Agile Cloud supporting the Tactile Internet Frank Fitzek Gerhard Fettweis in partnership with
Via Della Conciliazione 4.4.2005 12.3.2013 Source: http://www.spiegel.de/panorama/bild-889031-473266.html Gerhard Fettweis Source: http://www.spiegel.de/panorama/bild-889031-473242.html Quelle: japantimes.co.jp/news/2014/09/30/asia-pacific/hong-kong-democracy-protesters-set-deadline-for-demands/ Slide 2
Via Della Conciliazione 4.4.2005 12.3.2013 7 Billion Devices 2014 Source: http://www.spiegel.de/panorama/bild-889031-473266.html Gerhard Fettweis Source: http://www.spiegel.de/panorama/bild-889031-473242.html Quelle: japantimes.co.jp/news/2014/09/30/asia-pacific/hong-kong-democracy-protesters-set-deadline-for-demands/ Slide 3
Via Della Conciliazione 4.4.2005 12.3.2013 7 500 Billion Devices Billion Devices 2014 2022 Source: http://www.spiegel.de/panorama/bild-889031-473266.html Gerhard Fettweis Source: http://www.spiegel.de/panorama/bild-889031-473242.html Quelle: japantimes.co.jp/news/2014/09/30/asia-pacific/hong-kong-democracy-protesters-set-deadline-for-demands/ Slide 4
Via Della Conciliazione 4.4.2005 7 12.3.2013 Throughput Billion Devices Billion Devices 2014 2022 Source: http://www.spiegel.de/panorama/bild-889031-473266.html Gerhard Fettweis 500 Source: http://www.spiegel.de/panorama/bild-889031-473242.html Quelle: japantimes.co.jp/news/2014/09/30/asia-pacific/hong-kong-democracy-protesters-set-deadline-for-demands/ Slide 5
Wireless >2020 Outlook 100Tb/s 10 Tb/s? 1 Tb/s 100Gb/s 10Gb/s 1Gb/s 802.11ac/ad 100Mb/s 802.11ag 802.11n LTE Advanced 802.11b HSPA 10Mb/s LTE HSDPA 1Mb/s 802.11 3G R99 / EDGE 100Kb/s 10Kb/s GSM GPRS WLAN (10m) Cellular (100m) 1995 2000 2005 2010 2015 2020 2025 2030
Via Della Conciliazione 4.4.2005 7 Billion Devices 2014 12.3.2013 Throughput but there is m ore... Source: http://www.spiegel.de/panorama/bild-889031-473266.html Gerhard Fettweis 500 Billion Devices 2022 Source: http://www.spiegel.de/panorama/bild-889031-473242.html Quelle: japantimes.co.jp/news/2014/09/30/asia-pacific/hong-kong-democracy-protesters-set-deadline-for-demands/ Slide 7
The Tactile Internet Moving from 25ms RTT 1ms tomorrow http://ostsee-spezial.de/?p=148 Slide 8
Use Cases
Use Cases
Revolution Ahead: The Tactile Internet 4G: Content Communications 5G: Steering & Control Communications Health & Care Traffic & Mobility Sports & Gym Edutainment Manufacturing Smart Grid
5G Massive Requirements Massive throughput State of the art Massive low latency Massive sensing Massive resilience Massive safety and security 5G The Tactile Internet Massive fractal heterogenity > 10Gbit/s per user < 1ms RTT > 10k sensors per cell < 10 8 outage < 10 12 security 10x10 heterogeneity
Communication Networks Circuit Switched Networks Voice Places
Communication Networks Circuit Switched Networks Voice Places
Communication Networks Circuit Switched Networks Packet Switched Networks Revolution Voice Voice Data Places People
Communication Networks Massive throughput Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing Revolution Massive energy saving Internet of Things (IoT) Smart Grids Voice Voice Use Cases Remote Cars ehealth Data Flying Internet Places People Robotics
Communication Networks Massive throughput Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing Revolution Massive energy saving Internet of Things (IoT) Smart Grids Voice Voice Use Cases Remote Cars ehealth Data Flying Internet Places People Robotics
Communication Networks Massive throughput Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing Revolution Massive energy saving Internet of Things (IoT) Smart Grids Voice Voice Use Cases Remote Cars ehealth Data Flying Internet Places People Robotics
Communication Networks Massive throughput Code Centric Networks Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing A B Revolution Massive energy saving Revolution Internet of Things (IoT) Voice Voice Use Cases Smart Grids Remote Cars ehealth Things Voice Data Flying Internet Data Places People Robotics Control
Communication Networks Massive throughput Code Centric Networks Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing A B Revolution Massive energy saving Revolution Internet of Things (IoT) Voice Voice Use Cases Smart Grids Remote Cars ehealth Things Voice Data Flying Internet Data Places People Robotics Control
Communication Networks Massive throughput Code Centric Networks Circuit Switched Networks Packet Switched Networks Technical Challenges Massive reduction in delay Massive resilience Massive safety & security Massive heterogeneity Massive sensing A B Revolution Massive energy saving Revolution Internet of Things (IoT) Voice Voice Use Cases Smart Grids Remote Cars ehealth Things Voice Data Flying Internet Data Places People Robotics Control
One code to rule them all!
KODO Fast track to utilize network coding for research and industry by flexible, tested, maintained & high performance solution. Hardware and multi core support Library source code fully available. Licenses free for research/educational but paid if commercial interest exist. C++ but bindings for C, Java and Python
KODO Coding Speeds Industry trend F=GF(2^8) P=1MB Kodo 17 MT (sparse=0.5) Kodo 17 (sparse=0.5) ISA-L Jerasure 2.0 OpenFEC G=8 (12) 3096/2980 3096/2980 2255/2635 1250/1365 353/292 G=9 (13) 2542/2559 2752/2898 1961/2252 1096/1185 305/264 G=10 (15) 2136/2227 2025/2126 1724/1796 997/1072 285/245 G=16 (24) 1807/1496 1264/1239 1075/1180 628/644 179/160 G=30 (45) 950/647 672/513 266/271 349/361 96/90 G=60 (90) 594/329 359/256 123/122 184/184 48/46 G=100 (150) 383/209 226/159 74/73 111/111 29/28 G=150 (225) 266/141 153/107 47/46 74/74 19/19 Measured on Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz
Wireless Mesh Foto: Torsten Proß, Jeibmann Photographik
Kodo-Python 2-Hop Communication Example E D1 D2 encode recode decode Wanna learn more? 5G summer school @ TUD.
Recoding R S 40% 40% 80% D R 60% 60% 64 packets No need for signalling! prob. D has not received all 64 after X trans. impact of recoding impact of field size 76 81 102 105
Cloud Evolution Single/Static Distributed/Static Distributed/Agile
Example: Distributed Cloud
Example: Distributed Cloud Heterogenity (4 clouds) Clouds behave differently Speed-Up (5 clouds) RLNC does not need full degree of freedom M. Sipos, F.H.P. Fitzek, D. Lucani, and M.V. Pedersen, Dynamic Allocation and Efficient Distribution of Data Among Multiple Clouds Using Network Coding, in IEEE International Conference on Cloud Networking (IEEE CloudNet 14), Oct. 2014. M. Sipos, F.H.P. Fitzek, D. Lucani, and M.V. Pedersen, Distributed Cloud Storage Using Network Coding, in IEEE Consumer Communication and Networking Conference, Jan. 2014.
Mobile Edge Cloud / Micro Cloud / Cloud
Dynamic Distributed Cloud No Coding RS Coding Network Coding reliability reliability reliability storage traffic storage traffic storage traffic reliability storage traffic F. Fitzek, T. Toth, A. Szabados, M.V. Pedersen, D. Lucani, M. Sipos, H. Charaf, and M. Medard, Implementation and Performance Evaluation of Distributed Cloud Storage Solutions using Random Linear Network Coding, in IEEE International Conference on Communications - Cooperative and Cognitve Network Workshop - CoCoNet6, June 2014.
Dynamic Distributed Cloud No Coding RS Coding Network Coding reliability storage traffic
Software Defined Networks End to End Coding Schemes: Store and Forward
Software Defined Networks Hop by Hop Coding Scheme: Store and Forward
Software Defined Networks Network Coding Scheme: Compute and Forward
37 Software Defined Networks Latency gain of e2e vs RLNC (left) and hbh vs RLNC(right)
Results Packets 64 Size 250 B Bitrate 0.25 Mb/s RLNC gain over E2E RLNC gain over HBH
Radio Network Controller Backbone Radio Network Controller COMP Multicast MASSIVE MIMO BS MASSIVE MIMO BS IoT MASSIVE MIMO BS Relay MASSIVE MIMO Relay Multi-Hop Distributed P2P
Thank you 5G L A B GERMANY