Shulou(Shulou.com)05/31 Report--

This article focuses on "Application case study based on Kubernetes and OpenStack". Interested friends may wish to have a look. The method introduced in this paper is simple, fast and practical. Let's let the editor take you to learn "Application case study based on Kubernetes and OpenStack".

Open source software

The whole platform is based on existing open source solutions, and the platforms we want to use include: OpenStack,Kubernetes,Docker,OpenContrail and so on.

HW and independent vendors

No supplier is positioned to include both software and hardware. The Internet of things gateway CPU is either x86 plus 64 architecture or ARM architecture. We don't want to be locked in to suppliers who use expensive specialized equipment.

Interoperable

The Internet of things platform is universally available for a variety of uses. For example, Internet of things gateways can also be used in street lamps for computing objects, just as they are used in smart factories or industrial 4.0 applications. Therefore, we designed the following high-level architecture, which uses open source projects, OpenStack,Kubernetes,OpenContrail and Docker.

Sensor

Any sensor used to read environmental data (temperature, humidity, carbon dioxide levels), objects, counts, etc.

IQRF network

IQRF is a powerful wireless grid technology for operating sub-gigahertz ISM bands. Can be easily integrated with the sensor.

Gateway

At the Internet of things gateway, the IQRF coordinator processes the data from the sensor. These can be any x86 Docker,Kubernetes,OpenContrail vRouter 64 or ARM architecture, which is based on Docker,Kubernetes,OpenContrail vRouter and Debian and runs the system in terms of pre-built images.

The network

The gateway can connect over any network (GSM/LTE/WiFi) because SDN creates a dynamic L3VPN between the Docker service and the data center.

Data center

The data center includes OpenStack and Kubernetes control panels. Both orchestration solutions use OpenContrail as a single SDN (Software defined Network). This allows the entire platform to be managed from a central point. We can open any Docker container locally and remotely, and then create a service that dynamically connects to the OpenStack. OpenStack cloud container data storage and big data process services.

Virtualization / API acc

All data has been virtualized on the network port and can be accessed through REST API service.

Next, let's look at two cases.

Case 1: smart City Smart City

The first case is the SmartCity project in the city of Pisek, Czech Republic. The SmartCity concept and architecture will deploy more than 3000 endpoints and approximately 300 Internet of things gateways that operate in highly available mode in Kubernetes-driven containers. Part of the solution is an open data portal, where data API is available to third-party companies that provide information:

Traffic, path, parking

Monitoring, management, energy saving

E-commerce, market, tourism information

Environmental analysis

Lifestyle, social services, social networks

The target solution uses the Internet of things gateway based on raspberry pie 2 as the Internet of things gateway service. The gateway's data is stored in Graphite and processed by custom data mining applications, and the results have been displayed in the LeonardoCMS-based Urban Citizen Portal, a web service that allows complex visualization and arbitrary content. This open data portal enables data to be accessed through visual instrumentation or API.

The following screen shows a simple Kollarova X Zizkova output from an intersection with vehicles and pedestrians for a specific period of time.

Case 2: smart meeting at Austin OpenStack Summit

To prove that our IoT platform is independent in the application environment, we took an IoT gateway from the Smart City project and then put it into the Austin Convention and Exhibition Center at the OpenStack Summit together with sensors connected to the IQRF mesh network (measuring humidity, temperature and carbon dioxide levels). This illustrates the ability of the Internet of things gateway to manage and collect information, which can be obtained from any Linux-based communication technology such as IQRF,Bluetooth,GPIO.

We used a single, active Internet of things gateway to deploy 20 sensors and 20 routers on the third floor of the conference to receive data from the entire IQRF mesh network and transmit it to a dedicated time series database, in this case Graphite. The collector runs on the Docker container MQQT-Java bridge managed by Kubernetes. The most interesting are the Docker containers running on the raspberry pie and the virtual machines running in the Europe data center. OpenContrail SDN provides dynamic network coverage tunnels.

The following picture shows a single wireless IQRF mesh network sensor and router discovery.

IQRF is a wireless mesh network technology operating in the sub-gigahertz band. It provides very simple integration, product interoperability, a robust mesh network with a maximum of 240 hops, a range of up to several hundred meters, and ultra-low power operation.

The screenshot below shows the CO2 values at different times from different rooms on the second floor. The historical chart shows what happened on Monday.

From the collected Austin data, the following figure covers services.

Technical overview

The so-called "Internet of things platform" is created from a general perspective, which is based on collection, management and secure, dynamic handling of thousands of endpoints for centralized management. Therefore, the architecture is divided into two main parts:

1 data Center

The data center is the focus of the management of the whole platform of the Internet of things. The OpenStack IaaS cloud runs with the virtual machine on the SDN dashboard. These machines include time series storage, processing data cluster, data API proxy service, virtualized web service and so on.

2 Gateway

Internet of things gateways are located in target places such as street lamps, industrial equipment and household appliances. SDN provides the transport layer to connect remote Internet of things gateways with cloud services. Gateways can be multi-platform and, if possible, mix x86GP64 with ARM devices. Multiple sensor platforms are processed for multiple customers in a single gateway due to micro-service segmentation (Docker container) and Kubernetes multi-tenant support. This platform can provide scalable multi-tenant space.

The following chart shows the data center layer and components in terms of gateways. The details of this chapter show more information.

Detail chart

The detail chart provides a local architectural view of the entire Internet of things platform. The data center is shown on the left and the gateway mentioned earlier is shown on the right.

As you can see below, OpenStack is used as a cloud to handle all control service, as well as big data processing, and the virtual units mentioned earlier. The Kubernetes at the gateway is used for micro-segmentation of services, which is necessary for security between multi-tenancy and different sensors.

OpenContrail is used to connect the two sides and provides network segmentation between Kubernetes PODs and OpenStack project virtual machines.

As mentioned above, the SDN overlay has been segmented. What matters is that there is only an IP connection between the data center border router and the Internet of things gateway. The bottom layer is the VPN between the gateway OS and the data center boundary router, where the OpenContrail can communicate directly between the virtual machine (OpenStack cloud) and the container (gateway). This approach allows you to choose from different sensors and actuators, give them privileges, and securely connect to the cloud within the processing application.

The data center includes the following services:

Manage services

The HW cluster uses all the control service to run virtual machines: OpenStack controller,OpenContrail controller (SDN), Kubernetes master,SaltMaster.

OpenStack cloud

The OpenStack project provides segmentation for different virtual machines, including database (graphite,influxdb,openTSDB), big data processing (Hadoop), and data virtualization (Grafana,LeonardoCMS). It runs on the KVM super manager and uses the OpenContrail neutron plug-in for the network.

Border router

OpenContrail creates iBGP,iBGP peer-to-peer with the data center border router, where dynamic network routes can be propagated from the OpenStack virtual machine and Kubernetes pods on the Internet of things gateway. It creates the standard L3VPN core MPLSoverGRE or MPLSoverUDP.

The components of the remote gateway include:

Kubernetes Minion

Kubernetes minion communicates with Kubernetes master in the data center and manages PODs through kubelet. Kubelet uses the opencontrail plug-in, which connects to the Docker container using a vRouter proxy.

Kubernetes PODs

The Kubernetes PODs is connected to one or more container groups of the vRouter. PODs is classified by tags. This allows you to open different applications that can be read from message buses like IQRF,Bluetooth or GPIO.

Docker container

The Docker container in Kubernetes PODs brings the huge benefit of an easy operating system without any special devices. For example, IQRF uses a specific version of a simple Java application that can be delivered in a container in minutes without causing a mismatch with the operating system gateway.

Application View

The following figure provides the mode of applying the view. It shows how the OpenStack cloud inside the virtual machine can contact the Docker container on any geographic location L2 or L3, thanks to the coverage of OpenContrail. As a result, application developers can use the same tools as they do in the standard cloud.

For example, we collect data from environmental sensors. The sensor is directly connected to the container, the data is processed in the container, and then sent to the Graphite time series database. Because we want to display the data vividly in real time, we use Leonardo CMS that reads the GraphiyeAPI interface to use another virtual machine and display the data on the web page. Based on this, we can create different projects based on the same guidelines in the same cloud with multiple inputs and outputs.

At this point, I believe you have a deeper understanding of the "application case study based on Kubernetes and OpenStack". You might as well do it in practice. Here is the website, more related content can enter the relevant channels to inquire, follow us, continue to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.