Trends

4 minute read

Updated:

t

Clouds for the future internet and social networking

  • Cloud comp is seen as the base for anything you like to develop in the future
  • Internet of Things (IoT) - In addition to traditional access to machines, and the imp of objects and things being connected
    • Garage, car, house etc
    • If you are about to install a boiler, plumber will ask if I want access it on my phone
  • One thing likely is that this will generate massive amt of data to be stored, analysed and processed → Cloud is a foundation of these kind of applications
    • We need framework like Hadoop for data processing quickly
  • Cloud and IOT will reshape human factors and the way we live

What IoT actually means

  • Kevin Ashton said that the internet was created to connect computers, but we can connect more than that. We connect physical world so we have some form of ubiquitous computing and they’ll likely to have sensors and data
  • IDC report
    • Looks like IOT installed base will grow to 40 billion units
    • Amt of data it is gonna generate – 80—zettabytes

Architecture of the internet of things

  • image
  • In order for IoT to work, we use bottom up approach
    • Sensing layer – about things and data – to collect data
    • Network layer - Data is sent to a cloud through mobile network, internet and information network
    • Application layer – that make use of sensing data and process this data once its available
      • Smart home, environment protection, telemedicine, Intelligent traffic – cities – with sensing and traffic lights and intelligent parking system
  • A layered architecture where application relies on a network and sensing data

IOT: Smart everything

  • Take data, cloud and application approach and do sth with it
    • Spain – a lot of IoT based application on ‘smart agriculture’ to prevent for planned diseases
    • Monitoring radiation level generated by sun storms etc
    • Monitor radiation level in nuclear power plant - relies heavily on data
    • Pollution monitoring
      • More interesting bc they sense level of co2 and else related to pollution. And if it reaches certain threshold, there’s a notification saying that you need to do sth
      • Quite imp example
    • Monitor water quality – same as smart agriculture
  • Looots of data to process and cloud is there as a support foundation

Disruptive applications: New requirements

  • These applications are know as disruptive. They are disruptive bc the underlying comp structure isn’t designed to fulfil requirements of these applications
  • Sometimes, it will need to sense, store and process fairly quickly !! and have a time delay attached to the processing of the data
    • So low latency could be one of the demands to make the application successful
  • Send it to the cloud but it experiences data bottlenecks
    • Sensitive data and security concerns - you don’t want the data to be in the cloud
  • Increase network traffic and you increase the congestion in the network
  • Data analytics to have in place – will help you get new business model in order to make sense of the data

Example: Driverless cars

  • Number of sensors all around the car
  • Combination of enormous amount of sensor data and critical local processing power
  • Radar sensors to tell you what there is through a video camera, lidar sensors for light detection, GPS to know where it is, rear cameras, odometry sensors
    • Odometry – to do with how robot sense the data around it to estimate next position
  • A lot of real time applications and big data analysis
    • Decisions need to be made in milliseconds
  • Key REAL time requirement!! To ensure safety of ppl
  • Car is constantly connected to internet and the cloud – and rely heavily on cloud processing. Only problem with cloud is it could be late when receiving data from the cloud

Fog / edge computing

  • Data is processed in the closest computational resource, so the latency is low
  • Could have edge devices here and cloud data centres and in the middle have extra layer of fog layer –
    • Means that the driverless car to compute process and done quickly, you will rely on this middle layer close to the car that fulfils the requirement of the car
  • By pushing the intelligence and processing capabilities closer to where the application is!! → high bandwidth and low latency
  • Difference btwn the
    • Fog – could have many applications and they have to send data to cloud or LAN. The fog is a small cloud on LAN
    • Edge – simply send the data straight to an edge device – one device at a time.
    • They both support application, fog is LAN, number of elements, but edge is 1 device.

Network-edge computing

  • Car with embedded systems and sensors and data is generated.
  • Can send the data to cloud AND edge devices
  • you need to make calculation -Bandwidth from car to device
    • Bc - bandwidth to cloud
    • Be – bandwidth to edge
  • Driverless car would send data to cloud or edge depending on whatever!
    • Edge is close to the car so low latency and high bandwidth but cloud with higher latency

Summary

  • Computing clouds and IoT affect the entire service industry and thus the future internet evolution
  • The cloud ecosystems demand ubiquity, efficiency and security
  • Clouds are crucial for the support of future applications and shape the future of Internet.

Leave a comment