-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
1/44
CETTM MTNL
Traffic Concept
&Traffic Engineering Planning
(Mod Id:GSSFTRF005)
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
2/44
CETTM MTNL
Traffic Concepts Technological Developments
Services Developments
QoS - Quality of Service
Traffic engineering principles Traffic characterization
Traffic engg. Task
Traffic demand charecterisation
GOS objectives Traffic controls and dimensioning
Contents
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
3/44
CETTM MTNL
Telephone traffic is originated by the individual
needs of different subscribers and therefore is
beyond the control of administration
Telephone traffic for a particular exchange follows apattern of activity in that area
Normally there is a peak in morning and
afternoon and a dip during lunch period
Any of the subscriber ( or every subscriber ) canoriginate a call at any given moment.
The duration of calls are not known.
Telephone Traffic
Traffic Concepts
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
4/44
CETTM MTNL
2500
2000
1500
1000
500
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Calls originated Time
per hour Variations in Calling Rate
On a network, load & traffic pattern varies during
the day with heavy traffic and low traffic durations
Traffic Pattern
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
5/44
CETTM MTNL
Busy Hour:The hour in which maximum trafficusually occurs in an exchange is known as busyHour
Busy Hour varies from day to day or over a
number of days Busy Hour Traffic is the average value
of maximum traffic in the busy hour
One hour period starting at the same
each day for which the Average TrafficVolume or Number of Call Attempts is
greatest over the days under
consideration
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
6/44
CETTM MTNL
One hour period starting at the same each day for
which the Average Traffic Volume or Number of CallAttempts is greatest over the days under
consideration
Busy Hour Call Attempt :No. of Call Attempts in a
busy hour Call Completion Ratio : Ratio of Number of Successful
Calls to Number of Offered Calls
Busy hour calling rate: No.of calls originated per
subscriber in the busy hour
Cost Constraints:Cost of the line and certain
individual equipment is independent of the volume of
traffic
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
7/44
CETTM MTNL
Exchange with 2000 Subscribers of averageBHCA 10,000 & CCR 60%. Calculate the BHCR.
Avg. BH Calls = BHCA x CCR
= 6000 CallsBHCR = Avg. BH Calls
Total No. of Subscribers = 3
Erlang:If one circuit is held continuously for onehour then the traffic carried by that circuit
amounts to one Erlang ( 1 Traffic Unit )
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
8/44
CETTM MTNL
Average traffic: Ratio of Sum of HoldingTimes to Period
A = S/T
If C = total number of calls during the period Tthen Avg. holding time t = S/C hrs per call
Therefore A = C * t/T
Traffic Intensity :Ratio of Period for which anequipment is occupied to total period ofobservation
Erlang traffic: Period of observation is
generally considered one hour
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
9/44
CETTM MTNL
A subscriber makes 3 calls of 3 min, 4 min and 2
min duration in 1 hour period.
Calculate subscriber traffic in Erlangs.
Solution:Traffic = Busy period /Observed period
= (3+4+2)/60
= 9/60
= 0.15 E
Example
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
10/44
CETTM MTNL
Traffic carried by network is usually lower than
traffic offered to the network The overload traffic is rejected and not carried by
network
The traffic rejected by network is the index of QoS
Grade of Service
GoS = Lost Traffic / Offered Traffic= (A-A0) / AA - Offered traffic, A0 - Carried Traffic
(A Ao) - Lost Traffic
Recommended GoS = 0.002
i.e. 2 out of 1000 calls allowed to be lost
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
11/44
CETTM MTNL
More general term than GoS includes other factors
like Quality of Speech, Error free transmission
capability, etc.
Loss system: Circuit switched networksWhen overloaded, a user call is blocked and userhas to make a retry
Delay System:Packet switched (store & forward)
networkDelay when extended beyond limits becomes a losssystem In a store & forward network if the queue
becomes full, then further requests have to be rejected
Quality of Service
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
12/44
CETTM MTNL
Parameters for Loss system :- Grade of Service
- Blocking Probability
Blocking Probability
Probability that all equipments in a system are
busy When all equipments are busy no further traffic
can be carried by the system and the further
arrival traffic is blocked
Blocking Model
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
13/44
CETTM MTNL
Queuing Model:Parameters for Delay System
Service Delays
Flow ControlTo prevent loss, queue of traffic is cleared to an
acceptable limit Flow Control technique used to
prevent loss of traffic
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
14/44
CETTM MTNL
Designing a cost effective network which
provides the required Quality of Service under
varied traffic conditions demands a formalscientific basis
Traffic Engineering provides a means to
determine the quantum of exchange equipmentsrequired to provide a particular level of service for
a given traffic pattern and volume.
Traffic Engineering
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
15/44
CETTM MTNL
Erlang-B formula
A: User traffic described by offered traffic AN: Network described by number of channels n
E: Quality-of-Service described by blocking
probability E
Robust to the traffic process
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
16/44
CETTM MTNL
Economy of scale
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
17/44
CETTM MTNL
Packet based transfer mode
Packetized voice
Wireless access networks Mixed core networks
Photonic backbone networks
Centralized & decentralized control
Networking development
Technological developments
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
18/44
CETTM MTNL
Differentiated services Narrowband & broadband
Real-time services:
Delay sensitive
Jitter (delay variation) sensitive Non-real-time services
Packet loss sensitive
Best effort services
Services development
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
19/44
CETTM MTNLQoS Quality of Service
User perceived QoS Operator perceived QoS
System perceived Qos
Differentiated QoS
Gold Silver Bronze in UMTS Other classifications in e.g. ATM
Service Level Agreements
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
20/44
CETTM MTNL
Traffic engineering principles
QoS can only be guaranteed by ressource reservation End
toend
1. Bandwidth based mechanism
Separation: Imply low utilization >= high cost
Minimum bandwidth guaranteed => worse case
guarantee Sharing : Imply high utilization >= low cost
Minimum guaranteed & Maximum bandwidth
We may get obtain both QoS and low cost
Virtual circuit switched networks (ATM, MPLS)Packet streams are characterized by their effective
bandwidth
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
21/44
CETTM MTNL
2. Priority mechanisms: split services into priority classes
High priority traffic:Preemptive-resume:
High QoS to limited amount of traffic
* Non-preemptive:
Lower QoS to limited amount of traffic
Low priority traffic: Best effort traffic
Requires Admission Control and Policing: specification
of traffic characteristics + control of these
Bandwidth based mechanism has built-in access
control and policing
Traffic engineering principles
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
22/44
CETTM MTNLPriority Queueing system
Type 1: Load 0.1 erlang, mean service time 0.1 sType 2: Load 0.8 erlang, Mean service time 1.6 s
No priority: W = 12.85 s (for everybody)
Non-preemptive: W1 = 1.43 s
W2 = 14.28 sPreemptive resume: W1 = 0.0056 s
W2 = 14.46 s
(twice as many type 1 jobs as of type 2)
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
23/44
CETTM MTNL
Processor sharing - Generalized
Processor sharing: all users share the
available capacityGeneralized Processor sharing: maximumcapacity for each user
Robust to the service time (file size)
Mean performance measures are the same as
for Erlangs waiting time system
This model is applicable for Best Effort traffic
(Web traffic)
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
24/44
CETTM MTNLTraffic and service characterization
A service type is characterized by
Qos parameters (discussed above) Traffic characteristics
Traffic characteristics are in general statistical
(random variables)
Examples are:
Bandwidth demand (simple):
Packetetized services(e.g. Web browsing): fluctuating
Streaming services: constant
VoIP: On/Off (two-level)
Packet arrival process (complex): Leaky bucket control
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
25/44
CETTM MTNL
Bundling (QoS point of view)
Different services should be kept separate logically. Connections with same characteristics should be
bundled
Grooming (ressource utilization point of view)To save multiplexing equipment and to increase
utilization.
This is important in core and backbone networks
Recent development in traffic modelling
Traffic and service characterization
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
26/44
CETTM MTNLTraffic Engineering Tasks
Traffic demand
characterization
Grade of service
objective
Traffic control anddimensioning
Performance
monitoring
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
27/44
CETTM MTNLTraffic Demand Characterisation
Traffic
Modelling
Traffic
Measurement
Traffic
Forcasting
Simplifying assumption
Relevant parameters
Validation of models
Estimation of parameters
value
Forcasting demands for the planning period
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
28/44
CETTM MTNL
Modelling of the user demand (E.711, E.716)
Modelling of the call demands
Call attributes: information transfer mode,
comunication configuration, etc.
Call pattern: call-level and packet-level traffic
variables Modelling of their arrival process
Modelling of the traffic offered to a group of resources
In the user plane (E.712) and in the control plane
(E.713) Modelling in mobile networks (E.760)
Estimate traffic demand in each coverage area
Estimate handover and location updating rates
Traffic Demand Charecteristics Modelling
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
29/44
CETTM MTNL
General and operational aspects
Short, medium and long term applications Operational requirements
Direct measurements vs. call detailed records.
Technical aspects
Measurements principles Criteria to choose the length of the read-out
period: statistical confidence, stationarity of the
process
Normal and high loads Estimation of offered traffic
Measurements requirements PSTN & N -ISDN , B-ISDN and SS No.7
TRAFFIC DEMAND CHARACTERISATION MEASUREMENTS
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
30/44
CETTM MTNL
Forecasting of traditional services
Principles and pre-requisites Base data: traffic, economic, social,
demographic data
Strategies for dealing with missing data
Mathematical techniques Models: curve-fitting, autoregressive,
ARIMA, Kalman filtering, etc
Methods for the choice and evaluation of
models implementation
TRAFFIC DEMAND CHARACTERISATION FORECASTING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
31/44
CETTM MTNL
Forecasting of new servicesThere are not historical data
Techniques: market research, expert opinion
and sectorial econometrics
Combination of techniques, adjustments afterimplementation
TRAFFIC DEMAND CHARACTERISATION FORECASTING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
32/44
CETTM MTNL
Finitetraffic handlingcapacity
Stochastictrafficdemands
Losses,delays..
GOSObjectives
QoS Requirements
User Oriented
Described in network
independent terms
GOS objective
Traffic related NP
objective
GOS Objectives:Principles
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
33/44
CETTM MTNL
Traffic routing Hierarchical or non-hierarchical
Fixed or dynamic (time-, state- or event-
dependent routing)
Network traffic management controls
Maintain the throughput under overload or
failure conditions
Protective or expansive controls
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
34/44
CETTM MTNL
Service protection Discriminatory restriction of the access to
circuit groups with little idle capacity
For restricting overflow traffic and for
balancing or differentiating GOS
Packet-level traffic controls
Assure packet-level GOS objectives of the
accepted calls
Provide packet-level GOS differentiation
Signalling and IN controls
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
35/44
CETTM MTNL
Dimensioning of circuit groups
Only-path circuit groups and high-usage/
final circuit group arrangements
Single-rate and multi-rate connections
Service protection methods: comparison and
parameter optimisation
GOS objectives and cost optimisation criteria
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
36/44
CETTM MTNL
Dimensioning of cirtuit groups with DCME equipment Impact of special N-ISDN features
Attribute negotiation, service reservation,
multipoint connections
Network dimensioning using end-to-end GOS
objectives
Fixed and dynamic routing
Decomposition of the network into independent
blocks
Interactive procedures for network costoptimisation
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
37/44
CETTM MTNL
TRAFFIC CONTROLS & DIMENSIONING
PACKET-SWITCHED NETWORKS (I)
New
connection
Packet-level
GOS evaluatedPacket-level
GOS satisfied Accepted
Accepted
Yes
No
Connection-oriented networks with Connection
Admission Control (CAC)
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
38/44
CETTM MTNL
Packet-level perspective
Packet-level independent from:
Connection-level offered traffic
Connection-level traffic controls
Network dimensioning
Connection-level perspective
Effective bandwidth summarises the packet level
Connection-level independent from:
Packet-level offered traffic
Packet-level traffic controls
Similar to a multi-rate circuit-switched network
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
39/44
CETTM MTNL
TRAFFIC CONTROLS & DIMENSIONING
PACKET-SWITCHED NETWORKS (I)
New
connection
Effective
bandwidth disignAvailable
bandwidth Accepted
Accepted
No
Connection-oriented networks with Connection Admission Control
(CAC)
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
40/44
CETTM MTNL
Packet-level perspective
Packet-level independent from:
Connection-level offered traffic
Connection-level traffic controls
Network dimensioning
Connection-level perspective
Effective bandwidth summarises the packet level
Connection-level independent from:
Packet-level offered traffic
Packet-level traffic controls
Similar to a multi-rate circuit-switched network
TRAFFIC CONTROLS & DIMENSIONING
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
41/44
CETTM MTNL
TRAFFIC CONTROLS & DIMENSIONING
PACKET-SWITCHED NETWORKS (II)
Framework
Principles and definitions: CAC role Strategies for logical network configuration
Packet-level traffic controls
Methods for CAC
Methods for GOS differentiation: loss and delay priority Methods for adaptive resource management: ABR and
ABT
Dimensioning
Circuit group dimensioning and network dimensioning
Traffic routing and service protection methods
Particular features of packet-switched networks
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
42/44
CETTM MTNL
TRAFFIC CONTROLS & DIMENSIONING
SIGNALLING & INTELLINGENT NETWORKS
Dimensioning of signalling networks
Performance under failures and traffic overload
Maximum design link load max
Acceptable performance at load 2max
Allocation and dimensioning of intelligent networkresources
Particular features, e.g., mass calling situations
Fast implementation of new services with uncertain
forecast Quick and flexible procedures for allocation and
dimensioning
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
43/44
CETTM MTNL
TRAFFIC CONTROLS & DIMENSIONING
SIGNALLING & INTELLINGENT NETWORKS
Traffic controls
Guidelines for the choice of control parameters
Requirements on node-level overload controls
Harmonisation principles: multivendor, multioperator
-
7/29/2019 Traffic Concept Trafic Engineering Planning (1)
44/44
CETTM MTNLPERFOMANCE MONITORING
Detect performance degradations for taking
feedback actions
Degradation reason Feedback action
Short term Overload/Failure Management Control
Medium term
Long term Traffic growth / New Networkplanning
error,.Modelling
approx
Network reconfiguration,routing
changes,control adjustments
Common aspects with traffic measurements Traffic reference periods Monitored GOS must meet GOS objectives for
normal and high loads
Consistent with traffic intensity read-out periods End-to-end GOS monitoring Methods to approximate end-to-end delays by
means of local measurements