jakab
TRANSCRIPT
-
8/13/2019 jakab
1/27
Planning SwitchedOptical Networks
Tivadar Jakab, Zsolt [email protected]
Budapest University of Technology and Economics
Department of Telecommunications
-
8/13/2019 jakab
2/27
4/8/03 Third Hungarian WDM Workshop 2
Outline
Network planning: Traditional approaches and newchallenges
Planning dynamic networks: Requirements and
solutions Network resilience: Service and operational
considerations
Illustrative numerical examples
Summary and conclusions
-
8/13/2019 jakab
3/27
4/8/03 Third Hungarian WDM Workshop 3
Network Planning
Planningprocess
Service demands Available technologiesand equipment
Business considerations
Co-operation
requirements
Technical
constraints
Current network
Optimalnetwork
plan
-
8/13/2019 jakab
4/27
4/8/03 Third Hungarian WDM Workshop 4
Traditional Network PlanningApproaches
Services
Forecast
Demands
Planning
Network extension
Emerging new services
Modelling and forecasting
difficulties
Uncertainties
-
8/13/2019 jakab
5/27
4/8/03 Third Hungarian WDM Workshop 5
Planning Approaches for DifferentTime ScalesLTP demandforecast T0
Installedplan T0
T0 T1 T2 T3 T4
a) Long-term planning approach
b) Medium-short term planning approach
Installedplan T0
T0 T1 T2 T3 T6T4
Target LTPplan T0
ST-MTplan T0
ST-MTplan T1
ST-MTplan T2
New long term planning processSTP-MTP
demand forecastT0
. . . . .
Installedplan T2
T6
. . . . .
Target LTPplan T2
Long term planning process
T0-T4
LTP demandforecast T2 Long term planning process
T2-T6
LTP planT0
LTP planT2
Medium-short term planning process
-
8/13/2019 jakab
6/27
4/8/03 Third Hungarian WDM Workshop 6
New Challenges in NetworkPlanning
Soft-permanent OCh services More dynamic networks
Client traffic (IP) modelling and forecasting
difficulties Sharp competition
Strict economical conditions
How to save money? How to make money?
-
8/13/2019 jakab
7/27
4/8/03 Third Hungarian WDM Workshop 7
Planning Dynamic Networks:Requirements
Flexibility Manageability
Fast provisioning
Advanced resilience
Service differentiation
-
8/13/2019 jakab
8/27
4/8/03 Third Hungarian WDM Workshop 8
Planning Dynamic Networks:Solutions
Network
Real-time traffic control
Configuration management(Traffic management and resourceconfiguration)
Network planning, and development
(Traffic and transport)
Network operation, management and planning
Measured traffic data
Link capacities(Traffic and transport)
Routing strategy
(Traffic and transport)
Months - years
Days - weeks
Minutes - seconds
Traffic
-
8/13/2019 jakab
9/27
4/8/03 Third Hungarian WDM Workshop 9
Planning Dynamic Networks:Solutions
Optimal
network extension
Network consolidation
Rearrangements to achieve
the optimal configuration
Provisioning
Network configuration
to serve dynamic requests
-
8/13/2019 jakab
10/27
4/8/03 Third Hungarian WDM Workshop 10
Illustrative examples (1)Penalty on Dynamic Behaviour 2/1
Permanent channel requestsarriving randomly in time and
space
A simple distribute control
mechanism is to decide on
connection set-up
Routing: shortest path based
Wavelength assignment:
sequential (first fit)
Connection set-up is based onlocal, sub-optimal decisions
(requests should be served -
minimal blocking)
All channel requests areknown in advance
Pre-planned path selection
and wavelength assignment
Routing: shortest path based
Wavelength assignment: after
path selection (fixed route)
Optimal solution (targetfunction: minimal resource
need)
Dynamic Network Theoretical optimum
-
8/13/2019 jakab
11/27
4/8/03 Third Hungarian WDM Workshop 11
Network example:Hypothetical Hungarian Backbone
Network structure:
9 nodes, 16 links
average nodal degree: 3.5
Demand pattern:
Heavy star + light mesh
-
8/13/2019 jakab
12/27
4/8/03 Third Hungarian WDM Workshop 12
Illustrative examples (1)Penalty on Dynamic Behaviour 2/2
0
500
1000
1500
2000
2500
3000
Dynamic Pre-planned OTN
Network Type
OCh*h
op
Extra OCh to be installedOCh in use
Same routing
without effective
capacity constraints
(simplified case)
30% penalty on
dynamic behaviour
Evaluation of examples based
on link characteristics:
number of channels in use
number of channels to be
installed w/o WLS
-
8/13/2019 jakab
13/27
4/8/03 Third Hungarian WDM Workshop 13
Illustrative examples (2)
A More Perspective Case - Penaltyon Dynamic Switching
Automatically Switched Optical Network Switched OCh service
Illustration of potential statistical gains (traffic
concentration) and penalties on dynamic behaviour Switched optical channels based service for on-
demand connection requests
Instead of providing leased lines result in statisticalgains for the operator
-
8/13/2019 jakab
14/27
4/8/03 Third Hungarian WDM Workshop 14
Illustrative examples (2)
Automatically Switched OpticalNetworks - ASON
Based on advanced signalling (GMPLS) and switchingcapabilities which facilitate the introduction of intelligentflexibility and distributed management into optical networks
Optical
switch
Transport plane
PI
OCC Control plane
NNI
CCI
Clientequipment (IP
router, ATMswitch, )
OCC
Optical
switch
PI
UNI
NMI-T
NMI-A
Optical
switch
OCC
EM/NM
Managementplane
Optical
switch
Transport plane
PI
OCC Control plane
NNI
CCI
Clientequipment (IP
router, ATMswitch, )
OCC
Optical
switch
PI
UNI
NMI-T
NMI-A
Optical
switch
OCC
EM/NM
Managementplane
CCI: Connection Control Interface
NMI-A: Network Management Interface for the ASON Control Plane
NMI-T: Network Management Interface for the Transport Network
NNI: Network to Network Interface
OCC: Optical Connection Controller
PI: Physical Interface
UNI: User to Network Interface
-
8/13/2019 jakab
15/27
4/8/03 Third Hungarian WDM Workshop 15
Comparison Model On-off sources:
permanent connection in OTN case
switched connection in ASON case
A pont
B pont
C pont
D pontA pont
B pont
D pont
ton1 toff1 ton2
B2 forrs
Time ton1 toff1 ton2
B2 forrs
Time
OTN with permanent connections ASON with switched connections
C pont
-
8/13/2019 jakab
16/27
4/8/03 Third Hungarian WDM Workshop 16
OTN penalty for WLA
without WLC (about 60%)
Statistical gains due to switching
below 0.7 source activity ratio
Higher penalty for dynamic WLA
without WLC (about 100%)
Statistical gains from
protected switchedOCh based services
-
8/13/2019 jakab
17/27
4/8/03 Third Hungarian WDM Workshop 17
Statistical gains from protectedswitched OCh based servicesSource activity ratio
25%
Onlyswitched OCh
Only
permanent OCh
-
8/13/2019 jakab
18/27
4/8/03 Third Hungarian WDM Workshop 18
Statistical gains from switchingin case of mixed services
0
500
1000
1500
2000
2500
3000
3500
4000
4500
"75:25""50:50""25:75"OTN 0:100
permanent
onlySwitched
to permanent rate
OCh*hop
Och to be installed
Och in use
Different switched to permanent rate
(switched with 25% source activity ratio)
100:0
switched onlyIncreasing rate of switched traffic
100% 100% 87% 75% 57% 44%
100% 112% 97% 82% 58% 46%
-
8/13/2019 jakab
19/27
4/8/03 Third Hungarian WDM Workshop 19
Network Resilience Dynamic application of extra network
resources to limit the impact of failures Based on dedicated or shared network
resources
Requires switching function to be supportedin nodes
Basic schemes: 1+1 dedicated protection
n:m shared protection
restoration (dynamically configured shared capacities)
-
8/13/2019 jakab
20/27
4/8/03 Third Hungarian WDM Workshop 20
Illustrative examples (3)
Resilience in Full FlexibleNetworks
0%
50%
100%
150%
200%
250%
300%
not
prot.
optimal
path
rest.
1+1
Resilience Cases
RelativeHop*OCh
Extra for resilienceWorking
0%
50%
100%
150%
200%
250%
300%
not prot.
full flex.
1+1 term.
switch.
1+1 full
flex.
optimal
path rest.
full flex.Resilience Cases
Relative#SwitchPort
sExtra due to resilience
Working
Full flexible network: each capacity unit terminates onswitching-capable node equipment (e.g. flex. OADM
or OXC)
Comparison of link capacities Comparison of switch capacities
Extra for
dedicated
resilience
Savings on
capacitysharing Switches
not used
for resilience
-
8/13/2019 jakab
21/27
4/8/03 Third Hungarian WDM Workshop 21
Network ResilienceService Considerations
Different applications may need resilience withdifferent characteristics, such as
recovery speed
rate of recovered capacity (partial/entire) Different resilient classes can be specified
according to the different needs
Aim: meet different resilience requirements on thesame technical basis and lowest cost
-
8/13/2019 jakab
22/27
4/8/03 Third Hungarian WDM Workshop 22
Illustrative examples (4)
Restoration with Tailored RecoveryTime
Recovery time is assumed to be proportional with thenumber of active switching nodes involved in theprocess
Some switches can be pre-set and fixed to speed upthe recovery process
Reduced flexibility results in less efficient capacitysharing, therefore the amount of extra resources forrestoration is increasing
The joint optimisation of different classes may reducethe penalty
-
8/13/2019 jakab
23/27
4/8/03 Third Hungarian WDM Workshop 23
Illustrative examples (4)
Restoration with Tailored RecoveryTime
Impact of Tailored Recovery Time
on Network Configuration
0%
20%
40%
60%
80%
100%
Case 0 Case 1 Case 2 Case 3 Case 4 Case 5
Service Class Cases
Rateoflogicalhopswith
differentnumberofpre-set
switches
4 pre-set switch
3 pre-set switch
2 pre-set switch
1 pre-set switch
0 pre-set switch
Impact of Tailored Recovery Time
on Network Resources
0
50
100
150
200
250
300
350
400
450
Case 0 Case 1 Case 2 Case 3 Case 4 Case 5
Service Class Cases
R
esourceNeeds[OCh*hop]
spare for resilience
working
Average hop count of recovery paths
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
Case 0 Case 1 Case 2 Case 3 Case 4 Case 5
Service Class Cases
Averagehopcount
Traditional
restoration
1+1dedicated
protection
-
8/13/2019 jakab
24/27
4/8/03 Third Hungarian WDM Workshop 24
Network ResilienceOperational Considerations Guaranteed Capacity Optimal Restoration (traditional)
centralized management, explicit routing
Minimal Route Restoration
fits to the distributed environment - easy to implement
Disjoint Route Restoration
source routing
simple to return to the working route after clearing the failure
1+1 Dedicated Path Protection
source routing
Shared Path Protection
source routing, at least 3-connectivity of the network structure isrequired
-
8/13/2019 jakab
25/27
4/8/03 Third Hungarian WDM Workshop 25
Illustrative examples (5)
Restoration with DifferentOperational Options
0
200
400
600
800
1000
1200
1400
1600
1800
2000
NotProt Capac. Opt.
Rest
Min. Route
Rest
Shared PP 1+1 Disj. Route
Rest.Protection Options
TotalOC
h*hop
extra # channels toimplemented the
network w/o WLC
#channels in use
100% 91% 113% 131% 139% 149%
100% 140% 242% 203% 248% 257%
-
8/13/2019 jakab
26/27
4/8/03 Third Hungarian WDM Workshop 26
Summary and Conclusions
The impact of intelligent switching capabilities onthe planning of dynamic optical networks has beenoverviewed
Network resilience has been studied taking into
account service and operation orientedconsiderations
Some small numerical examples have been
presented to illustrate the presented planningapproaches
-
8/13/2019 jakab
27/27
4/8/03 Third Hungarian WDM Workshop 27
Related Publications
T. Jakab, Zs. Lakatos: Protection and restoration based resilience in Automatic Switched Optical
Networks, Proceedings of NETWORKS 2002, Munich, Germany, June 2002, pp. 459-467 T. Jakab: 4.3 Network Planning in TELECOMMUNICATION NETWORKS AND INFORMATICS
SERVICES Chief Editor: dr Gyrgy Lajtha, http://www.hte.hu/ob/eng/4.pdf, Scientific Association forInfocommunications (HTE), 2002.
T. Jakab, Zs. Lakatos: Protection or restoration: a generic study on the impact of line costs andswitching costs on the strategic decisions, Proceedings of DRCN2001, June 2001, Budapest, Hungary
T. Jakab, R. Clemente, C. Mas, H. Nakajima: Techno-economical analysis of ASON: Comparison of
resilience options - First Results from EURESCOM FASHION Project, Proceedings of DRCN2001,June 2001, Budapest, Hungary
T. Jakab, Zs. Lakatos: Protection or restoration: a generic study on the impact of line costs andswitching costs on the strategic decisions, Magyar Tvkzls (Hungarian Telecommunication Journal),XII. (2001), Selected Papers
EURESCOM P1012 FASHION Project Publications http://www.eurescom.de/public/projects/P1000-series/p1012/default.asp