Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248 Performance Assessment of a Spectrum and Radio Resource Management Architecture for Heterogeneous Wireless Networks Kostas Tsagkaris N.Koutsouris, A.Galani, P.Demestichas Telecommunication Networks & Integrated Services Lab Dept of Digital Systems, University of Piraeus, Piraeus, Greece
25
Embed
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248 Performance Assessment of a Spectrum and Radio Resource.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Performance Assessment of a Spectrum and Radio Resource Management Architecture for
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Motivation, problem area (cont.)
??
????
RAT?spectrum?frequency? radio resources?…
??
??
????
RAT?spectrum?frequency? radio resources?…
??
Pac
ket B
ased
Cor
e N
etw
orkRAN1
RANN
……
(Cog
nitiv
e) T
erm
inal
????
????
RAN/RAT?spectrum?frequency? radio resources?…
(Cognitive) (Cognitive) Access PointAccess Point
(Cognitive) (Cognitive) Base StationBase Station
• How to optimally select and use spectrum and radio resources?
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Motivation, problem area (cont.)
• The IEEE/SCC41 P1900.4 Working Group was born to solve the problem
• In Jan 2009, the group has standardised [*] a management architecture (System/Functional Architecture + Information model) for efficiently managing cognitive/heterogeneous wireless networks and terminals
??
????
RAT?spectrum?frequency? radio resources?…
??
??
????
RAT?spectrum?frequency? radio resources?…
??
Pac
ket B
ased
Cor
e N
etw
orkRAN1
RANN
……
(Cog
nitiv
e) T
erm
inal
????
????
RAN/RAT?spectrum?frequency? radio resources?…
(Cognitive) (Cognitive) Access PointAccess Point
(Cognitive) (Cognitive) Base StationBase Station
??
????
RAT?spectrum?frequency? radio resources?…
????
????
RAT?spectrum?frequency? radio resources?…
??
??
????
RAT?spectrum?frequency? radio resources?…
????
????
RAT?spectrum?frequency? radio resources?…
??
Pac
ket B
ased
Cor
e N
etw
orkRAN1
RANN
……
(Cog
nitiv
e) T
erm
inal
Pac
ket B
ased
Cor
e N
etw
orkRAN1
RANN
……
(Cog
nitiv
e) T
erm
inal
????
????
RAN/RAT?spectrum?frequency? radio resources?…
????
????
RAN/RAT?spectrum?frequency? radio resources?…
(Cognitive) (Cognitive) Access PointAccess Point
(Cognitive) (Cognitive) Base StationBase Station
• Little work on the performance assessment of 1900.4 standardised architecture so far
• This work places focus on the assessment of signalling load/delays that the management architecture causes to the network that it operates in.
[*] [*] IEEE Std 1900.4™-2009, IEEE Standard for Architectural Building Blocks Enabling Network-DeviceIEEE Std 1900.4™-2009, IEEE Standard for Architectural Building Blocks Enabling Network-Device Distributed Decision Making for Distributed Decision Making for Optimized Radio Resource Usage in Heterogeneous Wireless Access NetworksOptimized Radio Resource Usage in Heterogeneous Wireless Access Networks
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Research approach, Methodology
Methodology:
1. Define the exact information flowing in the above interfaces• Parameters
2. Setup experimentation platform for the performance assessment of the management architecture • 1900.4 architecture implementation based on distributed agents
• After identifying and describing the exact info to be conveyed…
• Setup a platform to experiment on a IEEE 1900.4-based architecture, functionality and interfaces
• Multi-agent environment based on JADE (Java Agents DEvelopment framework)
– fully JAVA and FIPA compliant framework for developing distributed agent systems
• Agents act as high level interfaces between the involved entities and the environment and ensure their communication through the asynchronous exchange of messages (according to the FIPA ACL message structure specification)
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
RANagent
GENERATORagent
NRMagent
OSMagent
RAN agent
Generated Traffic Information
Spectrum Assignment &Rules
NetworkStatus &Policies
MTagent(s)
MT agents
Network Status & PoliciesSpectrum
UsageInfo
MT StatusNotification
RANagent
GENERATORagent
NRMagent
OSMagent
RAN agent
Generated Traffic Information
Spectrum Assignment &Rules
NetworkStatus &Policies
MTagent(s)MTagent(s)
MT agentsMT agents
Network Status & PoliciesSpectrum
UsageInfo
MT StatusNotification
Experimentation platform (cont.)
• In order to carry out the experiments in the JADE based platform:
• A set of messages have been defined
– structures that contain the defined parameters
– proper message names
• These messages are then encapsulated in the ACL messages according to the FIPA specifications and in order to flow within the agent platform
• Overhead in the total number of bytes that are conveyed for management reasons is taken into account
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Indicative scenario – Description
• Heavy loaded RAN situation
• Trigger for RAN reconfiguration
• Messages (parameters) are exchanged until all MTs will be informed about the new RAN context and are provided with the new policies
Objective: Measurement of signalling load
and time delays induced to the managed network
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Indicative scenario – Description (cont.)
• RAN becomes overloaded and sends a Context&ConfigurationNotification message to the NRM.
• Information Collection & Processing Function sends to every other RAN that covers the same geographical area a Context&ConfigurationRequest message, so as to receive the corresponding Context&ConfigurationResponse with the network status information
• SpectrumEvaluationFunction detects that a spectrum reallocation may be necessary, so it sends the evaluation results to OSM agent with SpectrumUsageEvaluation
• OSM agent sends to NRM the new allocation of the available frequencies to RATs with NewSpectrumAssignment message.
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Indicative scenario – Description (cont.)
• Network Reconfiguration Decision & Control Function makes the decision on how the overloaded RAN should be reconfigured
• NRM sends ReconfigurationRequest message to the RAN
• RAN sends to the NRM a ReconfigurationExecutionNotification message after reconfiguration takes place
• NRM derives new policies (Policy Derivation Function) and sends them together with new RAN context information to MTs using NetworkStatus&Policies message
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Results - Bandwidth calculation
• The management burden is low, providing evidence on the viability of The management burden is low, providing evidence on the viability of the architecturethe architecture
• Of special importance are the results concerning the NRM – MT Of special importance are the results concerning the NRM – MT interface since it is the interface encompassing the air-interface interface since it is the interface encompassing the air-interface
Bandwidth needed for management
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
USERUSERGenerator agent RAN agentNRM agentOSM agent MT agent
Context&ConfigurationResponse
Context&ConfigurationRequest
Context&ConfigurationNotification
Context&ConfigurationNotificationAck
SpectrumUsageEvaluation
NewSpectrumAssignment
ReconfigurationRequest
ReconfigurationRequestAck
ReconfigurationExecutionNotification
ReconfigurationExecutionNotificationAck
NetworkStatusAndPolicies
Simulated Traffic
RAN agentRAN agentRAN agent
Simulated Traffic
RAN Reconfiguration
Decision Function
Spectrum Evaluation Function
Information Collection & Processing
Function
Policy Derivation Function
Overloaded RAN
Other RANs in the area
NR
M reaction tim
e
Other RAN Reconfiguration
Results – Reaction time of the NRM
• Focus is now placed on the “NRM reaction time” i.e. time that the NRM, needs in order to resolve a problematic situation
• Dependence on – the number of APs in the
area– the number of active MTs – processing power of the
machine where the algorithmic part of the described functionality is executed
• Focus on the influence of the number of APs and MTs.
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Results – Reaction time of the NRM (cont.)
• The general outcome drawn is that the architecture is able to optimize The general outcome drawn is that the architecture is able to optimize efficiently areas encompassing several APs and a high number of MTs as efficiently areas encompassing several APs and a high number of MTs as wellwell
• The dependence on the number of the involved APs is greater than the The dependence on the number of the involved APs is greater than the dependence on the number of MTs that exist in the areadependence on the number of MTs that exist in the area
Influence on the management reaction time of the number of APs and MTs in the area
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
USERUSERGenerator agent RAN agentNRM agentOSM agent MT agent
Context&ConfigurationResponse
Context&ConfigurationRequest
Context&ConfigurationNotification
Context&ConfigurationNotificationAck
SpectrumUsageEvaluation
NewSpectrumAssignment
ReconfigurationRequest
ReconfigurationRequestAck
ReconfigurationExecutionNotification
ReconfigurationExecutionNotificationAck
NetworkStatusAndPolicies
Simulated Traffic
RAN agentRAN agentRAN agent
Simulated Traffic
RAN Reconfiguration
Decision Function
Spectrum Evaluation Function
Information Collection & Processing
Function
Policy Derivation Function
Overloaded RAN
Other RANs in the area
Other RAN Reconfiguration
Informing MTs
Results – Handling a large number of connected MTs
• Handling a large number of connected MTs
• Focus on the ”Informing time (delay)” = the time needed to inform all the interested MTs about the new network status and policies
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Results – Handling a large number of connected MTs (cont.)
Less than 1 sec is needed in most of the cases, confirming that the Less than 1 sec is needed in most of the cases, confirming that the management information flow will not be blocked or bottlenecked and management information flow will not be blocked or bottlenecked and that the MTs will obtain all the necessary information in timethat the MTs will obtain all the necessary information in time
Time needed to inform all interested MTs about the new network status and policies
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Conclusion and outlook
• An experimentation platform was setup to assess the performance of a IEEE 1900.4-based management architecture
• The platform is based on distributed agents implemented in the Java Agents Development (JADE) framework
• Indicative scenario and results show that the signaling and time delays induced into the managed network will not aggravate the overall network operation
Session 12a, 16 June 2010 Future Network & Mobile Summit 2010 Copyright 2010 E3 ICT-2007-216248
Further reading
[1] A.Galani, K.Tsagkaris, N.Koutsouris, P.Demestichas, “Design and Assessment of Functional Architecture for Optimized Spectrum and Radio Resource Management in Heterogeneous Wireless Networks”, International Journal of Network Management, to appear
[2] S.Buljore, H.Harada, P.Houze, K.Tsagkaris, et al., “Architecture and Enablers for Optimised Radio Resource usage: The IEEE P1900.4 Working Group” Communications Magazine, IEEE, Vol 47, no. 1, pp. 122-129, January 2009
[3] IEEE Std 1900.4™-2009, IEEE Standard for Architectural Building Blocks Enabling Network-Device Distributed Decision Making for Optimized Radio Resource Usage in Heterogeneous Wireless Access Networks, Jan. 2009