Skip to content

Latest commit

 

History

History
192 lines (173 loc) · 9.25 KB

File metadata and controls

192 lines (173 loc) · 9.25 KB

* Fog Networks and the Internet of Things

http://www.gns3.com/

Lecture 1: Fog Networking - Going from Cloud to Fog

Part A (8:51)

Part B (21:28)

▶ What is Fog network A network architecture that uses one or a collaborative multitude of end-user clients (phones, laptops) or near-user edge devices (set-top boxes, Femto base stations, home gateways) to carry out a substantial (էական, nor 100% but substential amount of something) amount of
  1. storage (rather than stored primarily in cloud data centers),
  2. communication (rather than routed over wide-area so called backbone networks),
  3. control, configuration, measurement and management (rather than controlled primarily by network gateways such as those you can

find in the so-called core network of 4G LTE). ▶ What is Fog network Is the architecture who describes

  • Who does What?
  • How to connect them?
  • Architecture - not allocating resources but allocating functions in network.

And Fog network is an architecture that use could be just one or could be more than one that work with each other - thet is what we mean by collaborative multitude.

Part C - Use Cases (8:29)

Что такое фемтосота(Femtocell)? Use cases of fog networks ▶OTT Service Charging Clients can crowd-source network inference/measurement and overlay billing ▶Client-Side HetNets Control - Clients can autonomously manage/control their own configuration How car choose which line to go over 3 line? HetNets - Hetergeneous Networks WiFi AP - LTE RNS(eNodeB) - HNS(Home Gateway, HomeNodeBN) - FemtoCell ▶Client Resource Pooling Idle resources on the edge pooled but shared unpredictably

Part D Fog as a Network Architecture (10:37)

▶Fog as network architecture 5G, IoT, Big Data, EDD(Edge-driven “Data center”), EDC(Edge-Driven Control plane)

▶Examples • Over The Top (OTT) content management and SDP • Client-centric HetNets Control • Client-controlled cloudstorage • Session management and signaling load optimization • Crowd-sensing inference of network states • Edge analytics and real-time stream-mining • Clients’ idle computing/storage/content pooling • Edge caching/BW management at home gateway/small cell • FlashLinQ/LTE Direct/WiFiDirect/AirDrop • Cloudlets/Micro Datacenters/Mobile CDN • Client-driven distributed beam-forming(Проектирование фазированных антенных решёток) ▶NFO - Network Function Optimization, there are several threads to rethinking • Redefine Network Functions - CCN(Content Centric Network) or ICN(Information Centric Network) Manipulating digital objects not packets, and this redefine a lot functions that we need to think about • Virtualize the network not just computing infrastructure- NFV(Network Function Virtualization) • Relocating the network functions - FOG

▶Research Challenges • Trustworthiness/ verification of client/edge software & hardware - Do I trust this cients or things or Edges? • Incentivization(Խթանել) of client participation - maybe they do not want to participate on resource pooling, crows-sourcing, how do I Incentivization their participation? • Interactions w. client/IoT OS and definition of network service APIs • Interfaces: Cloud-to-cloud and cloud-to-fog • Oscillation/divergence and global configuration consistency during the interactions of local actions • Tradeoff of Local vs. Global architecture, Distributed vs. Centralized • • ▶Inter-Disciplinary Solutions - http://fogresearch.org ▶Impact on Value Proposition along Ecosystem Food-chain • End user experience providers? • Network operators? • Equipment vendors? • Cloud service providers? • System integrators? • Edge device manufacturers? • Client/IoT device manufacturers/OS? • Chip suppliers?

Lecture 2: From IT(information technology) to IOT

Part A (17:48)

5G, IoT, Big Data, EDD, EDC IoT = I & T • Internet
  • Connected

• Things

  • Communicate (port or antenna)
  • Sense (sensors …)
  • Actuate - cahnge something in environment(embedded controllers…)
  • Compute (processor)
  • Store (memory)
  • Interact (touchescreen)

IoT Applications’ Classification ▶Consumer oriented • Consumer & Wearable (virtual reallity devices, glasses, colothes) • Home & Building

▶Industrial Oriented • Healthcare • Vehicular & Transportation • Smart grids & Energy • City & Urban • Environment & Agriculture • Industrial & Logistics

Part B (14:23) The heterogeneity of out IT application need, is a major chanllange that researchers in this filed have to tackle.

We should ask ourselves in this particular IoT application, What kind of energy, delay, throughput, signaling cost, and human in the loop conditions are required.

▶Heterogeneous needs • Energy • Delay - • Throughput - has many order of magnitute variation within IoT applications. • Signaling overhead - • Cost • Human in the loop

▶ IoT use cases requirements

Smart meters(cities)eHealthITS*Surveilance
MobilitynonePedestrian/vehicularVehicularnone
Message sizeSmall(few Kb)Medium?Meduimlarge
Traffic PatternRegularRegular/irregularRegular/irregualrRegular
Device densityVery high(10 000 per cell)MeduimHighlow
Latency requirementslow(up to hours)Meduim(seconds)Very high(few ms)Meduim(<200ms)
Power efficency requirmentsHigh (battery powered meters)High(battery power devices)Lowlow
RelialbilityHighHighHighmedium
Security requirementsHighVery highVery highmedium
  • ITS - Inteligent transportation system

Part C (19:05) - IoT Layers

▶ IoT API and Architecture

Part D (16:08) - Interview: Dr. Adam Drobot

Lecture 3: Network Service Pricing

Part A (21:28)

▶ “Dump” pricing
  • Light users subsidizing heavy users
  • Consumers subsidizing(large) businesses
  • Bandwidth “hogs” damaging all other users

▶ Smart Data Prcising(SDP)

  • usage-based pricing (first problem is solved)

▶SDP Dimensions A. How to charge?

  1. Usage-based, quota rollover/trading
  2. Time/location/congestion - dependent
  3. Prepaid vs. post-paid, Contract vs no contract

B. Whom to charge??(2-sided, 1-800, toll-free, zero rating, sponsored data, unlimited,…)

  1. Closed
  2. Open

C. What service are you charged from?/ Coount byte

  1. Paid Priority - queueing systems
  2. Transaction-based (instead of counting bytes, we ask have you finished, Pioners are amazon->Kindle, you oay not for bytes)
  3. Cloud service pricing
  4. IoT service pricing

Part B (10:01)

Part C (11:03)

Part D (17:54)

Lecture 4: Principles of P2P Networking

Part A (21:28)

Part B (10:01)

Part C (11:03)

Part D (17:54)

Lecture 5: Client-Driven Control & Configuration

Lecture 6: Client-based Measurement IoT Signaling

Lecture 7: Edge Pooling and Caching

Lecture 8: Fog/IoT Security and Privacy

Lecture 9: Consumer/Wearable/Home IoT

Lecture 10: Connected Cars

Lecture 11: Smart Grids

Lecture 12: Healthcare