An SDN-based Network Architecture for Extremely Dense Wireless Networks H. Ali-Ahmad, C. Cicconetti, A. de la Oliva, V. Mancuso, M. R. Sama, P. Seite, S. Shanmugalingam
An SDN-based Network Architecture for Extremely Dense Wireless Networks
H. Ali-Ahmad, C. Cicconetti, A. de la Oliva, V. Mancuso, M. R. Sama, P. Seite, S. Shanmugalingam
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 2
The good ol’ days
MSC
BSC (2G) RNC (3G)
BSC (2G) RNC (3G)
Works very well… …until the archenemy arrives
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 3
Whoops
Archenemy
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 4
3.5G
Temporary solution: Push intelligence to network edges (= RAN)
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 5
But the Big Problem is ahead
Traffic
Revenues
Data era Voice era
Telecom operator
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 6
Can we “just” increase tx rate?
• Today’s radio access technology is (largely!) overprovisioned:
– IEEE 802.11n: easily 100 Mb/s, can be 300 Mb/s
– HSPA+: up to 84 Mb/s
– LTE: up to 300 Mb/s
– Let’s not even mention LTE-A!
• Answer is: NO. WAY.
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 7
Only way: densification
From macro-cell only
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 8
Only way: densification
From macro-cell only to pico-cells
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 9
Only way: densification
From macro-cell only to pico-cells to small-cells
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 10
Densification’s issues
Very high density + heterogeneity =
1. Interference in the radio access network
2. Poor efficiency in the backhaul
3. High signalling for traffic management
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 11
CROWD
• Connectivity management for energy optimised wireless dense networks
– From Jan/13 to Jun/15
– Value: 4,5 M€
– Effort: 443 PM
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 12
High-level view
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 13
Example: LTE
LTE does not cope well with extreme density
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 14
Example: LTE
That’s better
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 15
Issues solved!
Very high density + heterogeneity =
1. Interference in the radio access network – Fast control with local visibility
2. Poor efficiency in the backhaul – Multi-RAT traffic management with regional
visibility
3. High signalling for traffic management – Optimisation of local handover. Decoupling of
data vs. control paths paves way to aggregation
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 16
Tangible output
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 17
CROWD controller
DB
North-bound (REST) API
Aggregator Virtualizator
Translator
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 18
Validation
Please follow us at: http://www.ict-crowd.eu/ website @FP7CROWD Twitter channel
Backup
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 21
Goal
To enable sustainable deployment of very dense and heterogeneous wireless networks
sustainable = cost effective + energy efficient
very dense = 1000x compared to current density (in users/sqm vs. users/BS)
heterogeneous = + diverse range (macro vs. pico vs. femto)
+ diverse technologies (LTE vs. WiFi)
+ diverse deployments (planned vs. unplanned)
+ diverse backhaul types (optical vs. wireless)
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 22
Roadmap
End of 2013
• Architecture final
• Preliminary solutions
• Numerical / simulation analysis
End of 2014
• First public release of CROWD controllers
• Final solutions validated with stand-alone testbeds
• Initial results from integrated testbed
June 2015
• Final demonstrator ready
• Full results available
SDN4FNS – Trento, Nov 11, 2013 Follow us at @FP7CROWD 23
Expected impact
• Raise awareness about technical challenges with extremely dense and heterogeneous wireless network – Workshops, conference sessions, special issues
• Propose architecture / solutions to the relevant venues – Standardisation committees, industry alliances,
stakeholder groups, demonstrations at public events
• Contribute the CROWD controller to the SDN/NFV community – Release of the source code under GPL, integration with
well-known platforms, tutorials and webinars