KRnet 2013 Keynote Speech Innovation Technology for Future Convergence Network Jinsung Choi, Ph.D. EVP, Head of ICT R&D Division, SK Telecom
Contents I. Key Trends Driving Network Evolution II. Innovation Technologies for Future Convergence Networks - Software & Cloud-based Network Architecture - SUPER Cell III. Conclusion 1
I. Key Trends Driving Network Evolution 1. Mobile & Connected 1. Mobile & Connected - Exploding data traffic growth originated from increased devices and LTE commercialization More devices always connected to N/W The number of connected devices per person will be 3 times more in 4 years LTE promising wider data transmission bandwidth More traffic induced by LTE user than by 3G users Mobile Data Traffic by Device Monthly Data Traffic per User 3 10 Exa Byte/Month Other portable devices (2.2%) M2M (4.7%) Home Gateways (4.8%) 2 3G 4G/LTE Non-smartphones (5.7%) 5 Exa Byte/Month Tablets (10.0%) Laptops & Netbooks (24.2%) 1 [Giga Byte/Month] Smartphones (48.3%) 2011 2012 2013 2014 2015 2016 10 11 12 13 Year Source: Cisco VNI Mobile, 2012 Source: SKT Internal 2
I. Key Trends Driving Network Evolution 2. Multimedia Contents 2. Multimedia Contents - Both demand and supply of multimedia contents are ever growing, mainly induced by LTE introduction and the emergence of contents platforms such as YouTube and Netflix Video is major internet traffic 51% Video Traffic, 23% Web Traffic (2010) Data Service Usage Pattern More multimedia traffic in LTE than 3G DNS Telnet OTHERS Newsgroup Email FTP Peer-to-Peer VIDEO 3 % 5 1 % Etc. SNS (11%) (7%) App. Market (11%) Multimedia (38%) SNS Etc. (12%) (7%) App. Market (10%) Sep. 2012 Multimedia (30%) WEB 2 3 % Web (33%) Web (40%) 1990 2000 2010 2 3 % 4G/LTE 3G (WCDMA) Source: CISCO based on CAIDA publication Source: SKT Internal 3
I. Key Trends Driving Network Evolution 3. Computing Evolution 3. Computing Evolution - Applications of IT Infra techniques are being extended with the help of processor cost down and evolution of computing technologies Processor cost diminishes Computing Technologies Processor cost halving time: 1.1 year Virtualization Distributed Processing Cloud Management Open API Security MapReduce HDFS Crawler Visualization Parallel Computing In-memory Computing 4
I. Key Trends Driving Network Evolution Key Innovation Technologies Key Trends Key Tech Mobile & Connected Always connected smart devices High bandwidth LTE/LTE-A M2M/IoT Multimedia Contents Mobile IPTV services Video streaming UCC sharing Computing Evolution Software & Cloud-based Network Architecture SUPER Cell for Beyond LTE-A/5G Network Cloud infrastructure and services Analytics/Big Data Personalized services 5
II-1. Software & Cloud-based Network SDN/NFV SDN and NFV allow to design and manage networks having potential to increase operator agility, to reduce cost, and to disrupt the vendor landscape Software Defined Network (SDN) Network Function Virtualization (NFV) Open Innovation Platform Lower cost, raise efficiency and increase agility Innovative Network Apps from 3 rd Parties Encourages innovation and competiveness among 3 rd Parties Virtualized Network Functions DPI CDN PE Router P-GW/S-GW Firewall WAN Accelerator CG-NAT Orchestrated, Automatic & Remote Install Control Plane Open Standards Standard/High-Volume Servers CapEx/OpEx saving Standard/High-Volume Storage Data Plane Standard/High-Volume Switches 6
II-1. Software & Cloud-based Network To-Be Image Migration to more efficient, agile, intelligent, but economic mobile network by exploiting disruptive future network technologies: SDN, N/W Virtualization, Cloud Legacy Network Software & Cloud based Network Vendor-specific Hardware, Low Scalability, Simple Policy-based Operation Virtualized, Service-Oriented and Flexible Network Analytics Platform Orchestrator SDN/Cloud Controller PCRF Analytics-based Control RU#1 RU#2 DU Pool (A-SCAN) #1 #2 #N EPC DPI Video Opt. CDN Internet RU#1 RU#2 S/W DU #1 DU #2 DU #N EPC IMS DPI CDN Video Opt. Web Accel. TCP Opt. S/W GPP, Standard H/W Virtualized/Cloudified Internet RU#N Access Network Core Network RU#N 7
II-1. Software & Cloud-based Network Use Case 1): Dynamic Service Chaining SDN Use Case: Dynamic service chaining enables a highly scalable and cost-efficient network structure As-Is Low Scalability - Most traffic passes through every service node every time Large CAPEX/OPEX To-Be More efficient N/W Operation & Time-to-market - Per-subscriber, per-application, per device differentiated data services Low Flexibility - Manual & static policy control & enforcement - Lack of flexibility and agility SDN Network SDN Controller Gi Service Network On Context A On Context B Device EPC Internet Device EPC Internet Dedicated appliances performing specific network functions Virtualized network functions running on standard H/W On Context C 8
II-1. Software & Cloud-based Network Use Case 2): Seamless Subscriber Mobility Use Case 3): Monitoring Network Control & Billing Innovator s Dream: SDN is capable of providing logically centralized control plane and common control protocol Seamless Subscriber Mobility SDN provides centralized and common control protocol working across different technologies - e.g. 3G, LTE, WiMax and Wi-Fi Monitoring for Network Control & Billing Packet handling rules in SDN switches can efficiently monitor traffic at different level of granularity - Per-subscriber/per-service statistics are collected easily at each SDN switch - Enable real-time control and billing SDN Controller enb 1 N-generation network enb 2 SDN Switch N+1-generation network 9
II-1. Software & Cloud-based Network SDN Requirement For the integration with mobile networks, 3GPP-compliancy, subscriber-awareness, serviceawareness and controllability are required?? S1-MME HSS S6a MME DPI? PCRF S11 SDN Controller S/P-GW Control Plane Gx SDN Protocol UE enb S1-U S-GW Data Plane P-GW Data Plane SGi Internet RAN Evolved Packet Core 10
II-2. SUPER Cell Motivation 5G has gained lots of attention for future radio access on developing LTE/LTE-A tech. Target - 100Gbps maximum cell capacity (1000 times larger than LTE) - 1000 times lower energy consumption - 1000 times higher capacity for connected devices 100 G 10 G.11ac.11ad 5G 1 G Wireless LAN.11n 100 M.11g.11a LTE LTE Advanced.11b 10 M 802.11 HSDPA 4G 1 M WCDMA Cellular 100 k 2G GSM CDMA 3G 1995 1995 2005 2010 2015 Year 11
II-2. SUPER Cell Vision SUPER Cell is the SKT s future network architecture combined of key elements from efficient radio resource utilization, smart cell-split, and new frequency bands SUPER Cell Vision Unified RAN (Cloud and Heterogeneous) High Rate Support Performance Optimized Energy Efficient (Power Saving) Reconfigurable (Adaptive Management) Smart (Hierarchical Scheduling) Mobility Support Smart Mobility/capacity enhanced by control/data separation Unified Cloud architecture in heterogeneous network Performance Optimized Dynamic interference coordination and management Energy Efficient Green architecture and functions for power saving Reconfigurable N/W operation mode changes interworking with SON SON: Self-Organizing Network, RAN: Radio Access Network 12
II-2. SUPER Cell Target 100 times capacity enhancement and 10 times cost reduction compared to current LTE N/W 100x Capacity Enhancement 10x Cost Down 100 Capacity 10 Cost Massive MIMO Small Cell MC TD-LTE Wi-Fi UE#1 UE#2 UE#3 f1 f2 fn f offload Freq. SON Optimization Analytics Server 1. Higher Radio Resource Efficiency 2. More Cell Splitting 3. More Frequency Bands 4. Lower Operation Cost MIMO: Multiple Input Multiple Output, UE: User Element 13
III. Conclusion - To come up with mobile data traffic explosion, SK Telecom has developed future convergence smart networks enforced by 1) SDN/NFV 2) SUPER Cell - R&D collaboration among university, industry and research organization is indispensible for global leadership and the successful development of future convergence networks [SUPER Cell demonstration in MWC2013] 14