected by
external factors like weather condition.
Thus channel fades and communication between HB and GT
cannot be maintained due to the incapability of the ground
terminal to properly predict and react to the rain fade situations on
its own.
In addition, the hub does not have any intelligence to assist the
ground terminals during rain fade situations.
It is extremely difficult to clearly engineer SLA metrics such as
throughput, delay and jitter when there are uncertainties.
Therefore, there is an inherent fuzziness in the Sat. Based network
congestion control .
15. Motivation
Some of the very well known advantages of fuzzy logic control:
1. Easy handling of inherent nonlinearities
2. Easy handling of multiple input signals
3. Less dependence on availability of precise mathematical models
Fuzzy control techniques are suitable to apply to the congestion
control problems in satellite networks due to:
1. The inherent difficulty in obtaining precise mathematical models
through conventional analytical methods.
1. Possibility to obtain simple linguistic rule curves for congestion
control.
16. Intelligent systems-Fuzzy Logic
The general fuzzy rule is in the form of
If A is x Then B is y
„x‟ is a fuzzy set of the input variable A and „y‟ is
the fuzzy set of the output variable B.
„max-min‟ operators are used for implication and
composition respectively in Mamdani inference
systems.
Triangular membership functions were used as
proven to be extremely effective in networks [3]
17. Design of the Fuzzy logic controller
(Fuzzy Inference system)
-Fuzzifier: Crisp values are
transformed for linguistic terms of
fuzzy sets.
- Knowledge Base (Fuzzy Rule
Base) :IF-THEN rules to
characterize the goals and
policies of the problem
considered.
-Fuzzy Inference Engine :
• Provide the decision-making
logic to the system.
• Respond to fuzzy input
provided by Fuzzifier and also
possibly previous inference
from the rule-base itself
•Identify one or more possible
control action
-Defuzzifier : Central method
18. Functional block diagram for the high-level
system design- case study
Hub Correlator
− Correlate relevant precipitation data
with fuzzy inference engine to
predict service degradation within
specific timeframe
− Initiate appropriate proactive or
reactive required actions
Link Detector
− Assign traffic on forward and return
link timeslots
− Adapted to fade & congestion
QoS Arbitrator
− Active queue management to discard
packets/flows and notify impending
network congestion
Traffic Shaper
− Connection admission control
− Packet conditioning and scheduling
The motivation of the design is to achieve high
performance communications and congestion
adaptation to impacts of dynamically weather
change, on system performance (SLA QoS), e.g.,
channel rain-fade.
19. RED-Fuzzy logic
A novel Fuzzy-based packet dropping algorithm
In [1] fuzzy inference systems applied to RED
for parameter tuning, based on the inherent
fuzziness in network congestion.
Making satellite-IP based network more adaptive
to dynamic weather condition change, e.g.,
channel rain-fade(rain attenuation)
In [2] tried to control complexity and non-
linearity inherent of satellite systems by coupling
FL and RED to prevent impending congestion
from happening and to maintain QoS.
20. Control problem scenarios [1]:
Three different scenarios (control problem)
for weather adaption are considered:
1. Adjustment of the Bandwidth
2. Adjustment of the minimum queue length
qmin
3. Adjustment of the maximum queue length
qmax
The current weather conditions, allocated
bandwidth and the queue length are the
input parameters to the above control
problems.
21. The universe discourse for input and output
parameters
1) Bandwidth range (input in scenario #1) is selected to vary between 16
and 128 kbps per user. The packet size is fixed at 512 bytes.
2) The minimum queue length qmin (input in scenario #2) is varied
between 4 and 16 packets.
3) The maximum queue length qmax (input in scenario #3) is varied
between 8 and 20 packets.
4) Current weather condition (input in all cases) is based on the total
precipitation rate.
5) The bandwidth adjustment (output in scenario #1) is done between 0
and 48 kbps per user.
6) The output parameters in scenarios #2 and #3 are adjusted to fall
within the ranges already mentioned under 2) and 3).
22. Fuzzy sets associated with input parameters
The synthesis of the membership functions depend on the choice of
thresholds of the weather condition parameters, e.g., the amount of
precipitation.
Change in weather conditions (mm o rain drops/hour)
23. Scenario#1 (bandwidth ranges )
Bandwidth range (input in scenario #1)
is selected to vary between 16 and
128 kbps per user. The packet size
is fixed at 512 bytes.
Scenario # 1 [Bandwidth Adjustment]
− IF (the current allocated bandwidth is low) AND (weather becomes worse)
THEN {heavily increase the bandwidth to commit the SLA QoS};
− IF (the current allocated bandwidth is medium) AND (weather becomes
worse) THEN {slightly increase bandwidth to commit the SLA QoS};
− IF (the current allocated bandwidth is high) AND (weather becomes worse)
THEN {do not change bandwidth};
………..
27. Fuzzy sets associated with output parameters
Defuzzification converts the output fuzzy result into a crisp and a singular output.
Center of area method is used as the defuzzification method.
28. Control Problem-scenarios [2]
eliminating the fixed threshold value, specifically the
qmax value and the drop probability value from RED
and replace with dynamic network state dependant
values calculated by the FI engine.
Inference engine keep dynamically calculating qmax
and drop prob. Behaviour based on 2 following
network state inputs:
1. Link Capacity of the channel between the gateway
and the destination (Relay Satellite )
2. The number of dynamic traffic sources accessing the
gateway :
It has a direct impact on changing queue sizes and
provides a direct measure of the future queue state.
29. Input –link capacity:
1.8 MB assumed to belong to fuzzy set High
0.18 MB assumed to belong to fuzzy set Med
0.16 MB assumed to belong to fuzzy set Low
30. Input – Number of traffic sources:
90 Sources assumed to belong to fuzzy set High
60 Sources assumed to belong to fuzzy set Med
20 Sources assumed to belong to fuzzy set Low
32. Rule -1 is read by the inference engine as
“IF Link Capacity is low AND Number of Traffic Sources is low
THEN
qmax is medium and p-drop is medium.”
33. Conclusion
Intelligent network congestion control mechanisms are
highly expected for Satellite IP-based networks due to
the inherent impacts of dynamical weather change on
system performance, e.g., channel rain-fading.
novel Fuzzy-based packet dropping algorithm may
satisfy the SLAs and QoS requirements under dynamic
weather conditions.
It needs to be tested in the real world field work and on
the real time satellite hub/terminal gateways.
Simulations have been done in single Hub-Terminal
connection .
Research on multiple terminal gateway connections
would be interested.
34. References:
[1]:Adaptive Congestion Control Under dynamic weather condition for
Wireless and Satellite Networks;Hongqing Zeng, Anand Srinivasan, Brian
Cheng, and Changcheng Huang; EION Inc., Carleton University;
ITC2007-PP.92-107
[2]:Congestion Control for Adaptive Satellite Communication systems
Using Intelligent Systems;Banupriya Vallamsundar, Jiesheng Zhu,Brian
Cheng; University of Waterloo; Carleton University; Signals, systems and
Electronic, 2007,ISEE07 International Symposium-pp.415-418
[3]: Design an active queue management algorithm based on fuzzy logic
decision.; Fan, Y., Ren, F., Lin, C.; Proceedings of IEEE ICCT 2003,
[4]: Steady state analysis of the RED gateway: stability, transient
behavior, and parameter setting, H. Ohsaki, M. Murata, and H. Miyahara,
IEICE Transactions on Communications, Vol. E85-B, Jan. 2002
36. Back up
The term “service” in the telecommunications:
A service in an IP based environment is defined by International
Telecommunications Unit (ITU) as “a service provided by the service plan to
an end user (e.g., a host [end system] or a network element) and which
utilizes the IP transfer capabilities and associated control and management
functions, for delivery of the user information specified by the service
level agreements” (ITU Recommendations, 2001).
The meaning of “quality” is very broad:
In telecommunications it is commonly used in assessing whether the service
satisfies the user‟s expectations. However, the evaluation depends on
various criteria related to the party rating the service. Customers assess it
on the basis of a personal impression and in comparison to their
expectations, while an engineer expresses quality in terms of technical
parameters.
This discrepancy may sometimes lead to misunderstandings between the
service provider and the customer.