Cloud Native Computing Foundation KubeCon + CloudNativeCon China 2019 (Shanghai), 5 Jul 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IoT Application Running on KubeEdge + ARM Platform - Xuan Jia & Bin Lu

Description

Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io

Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects

IoT Application Running on KubeEdge + ARM Platform - Xuan Jia, Edge Computing Architect & Bin Lu, ARM

We think k8s will have broad potential sitting between IoT edge devices/gateways, edge nodes and cloud platforms. So, we focus on this domain. Our initial goal is to evolve k8s for IoT edge applications: 1, Supporting industrial IoT use cases scaling to millions of edge devices 2, Via Edge nodes, bringing computing closer to data sources. Anticipated benefits include reduce latency, high performance and improved reliability. In this presentation, we will introduce and show our cloud-native broadband router CNF based on k8s edge computing platform. Major features including: 1, Hardware: we will demo a personal and portable edge cloud. This cluster consists of 4~5 Arm boards. 2, Container orchestration: Kubeedge + K8s 3, Deployment: helm 4, Container: minimal size container build 5, CNF: distributed gateway 6, other features: cpu manager/device plugin/...

https://sched.co/Nruo

A

I also have a partner to join me for this session. Mr. Looby, from.

B

A

I'm a Christmas offer two parts: one is kubernetes under now I'm a member of the communities. Another part is for secure computing, so today, I will focus on this purse, which is based on kubernetes and arm platformer. What are they potentially future applications? Also, we will know the needs of edge computing and also white communities fits more for this scenario.

C

So before we officially start I will make her serving all of you.

B

How many of the peers are you here? You work for IOT.

A

Related design and R&D raise your hands.

C

Thank you so, but before we up question, if there.

A

Is or how many of you use arm.

C

A

Or any other platforms to be integrated.

A

I owe t, do yours arm a platform or the other platforms.

D

A

Your stay on the other.

B

B

C

Deliver this thing, 6:07.

A

C

I will share with you about this again in terms the.

D

Operators, the edge computing.

A

Your 18 to purse one is the network side edge compute. Another.

C

A

Computer, so today we'll focus on the outside. In terms of you know, to architect your cloud and what a business won't age shoulder and how.

C

To reduce the cost of ego so see I will the on-site.

A

B

Edge computing in terms of.

C

A

Come across in our industry, actually this is very clear.

B

A

Corporate it with lots of enterprises.

B

A

They're, very smart city and smart beauties and also smart, fractious, a lot our business works or campuses. They actually don't want to enable.

C

B

Workflows to flow out.

A

Of their works.

C

And campuses and.

A

We cannot control so in terms of industrial scenarios. Their protocols are, my kin are.

B

A

Are independent or they private protocols, not the IP protocols, so these protocols are not suitable for purity.

C

B

To the exclude for.

C

B

So what should we do, how.

A

C

This capabilities.

A

Our industrial gateways to transfer the verticals.

D

And to connect the protocols and we.

A

Can take the AP flow, so this is a new cloud scenario and.

C

B

A

Distributed machines for better illustration: this is our first need and also low latency. A lot of people will talk about the deployment of 5g after this.

C

A

D

C

With the industry.

A

They have not very high requirement for latency but very high requirement and need you for.

C

The distance constant this.

A

At your home, we gave you a cable.

C

A

C

Video, the traffic.

A

And speech very sound and.

C

A

You can enjoy this, but if we offer the Internet service for the whole beauty there, sometimes the speed will be reduced. Broadband.

C

No problem, but the.

A

Speed has problems so, which means, if.

C

Nobody use it licensing very low, but if lots of people use it not easily.

A

Very high so can.

C

A

Comprehend this and it's Lowell Addison, yes, actually, it is in terms of industrial scenarios. What they require is.

C

A

The time within 10 milliseconds.

C

A

Milliseconds, your system will break down and they donate this distance.

C

So for the on-site devices.

A

If you won't manage them on-site for how to provide better.

C

In terms of standardization,.

A

C

We're promoting the TSM networks.

B

For the corresponding investigation.

A

And also the data volume and the bandwidth, so the data volume has exceeded a bit bandwidth and we have millions levels of the quantity of.

C

Devices and also the whatever network connections are not reliable and.

A

Very expensive, so in your home, sometimes we will buy this Moscow.

B

A

Server with a natural creature, HD.

D

B

That's all no need.

C

For Ronald.

A

Ron, it's on the operators, Network.

D

To play the video no.

C

Need to occupy the also take this.

A

From the customers.

B

Assuming we have deployed our services.

A

On the network side uniform edge computing, according to the pattern of the public vote, you need to to pay every month for connections and you need to buy the storage every month. You need to pay a bit money, 10.

B

C

A

Not sure, but you have to pay or you need to hey.

A

Master membership or you buy a device created at your home to provide very stable on video service. For you, lots of.

C

People are trying this.

A

For you, by a USB yourself.

C

The cost is 4050.

C

You can purchase.

A

The other devices so.

C

The users need, because of computing focus.

A

D

A

Instead of a technology.

B

A

On-Site will also show.

B

D

Value to serve.

C

A

Also in terms of information, safety relates to personal privacy, data and business data very sensitive and third-party apps and third party.

B

C

Mean designers now.

A

We have loss of a pilot's more than 90%.

B

A

C

In campuses, some perks, so China Mobile can provide us the.

B

D

Infrastructure but.

C

The traffic will not go out of their parks, so.

A

D

They're still using.

A

Their own private cloud, no.

D

Traffic or not.

A

They export it very.

B

A

In lots of vertical sectors, so this is their needs because in the verticals migration.

C

To cloud is very slow and also.

A

With the edge computing, we will come across a similar problem. Speed computing won't come across similar problems to the cloud computing in.

B

C

B

Security because.

C

A

C

Security promise are different to.

A

The private, if.

D

A

Them in fetch the Gateway or server services, their traffic would not be exported out to the park and they can be connected to the container, and this fits better for also.

B

A

It's for offline operation and on-premise.

D

A

Management administration.

D

A

Reliability because.

C

A

C

Has this self-made wanted it? We don't want to.

C

The breakouts, so we don't want to influence.

A

Segundo, maybe thousands.

D

Are for the IOT, my shrinks, otherwise,.

A

Our computing would have no so arch computing. Has this self-made.

D

C

Persistent lots of the people might have lots of solutions.

A

Can be used, but if the network come across problems because the reliability of the average is lower.

D

Than data center sorry.

C

A

I mean is through my relative in computing, cannot reach the reliability of the data center zone.

B

A

When we have the lower reliability of an engine,.

B

A

On-Premise, weather management user. Do you wait up war.

B

A

B

A

Are kind of weak.

D

Managed management.

A

Or EMM stration, we need to upload the data according to knees.

B

A

B

Kind of a weak connection.

C

With the cloud, another part is hardware.

B

A

B

A

Were Pyatt just launched the restroom pipe burst forth, very good, productive restroom I can draw arms container. This.

C

Guy from is from now the arm architecture is based.

A

In terms of IOT, it has the ecosystem and better than the other systems and lots of phone the wrong there.

B

A

Part is machine learning, Big, Data.

C

A

Have your confusions.

A

Is very limited, but you want to use it for big data processing. You based not like this, because in the eye and fit this, what we have reasoning and.

E

A

C

Terms of training for training, specifically, we students.

A

Think that it's reasonable to be put into the data center not to the edge and not seen, looks out of edge computing. Because of the.

C

B

Resources, so what can I do reasoning? We have the models.

A

B

We can input the amount of.

D

Into our phone, where.

A

The data center edge computing, so if we have the video streaming, we can use a model or show.

E

Process it directly and.

A

Then beckoned we can modify.

E

C

A

Of edge computing will have the AI and also.

C

A

Our phone, we have the identification services or the image identification relating.

B

Hiv I also here is.

A

A concern, for example,.

B

A

Course maybe my Deep Diver, not very clear, say we have a manufacturer from Hangzhou based.

C

In Hangzhou they work for K or the price is two thousand their camera prices, two thousand and they.

A

C

D

C

Service, which means when we excite data.

B

A

C

A

Anticipation or echinacea services for your face and the accuracy.

C

A

To 70 percent, because the cheap processing capability is very small without.

B

A

B

Processing they.

A

Need to migrate the traffic into the edge data.

D

C

Center, this is one mode. Another mode is.

A

B

C

A

Camera, maybe only 40.

B

Or 50 grand for.

A

The informational capture and the.

C

B

A

Data center in the future is the computing 5g. There would be loss of the traffic's, maybe.

B

A

What they should do is to rent the edge computer. You, the m4 CPU, to process the corresponding face recognition, information.

B

D

Means these are two.

A

B

A

In terms of the.

B

Manufacture I want to.

A

B

More and higher, but.

A

From the holistic so camera.

A

B

A

Lots of them a is on the camera. This.

D

C

The integrated cost and price of.

A

Our solution, because you buy a camera, the cost would be only 40 to 50 and if you integrate the eye applications on the front end, the camera would be 2,000 yen. But on the.

B

Road on the street in the intersection.

A

B

A

About 16 cameras, so the cost is very high if we offer this cloud service on the remote end, maybe.

D

The cost would be reduced because.

A

The payment method.

C

The charging method you.

A

Know is Satya so.

D

We won't take Carol, always and.

A

Now we need this infrastructure augmentation.

B

Which improves.

D

The capability.

A

Of edge the order to expand its ecosystem, this.

B

A

And edge computing, what kind of a connection.

C

A

Been forged in lots of the open-source projects you.

C

A

B

C

Project why we have this? Because, in the current given antis environment, all.

B

Notes are mouse.

B

A

Scale very small, maybe dozen or hundreds of lots.

B

C

A

Sure with you that we are responsible for maybe thousands of clusters, but actually there is no.

D

B

Quantity, very small, but on.

A

The edge they'll be anywhere, though.

B

A

I'll skip this part. The equivalent is architecture.

C

A

C

A

These devices work as.

C

A Korean I just knows.

B

You can see here there lots of devices and we have communities controllers.

A

Etc, but actually.

D

A

C

They're hard to stop.

D

A

Iot device: this is to launch, as I mentioned, before, lots of problems, including communication, so.

B

It will come across all this means.

A

D

These problems, we.

A

Can go through kubernetes to standardize these devices and.

C

To manage them all and.

A

To manage this as the Gateway to provide this to the customer next, my colleague looping will come to stage Thank You Dutchman.

E

My colleague mentioned that the smart cities and other.

B

Smart projects.

E

With IOT would bring us some challenges and one of them that Lucien CF can we manage so many IOT devices. So let's have an overall review of them. Well, actually it is already up to the industrial standard and we are thinking about using kubernetes to manage all of these infrared and Bluetooth devices, for example your Bluetooth speakers loudspeakers and infrared ACS to be all managed.

D

E

E

And we would but.

D

E

In each of these progresses, the corresponding mode use should still include network discovery, content.

C

D

E

Updated respective.

E

B

E

Jarrah mentioned here.

B

E

Have some selection to nice, the data not only on the API server on the masters.

E

So this is the current kubernetes architecture.

D

C

E

Maybe the couch.

D

C

E

D

E

Use kubernetes.

D

E

C

Tell them to you I'm a way how this slide.

E

Is about its main advantages.

E

Container orchestration.

B

D

E

Abstraction and about the dual.

E

Itil T cloud would be totally.

B

A

E

Traditional changes, so the first challenge would be a.

D

E

D

E

Gateway is compared to a server level hardware that were to have limited resources, for example, a vulnerable.

E

Which means in our.

D

Data we use that is to manage.

E

Our private route.

E

Machine of China Mobile, for example, in the community or neighborhood it might be.

E

Sure to communicate away, a certain could be master and a2 will be through in tonight by the latency to be severe, and it was a very highly trusted and it is not too reliable.

B

E

May cause a very long delay or even interruption and then the long.

E

E

In the IOT scenario,.

D

Because to be master is not reliable, it.

E

Is likely to be interrupted at a certain time.

E

B

E

Have a trimaran children come in attention and.

B

E

To disparage the primary uses of data with a new labels, an update together, this is about an economy and then about IOT device discovery. You know that in the kubernetes cluster it does not support IOT.

E

Device, you may know that show me devices into the Xiaomi system and it is similar. We could evolve a large number of infrared, GTI or Bluetooth devices into the gateway and the Gateway to descended as a primary data that could be master and then in the china, mobile machines could be master which could have an overall management of these millions of infrared and new tooth devices.

E

E

E

E

Them so we have.

D

E

Have the real-time, a synchronous, updating.

D

E

D

E

And protocol support to.

D

Support and it.

E

Also has our simplified device.

E

The overall framework for cube edge.

B

E

E

Discount, which means.

E

There is an H car to manage the application on the H nodes and to realize the runtime for the part.

E

It is a very lightweight equipment.

B

E

Is like the poor, all those so much so we need to pay attention to full things. The first.

E

E

E

E

Communication between.

E

Which is similar to.

E

E

Connection between the H node and the crowd is recovering, and then it could all be synchronized to the.

E

Synchronization is implemented, it is.

B

E

We used to have it is to synchronize the increment in the recent time period utilities the network flow, and the third would be about to the decentralized.

E

Make sure of communication between the.

E

The H nodes could be selected by their relative actually.

B

E

Could be synchronized.

E

Discovery of the devices for Bluetooth devices- it is okay, now.

B

E

For the time being,.

E

E

And you could download it and try it on the arm platform and then for the Cure.

D

C

Still need to further standards and features.

E

Like the unified management of the cloud edge.

D

C

The coffee mat and.

E

The more device recovery should be involved as well, and the roadmap is here for the link.

B

E

Edge core could completely simulate a slur function and undateable masters. You can.

C

See it's 100% IPR.

E

Compatible and it is not different at all from normal life.

B

I, do this and.

E

This edge call progress is very critical in the sloth.

E

Can I zoom in a so that you can see it more clearly.

B

B

Because all souls are.

D

You totally this.

E

C

E

Note we can even.

D

C

E

E

We're using a North Pole.

B

E

On arms, currently, now everything is ready for the next step. We.

B

Need to introduce new features so.

E

This is the end of my presentation.

B

E

Our demo any question.

E

Sorry, the interpreter didn't hear the question because there's.

B

E

B

E

Our thinking here is about the synchronization of the data metadata in by their increment.

E

It's also connected to the.

C

E

And it also has.

E

Actually, on the million level in order to change.

D

Wow you mentioned.

E

A

B

Want about the features.

E

Of H computing, you.

B

E

Be in the first- and it's mentioned.

D

B

Cata is ready on the arm and.

E

Tomorrow, I'm going to talk G wiser on arm.

A

I have another presentation on.

D

A

C

E

C

May contain some.

E

More sensitive data and.

C

D

E

Run the note or locally can you use.

E

In our group, you could initiate a 10-4.

B

E

Involving more people, it may involve the secure data discuss and have you compared H X.

D

E

With Gateway yet not yet.

C

Before we discuss this with.

A

Their operators, so kubernetes can work as a coordinator on the cloud, but Ajax is a kind of carrier on the edge on the edge.

A

Sorry I have used up my time so I can communication with the offline I personally think that to Cuba night is base for this. Thank you.