5.10 Edge Computing

Loading...

From the course by Yonsei University

IoT (Internet of Things) Wireless & Cloud Computing Emerging Technologies

35 ratings

Yonsei University

IoT (Internet of Things) Wireless & Cloud Computing Emerging Technologies

35 ratings

Course 3 of 4 in the Specialization Emerging Technologies: From Smartphones to IoT to Big Data

IoT (Internet of Things) devices are already abundant, but new products that include IoT modules are now a common trend. Also, almost everything is already connected to a Cloud, and much more will be in the future. Naturally, as this trend continues, in the near future almost all devices and appliances will include IoT modules which will use sensor data collection and control/management based on Clouds. Since we will live in an IoT world supported by Clouds, knowledge of the core technologies and platforms of IoT and Clouds will enable you with the tools to become a true leader in the future product and business world. In this course, the start-of-the-art IoT and wireless networks and Cloud technologies are introduced (for details on 1G to 5G mobile communications and smartphone and smart device technology, please take my course “Smart Device & Mobile Emerging Technologies”). This course ends with projects that teach how to analyze Bluetooth and W-Fi wireless networks and setup and use an EC2 (Elastic Compute Cloud) Virtual Computer in AWS (Amazon Web Service), which is the most powerful and popular Cloud technology in the world. Comparing to the human body, IoT is the neural network and the Cloud is the brain. Thus, I cordially welcome you into the brain and neural network of the future intelligence world!

From the lesson

Cloud Technology

The fifth module “Cloud Technology” focuses on the Cloud market analysis, Cloud service types, MCC (Mobile Cloud Computing), and Edge Computing technology. Frist, the characteristics of the world’s top cloud companies and their services including AWS (Amazon Web Service), Microsoft, IBM, Google, and Apple's iCloud are introduced. Then the characteristics of cloud models, which include public cloud, private cloud, community cloud, and hybrid cloud are described along with the differences in cloud service models, which include SaaS (Software as a Service), PaaS (Platform as a Service), and IaaS (Infrastructure as a Service). Based on the service models, the benefits and characteristics of Cloud services are introduced. More details on the operation process are introduced, which include the IaaS and VM (Virtual Machine) administration, PaaS Runtime Environment for application support, and Open SaaS applications access processes. Then the relation between IoT and state-of-the-art mobile cloud technology is introduced. First the differences in MCC (Mobile Cloud Computing) and Edge Computing are described, which includes details on Fog computing, MEC (Mobile Edge Computing), and Cloudlet technology. In addition, the functionality and characteristics of the Cloudlet architecture and its 3 layers are covered.

5.10 Edge Computing7:55

Meet the Instructors

Jong-Moon Chung
Professor, School of Electrical & Electronic Engineering
Director, Communications & Networking Laboratory

Edge computing. Cloud technology.

Comparing MCC; Mobile Cloud Computing to Edge computing.

Edge computing is a concept to bring the cloud computing technology closer to the users.

In terms of conventional mobile cloud computing in which we have

mobile users connecting through the mobile communication core network,

through the internet to the cloud servers,

that is what we're considering as conventional mobile cloud computing.

Cloud services to mobile devices are accessed through the Internet connection.

Some of the characteristics include long time delay,

and quality of service performance which becomes issues,

high usage of the network resources because

every request will need to go through the mobile communication network and the Internet,

and then a return must be received,

high battery usage of IoT and smart devices because they are sending it to

the local base station or something that connects them to

the wide area network such that you can access the Internet.

Significant packet inter-arrival time jitter.

This is an issue because,

when you send packets,

send packets, send packets,

then when you send them,

they may have a uniform distance in time between the frames and between the packets.

However, when they go through a complex network and they go through hop

by hop to hop to reach the server, and then the server,

the Cloud computes the result and sends it back to you,

even though when the packets are leaving the original device,

they have a uniform distance between packets and frames.

Once they arrive through the Internet to a device,

these inter distances between the frames and packets get totally messed up.

In some cases, some may be very densely bursty clogged up together,

where some may be separated far far apart.

This is what can increase the overall loss probability for delay bound violations.

Especially this is important for real time multimedia services like video games,

video streaming and things like that.

For example, when we receive video files through multimedia services,

the typical data rate is based upon how many frames per second of images am I receiving,

such that when I reconstruct them they look like a motion picture.

Then, typical data rates include 20 frames per second,

30 frames per second,

40 frames per second,

and these are typically the case for one frame is made up of multiple packets.

So therefore, each second for example,

I need to receive 30 frames and each frame is made up of multiple packets.

If you think about that,

then one second is divided into 30 segments,

and each segment I need to receive

all the packets that will help me reconstruct that one frame.

Because I need to reconstruct 30 frames per second to

show it on the screen so that it looks like continuous motion picture.

In that case, if a frame does not have its packets

arrive within that one over 30th type of time interval,

then the frame will not be reconstructable,

and then we will just need to go to the next frame.

In that case, the frames that were delayed will be wasted,

because they will have delay bound

violated their time duration of when they should have been received.

That is what we call a delay bound violation.

When it comes to real time services and multimedia services like video games and videos,

this becomes very important.

Not being able to match this performance, in other words,

having large inter-arrival time jitter of

the packets significantly degrades the overall quality of service.

Edge computing is a solution for this.

Computing storage resources are moved closer to the edge,

closer to the user.

In other words they're moved closer to like base stations or access points.

Edge computing characteristics include

very short time delay with high quality of service support because it's so close,

low usage of network resources and low battery usage of IoT and smart devices.

Because you're only communicating in a short range of

near distance device that is going to quickly give you results.

But on the other hand,

it is more complex.

You have to be able to predict what is going to be used,

you need to bring that over to the Edge computing device,

and if you do that in events,

then you will get maximum speed delivery when it is requested.

But on the other hand, you need to predict and fetch it over and bring it close enough.

So therefore, there are the challenges of being able to do this the right way.

Looking at some of the characteristics of

the conventional mobile cloud computing along with Edge computing techniques.

The deployment is based on MCC is centralized.

However, Edge computing is distributed.

Meaning that centralized MCC means that the original cloud,

where the servers are that is controlling the overall cloud computing support.

So it is centralized.

Where the Edge computing device is,

something like a cloud let or an MEC,

or a fog computing device,

these are distributed and abundant and many.

Distance to the user equipment, UE.

One again, UE is what we call our smart device and

all devices that are connected to the mobile communication network,

either 4G, LTE or LTA or 5G which is going to come very soon.

We will call these devices UE. User Equipment.

So, looking at the UE as in terms of MCC,

of course, the cloud is very far away.

However, Edge computing it's very close.

Latency wise, MCC is long because we need to go through the mobile communication network,

the wide area network.

You need to go through the Internet and go

into the location of where the cloud servers are.

Then you need to bring back your information.

However, when you use Edge computing,

it is right there right next to you,

and therefore the overall latency is short.

As in terms of inter-arrival packet jitter,

and MCCs are high,

which is not good.

However, Edge computing is low,

which is very stable and more reliable.

When it comes to computational power,

the MCC is using a cloud server that is in a server location.

However, Edge computing devices,

these are small portable and a lot of cases battery operated.

So therefore, their computational power is very limited.

Also in terms of storage capability, well,

the MCCs are at the major cloud at the major servers.

So, you can store a lot.

However, what you can store on Edge computing devices are limited.

So therefore, when we look into Edge computing,

we will focus on the main three.

Fog computing, Mobile Edge computing MEC,

and also Cloudlets in the following lectures.

Explore our Catalog

Join for free and get personalized recommendations, updates and offers.

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2019 Coursera Inc. All rights reserved.

Download on the App Store

Get it on Google Play