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

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: KubeEdge - Kevin Wang, Huawei

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

KubeEdge - Kevin Wang, Huawei

KubeEdge is recently accepted as a CNCF sandbox project. It is accepted as a Kubernetes based and cloud native reference architecture for IOT/Edge computing. Since KubeEdge is opensourced, there are quite some companies, academics and community members who showed interests and started to contribute. In this talk, Kevin Wang from Huawei will share the overall architecture/design and specially demo KubeEdge’s real life usage. Besides, we’d like to invite participants in the room to share their usage cases, thoughts and suggestions on how to facilitate the contributions/collaborations.

https://sched.co/NrqV

A

B

They are intention, for the event is fun /, that's what the empty sir.

B

Please raise your hand if you were doing edge computing and how many of you heard of cube edge project for how many of you are using footage.

B

A

B

Like to give you a general introduction about this project and I hope, you can go to our website for an information.

A

B

I would like to start my introduction.

A

B

Actually I already asked you these questions, I'm sure you're interested in edge computing.

B

A

B

Control team is a very point domain up to.

B

Special benefits with edge computing, of course, there are also tiny news.

A

B

Not really stable between the edges, sometimes it's a public network or some private networks. I need a there's, limited bandwidth.

A

B

All the nodes and devices on edge.

B

In the stroke on screen, there's.

A

B

Gin, if generous hardware and.

A

B

And also the devices my edge is quite small, so resource constrained.

A

B

Your data request can be sent to to the edge and come.

A

B

You so you can achieve data locality and the security.

A

B

About from touching it's more secure.

B

Pieces of information to.

A

B

Much this year to is.

A

B

By CN CF as a sandbox project and we're including this project.

B

A

B

Achieve from synergy between cloud and edge, so our main idea is that edge.

A

B

The extension of cloud, a lot of vendors already have existing.

A

B

Strong service, our cloud, so if you can connect the cloud and the edge users can enjoy more comprehensive service and also enjoy doing first localize to serve a sandwich. Yesterday we released our coverage.

A

B

100-Person we.

A

B

Had four minor releases.

A

B

In our whitepaper there.

A

Is listed as a reference architecture now.

B

There are over 100.

A

B

In our project.

A

B

Should say over 100.

A

B

Have contributed to our project, there are also a lot of companies contributing including cloud companies in China, such as.

A

B

Mobile and also international companies, it you see an Intel.

A

B

Activity in the past six months later on, the left is quite small, but, generally speaking, it's quite active and the level of activity is increasing.

A

B

Version and the previous release in check used a bigger support.

A

B

Some devices- and I will give you more details.

A

B

Is our vision and mission, our goals.

A

B

I already mentioned.

A

B

Is the edge computing platform based on cuba notches.

A

B

Hope we can use.

A

B

My sister was he'll manage the nodes on the couch, so we hope we as can adopt the same method, to manage the Nazi distributed on the edge and as.

A

Well, as the applications on.

B

A

B

Well, as connected devices, we hope this kind of management is can be used across.

A

Different fields that.

B

Means including.

A

B

You also would like to manage KPI. It should follow the design style of.

A

Youtuber notches and.

B

We also need to support given artists.

A

B

In terms of first scale, we would support.

A

B

Number from notes under.

A

Devices, this is the latest village architecture.

B

So the the edge part is enlarged, the part of the crowd is.

A

A presentation.

B

A

B

Main controller is edge controller.

B

Changes shadow management.

A

B

Not only I judge them harder is scheduled to to the edge.

B

We used H, comma 13.

A

B

To write latest information of these applications to Fibonacci server.

A

B

Other say the part: you know it.

A

B

Their configure can be processed. You.

A

See the same message as on code in order.

B

To achieve the coordination.

A

Bit you can edit each.

B

A

B

Send a message to crowd: you have cloud and edge cop are based around WebSocket.

A

B

Know where you can create connection WebSocket, it's can achieve two-way.

A

A

B

They can be in a private network. You don't need a public IP as long as you can access the crowd hub. You can persist until they communicate to.

A

Us crowd, and it's always in terms of touch.

B

Device management.

A

B

A

The rest management or device.

B

A

Computing platform there's also.

B

Concept of exchange with.

A

B

I have a device, you can control the device through device to change and you can map it on the cloud.

B

We can change some.

A

Commands own device.

B

To change to control the edge device.

A

B

We achieved the mapper.

A

B

Of pepper because edge there, many different types of device and not.

A

B

Of scenarios mqtt, but other devices is a bluetooth on oil.

A

B

Ua, which is industrial control, protocol and.

A

Many other particles.

B

Device vendors may have new protocol.

B

A

B

Choose one of us.

A

B

Course MQTT is embedded.

B

Device vendor would like to add.

A

B

A

Can create a mapper so.

B

They can have for abstraction of the.

A

B

Now currently,.

A

B

Is communicating with event bus.

A

B

Event was is mqtt, client.

A

B

Can also communicate with the device.

B

So this is the implementation of humour after some analysis or find the native cubic net had to such a lot of.

A

B

A

B

Example, the realistic part is not drilling new data content and a lot of furniture in.

A

A

B

Also not necessary, so we decided to rewrite lightweight kubernetes.

A

B

Manager which consistently managed metadata.

A

B

Is so up to synchronization and metal manager right.

A

B

B

B

Notice doesn't have consistent consistency of metadata, so it's there's a break.

B

From the crowd or not be recovered, but after our consistency and even if England is connected to scenario after the notary's recovered, the port and other data can.

A

B

A

B

Tree is responsible for controlling the information of other devices and copy is saved.

B

Consistent storage, perfect.

A

B

Is to control to TP.

A

Some devices are.

B

A

B

Portal is exposed so.

A

B

This box, each part.

B

B

The overall memory is 10.

A

B

A

B

Consumption is not just this component, the MQTT broker and the darker supported in the first version.

B

Is consumption may be over.

A

B

Megabytes I know now we are doing the support for container the under. We will continue to support more lightweight time and we also introduced our key feature: that's edge mesh. It's originated around the concept of service.

A

B

A

B

To achieve the.

A

Data communication.

B

Capacity cross edge in.

A

The early Russia.

B

We could only access the IP portal or you are. We couldn't service.

A

Discovery etc.

A

B

Makes the edge mesh.

A

Kevin edges under.

B

A

B

1.00, we can't have.

A

Cross network edge-to-edge communication, so the.

B

1.0 version.

A

B

Connection or flexibility of the razor and edges.

B

A

B

To connect Roberts networks and also the communication, was dark out. It's.

A

B

To be a very important way of communication.

A

B

Saying 1.0 version.

B

Change an early version of coverage.

B

More supportive was a be given to.

A

That device to see.

B

The station no.

A

B

A strong computing power there was no.

A

Specific support in.

B

The couch architecture.

A

Each connections.

B

Directly connected to wizard.

A

B

So when the notation broken from the master.

C

B

No capability for dynamic.

C

They have the requirement for the scalability of applications is so in one panel version we introduced.

A

C

A

C

We have community master to the edge after that.

C

Club hub and ng hub is not that important anymore and.

A

C

Importantly, master went yet go down to edge.

A

C

Master is in different network of ways. Note: how can we connect it? So in the later versions, we will support us through edge mesh, because each mesh has just launched and edge site as also just a launch, and they have not connected yet.

A

A

C

Is traditional version, a transitional version, so at that time we use this kind of a mode, but the problem is darker: consume a lot of resources. So in the 1.0 version we add the CRI interface back.

A

C

We have the verification of continuity, integration, verification and now we are in the alpha version. You can have a try and they turn.

A

C

The darker tools, we.

A

C

Eliminated and all will be supported by CRI and later we will use our cRIO so in terms of a resource consumption. It's lighter.

A

C

A

C

Evaluation framework of.

A

What we have done for.

C

Example, we haven't released all the data yet before 1.0 version.

A

C

Have a framework for the performance- and this framework is pretty easy. We use two layers of kubernetes. The lower level is kubernetes, every port will were regarded as edge node to simulate all the nodes in the edge and another one is independently deployed, kubernetes and cube edge component on the cloud. So through this two layers, we can simulate a huge amount of edge, no deployment.

A

Actually, through.

C

Our evaluation on the design for the communication between cloud and into edge will use H hub.

A

C

They can do convergence for messages between nodes and the synchronize at our cloud, so it.

A

C

Different from Kaiba's is that the nodes connecting will be reduced and Anita functional criminality.

A

C

A

Communication data.

C

Volume will influence.

A

Its scalability long.

C

A

C

Ensure the same amount of connecting node and you were ensured that they'd have volume and to ensure the scalability. Second, if your communication package Spaulding cannot be changed, then you ensure the long.

A

Connection you.

C

Can also ensure availability.

A

C

I think for edge.

A

C

To native kubernetes it will have a better scalability for.

A

This framework, it.

C

Will validate our hypothesis.

A

C

We have some people from the community, they have tested 4000 no, but it's only.

A

C

Terms of performances in.

A

C

Of communication between cloud and edge, we are using web sockets between cloud hub and edge hopper. We are using WebSocket.

A

C

We are studying and quit protocol.

A

C

Have version based on creative, but the performance is on evaluation and in terms of performance it will improve a little bit.

A

C

The application scenario.

A

C

A

C

A

C

Cool part of Poway cloud computation and open make it open source apart from good edge. We have the power cloud service and the application is in a smart park. The main demand from the customer is that they want to do the facial recognition and they.

A

C

The people's ID and.

A

C

The gate they want to monitor the traffic of the park as well and they for the camera. They do not use smart camera, they are using traditional IP camera. You use IP address. You can read.

A

C

A

C

Overall application is pretty simple, so.

A

C

Use facial recognition.

A

C

Traffic flow and synchronize data to add change, as we will know at edge. An original video flow have a big data volume if you have a application on edge.

A

C

The video frames you only capture a small picture of the head side, you make a comparison and the comparison is based on your business design and whether you want to build a big data cloud match, or you just added make it on the edge. In the initial scenario, the client.

A

C

Use the central big data model in cloud, so the official recognition is on clouds and for the edge they will.

A

C

A picture of head size from the video and upload to the cope, a cloud.

A

C

On the head image and they can do the face detection.

A

C

A

C

A case for CDN customer, they are it's not using each side and the user. The main city n is for ordering of videos generally. The case is to send video from central cloud to the edge and we slice the video. When you watch the video. Actually, you can see you can adjust. You can only change.

C

The pace, so you need to do slicing, and then you need to change the ship to the code and now maybe you're coding may be different from the original video, and you also need to provide hello media.

A

C

A

C

Make it directly connect to the cloud and then it can also okay, run video transcoding, rendering slicing at the edge managed as job and.

A

A

C

Take a look at the roadmap of the project. Actually, we started from November 2018 and we announced the open source of this project. We provide a lightweight. We are lightweight agent at edge and.

A

C

Past of versions, we provide support for the edge equipment.

A

C

Also support the synchronization of the state data synchronization of state.

A

C

We also have a mesh layer for the communication between cloud and an edge, and we also support CRI and continuity integration at 1.0 version. We need together do CSI, because in the current version we deleted a lot of storage type of support.

C

In the edge do not have a WS.

A

C

For the clients we met, they still need the demand of where it's eternal storage.

A

C

A

C

To have cross edge data sharing.

A

C

Full 1.0 version we already have done CSI.

A

C

Storage compatibility, so we did not have enough had come to work on for the work on it, so it will be done later.

C

We also have some remote control functions so that you can monitor the edge status and we will further optimize the openness at the community. Also, we have over 100 compute a commit contributors, but through previous survey, people think that to become a contributor, there is some threshold and we were in the guidance. All files show that people can contribute to it due to time limit previously. I planned a demo, but I will not do the demo here, but you can go to our booth and take a look and for the demo.

A

C

Demo is pretty simple.

A

C

To achieve the control of the equipment or the communication was implemented, and here we chat is just a photo to enter and it's mainly on the cloud and we have a web server through a web server. We, you can select a song.

A

C

A

C

The song name at capacitor.

C

Synchronize, the equipment updated to the edge and then we will connect to a Bluetooth, sound amplifier.

A

C

Topic message from the subscription and play music by the track.

A

C

Just to show you how to add the definition of equipment and how to update the equipment status and for the demo, you can go to our PI ways. Booths and I will not.

A

Show it here and.

C

The last part is about the basic information we have our community, you can see our website and we have slept Channel and every two weeks we will have a meeting and we will upload the meeting video to the YouTube and for domestic or Chinese.

A

C

You can scan the QR code and you can join a group and we can have a further communication in.

A

C

Any questions from the floor.

C

You choose to build HD.

A

C

Arthur, after evaluation, we think if we want to simplify or remove some cupola, and then we need to cut a lot of codes, so we decide to write by ourselves. My Chinese is not so good. I will speak in English. Welcome to NEMA. Your visual chose a home in routine, oh hi, yo, Ian, X, the panthan miracle George M corner since Idaho super shop.

A

C

Can kind of England around her? She chose a poco de.

C

Church matrix API Rocco, Nico TN, they say you sell veneno kubernetes.

A

C

Keep attaching Mao died while we bring him in sugar, warm Andy Mineo. You go Sam whoa, damn Porsche Cayenne, double Michelle, but house without her younger son, when she.

C

Sorry he's not using microphone. The interpreter cannot hear I.

A

C

Repeat his question: first, the first question is about hatch, edge, notes, management of how to provision act, notes actually for the crew edge projects. We do not do it fully because in H ecosystem, some other projects are taking care of. Currently, our starting point is that the edge notes focus on OS and a good edge release. Pack, we have an installer. It can help you to install the components in the edge node for adding notes you can edit manually, you can register on cloud and create note object.

A

C

A

C

You filling the cloud hub information in a for security.

A

C

Have very fundamental certificate-based, a security.

C

C

Security design we have prototype. Now we have some discussion in the community. The main idea is that we will use.

C

Its back in cloud to issue certificate. So when your cloud is disconnected to with edge, they are on the same certificate chain, they can still do the validation certificate, but when it is disconnected, you cannot issue no new certificate, but at the edge you do can do the certificate authorization and the validation.

C

In its extending from Center to edge so for at you just do the simple decision.

C

Would you extend to the controlling of external control.

B

Under the actual status can.

A

B

Reported back so.

A

B

Risk commendation: if you change some field value, there may be some problem.

B

It's a 2002 change.

A

B

I, we know that being an.

B

Inch on your teaching weekly to.

A

B

Jellybeans, which, in.

A

B

So you could touch it. I want me to kill your H Tibet.

A

B

A

B

Table yes, she don't do not take English when she should die. Grenadian mean she logical. You know functional she's, opinionated,.