The 3GPP Long Term Evolution Advanced (LTE-A) standard specifies a set of pioneer features such as relay nodes and carrier aggregation. At the same time, the Software Defined Networks (SDN) have become an emerging technology which provides centralized control and programmability to modern networks. In the current communication environment, cloud computing could combine the advantages of both technologies in order to create a novel cloud assisted Software Defined LTEA architecture with relay nodes. Moreover, due to the increased requirements of modern services, the optimal resource allocation is a necessity. In such a context, this paper describes a QoS aware cross carrier scheduler for downlink flows, aiming at the optimization of system resources allocation. The proposed scheduler is evaluated against the PF, MLWDF, EXP/PF, EXP RULE, LOG RULE, FLS and FLSA schedulers in a cloud assisted Software Defined LTE-A topology with relay nodes. Simulation results show that the proposed scheduler improves the real time services performance while at the same time maintains an acceptable performance for best effort flows.
QoS-aware scheduling in LTE-A networks with SDN control (presentation)
1. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
QoS-aware scheduling in LTE-A networks
with SDN control
Emmanouil Skondras1, Angelos Michalas2,
Aggeliki Sgora1, Dimitrios D. Vergados1
1Department of Informatics, University of Piraeus, Piraeus, Greece
2Department of Informatics Engineering, Technological Education Institute of Western
Macedonia
1
2. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Outline
• Introduction
• Resource Allocation Schemes & Algorithms
• The FLSA-CC scheduler
• Simulation Results
– Simulation Setting
– Simulation Results
• Conclusions
3. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Introduction
• A three level cross carrier scheduler, the FLS-
Advanced Cross Carrier (FLSA-CC), is proposed.
• Downlink packet scheduling in LTE-A networks with
Relay Nodes.
• Cloud assisted SDN architecture.
• FLSA-CC aims at QoS aware resource allocation, in
order to:
– Satisfy the requirements of strict real times services.
– Maintain an acceptable throughput for best effort flows.
3
4. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Scheduling Strategies for LTE-A
• Several downlink packet schedulers have been
proposed in the current literature.
• They can be classified into two groups:
– Non-QoS aware
– QoS aware
• A non-QoS aware scheduler does not take into
account parameters that affect the service quality.
• A QoS aware distributes resources considering the
specific constraints of each service.
4
5. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Scheduling Strategies for LTE-A
• In LTE info is transmitted in
10ms frames.
• Each frame is spit into 10
sub-frames of 1ms TTI.
• Each TTI consists of
resource blocks (RBs) - the
minimum allocation unit.
• A scheduler assigns an RB
to the user with the biggest
metric.
5
6. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Proportional Fair (PF)
Non-QoS aware scheduler
• Proportional Fair (PF)
di
k(t): Available throughput in the kth RB of the ith flow.
𝑅i(t-1): Past average throughput.
6
7. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
QoS aware schedulers
• Modified Largest Weighted Delay First (M-LWDF)
• Exponential/PF (EXP/PF)
DHOL,i: Head of line delay.
δi: Target packet loss ratio.
τi: Delay constraint.
Nrt: The number of active real time flows. 7
8. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
QoS aware schedulers
• LOG RULE
• EXP RULE
DHOL,i: Head of line delay.
Nrt: The number of active real time flows.
Γi
k: Spectral efficiency for the ith flow on the kth RB.
bi and c: Configurable parameters.
8
9. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
QoS aware schedulers
• Frame Level Scheduler (FLS)
– Two level QoS aware strategy.
– Upper level
• Estimates the ui(x) quota of
data that the ith real time
flow must transmit at the xth
frame to succeed its QoS
constraints.
– Lower Lever
• Uses the PF metric to
allocate RBs to flows
qi(x): Queue length in the xth frame.
Mi: the number of coefficients used.
ci(n): The nth coefficient value.
τi: The target delay.
9
10. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
The Proposed Scheduler – FLSA-CC
• Improved version of the Frame Level
Scheduler Advanced (FLSA)[3].
– Cross carrier scheduling.
– Real time flows receive higher priority than the
best effort ones to fulfill their constraints
– Maintains an acceptable level of performance for
BE flows
– Relay assisted LTE-architecture.
10
11. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
The Proposed Scheduler
11
12. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
The Upper Level
• Uses the formula of FLS
– To estimate the quota ui(x) of data that the ith real time
flow should transmit in each xth TTI, to succeed its QoS
constraints.
• The FLSA-CC estimates the coefficient value ci(n) using formula:
where N is the number of aggregated component carriers.
• ui(x) quota is estimated in each xth TTI of a frame.
– Whereas in FLS it is estimated once at the beginning of
each xth frame.
12
13. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
The Upper Level
• Performance improvement has been observed.
– Due to the fact that:
• In FLS, when a real time flow transmits its ui(x) quota of
data, it loses the opportunity to continue the
transmission until the beginning of the next frame.
– By recalculating the formula in each TTI (instead of
estimating it only at the beginning of each frame):
• The FLSA-CC provides more resources to real time flows
that have remaining data for transmission.
13
14. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
The Middle & Lower Levels
• The middle level uses a CC version of MLWDF in each
TTI.
– Realizes improved resource distribution among the
real time flows.
• In comparison with the FLS scheduler which at the
second level uses the non-QoS aware PF.
• The lower level uses a CC version of the PF
– Allocates the remaining RBs of each TTI to real time
and best effort flows.
14
15. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Performance Evaluation
• The performance of the FLSA-CC was
evaluated against the schedulers:
– PF
– M-LWDF
– EXP/PF
– FLS
– FLSA
– EXP-RULE
– LOG-RULE
15
16. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Performance Evaluation
• An extended version of the open source
simulator LTE-Sim used.
– The iCanCloud and the OpenFlow modules of the
Omnet++ simulator have been configured and
embedded to the Lte-Sim.
• Enabling the ability to include cloud infrastructure and
SDN controllers to the simulated LTE topology.
16
17. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Performance Evaluation
• The simulation parameters:
17
18. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Performance Evaluation
• The Cloud implements the
functionalities of the LTE Evolved
Packet Core (EPC).
• Flow forwarding as well as
resource scheduling in each
DeNB and RN.
– Are performed using a
centralized global controller.
• Placed into the SGW on
the cloud.
18
19. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Performance Evaluation
• The parameters considered in each scheduler:
19
20. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
VoIP and Video packet loss ratio using
different target delays
20
• The impact of the target delay parameter for the case of having 100 users per RN.
• FLSA-CC compared with the rest of the schedulers exhibits lower PLR independent
of the target delay parameter.
21. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
VoIP and Video packet loss ratio
21
• Considered target delays: 100ms for VoIP and 150ms for video flows, as
determined by the LTE QoS class specifications.
• FLSA-CC results in a lower PLR than the rest of the algorithms.
• Marginal decrease of its PLR for VoIP flows as well as up to 3% lower PLR for video
flows compared to FLSA.
22. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
VoIP and Video throughput
22
• FLSA-CC succeeds higher throughputs than the rest of the algorithms
providing rates of up to 800kbps for VoIP and up to 28Mbps for video services.
23. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
VoIP and Video fairness index
23
• FLSA-CC scheduler improves the fairness for both VoIP and video flows.
24. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Best Effort throughput and fairness
index
24
• FLSA-CC outperforms the other two schedulers and provides throughput up to
1.5Mbps for best effort flows even when the number of users increases.
• While the FLSA accomplishes only a 100kbps throughput.
• Additionally, the FLSA-CC scheduler significantly improves the fairness index of best
effort flows.
25. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
Conclusions
• FLSA-CC cross carrier downlink scheduler
• Performance evaluation against other
scheduling algorithms
– In a cloud assisted SDN architecture
– LTE-A network with relay nodes
• FLSA-CC achieves better performance in terms
of PLR, attainable throughput and fairness.
25
26. The 7th International Conference on Information, Intelligence, Systems and Applications (IISA 2016)
26
Thanks for your Attendance!
Comments, Questions?