Today's topic is how to quickly create a batch resource mix, plan, and build systems using the Qingyun resource Orchestration Service, while talking about how resource orchestration can help us replicate a complete set of IT environments and how to make copies of the same architectural resources across regions.
What exactly is a resource orchestration?
As you know, cloud services are essentially a work of it architecture software and it platform intelligence.
The traditional way we build a system, whether it is medical, education or finance, media, etc., the top layer needs to deploy its own business code;
The middle layer is the professional software, platform middleware and service component of the business dependence;
The lower layer is the deployed hardware system, including compute nodes, storage resources, networks, and then the operating system, and so on.
The lower two layer is the operational dimension of the content, with the traditional IT architecture, the enterprise may need a lot of engineers, including on-site to do hard work, dirty, dirty live operations, development engineers.
Because some operations need to chew the hardware manual, understand the motherboard chip interface, router instructions and so on.
The intelligence provided by cloud service is to implement each layer in the way of software, to standardize the interface to provide, so many hard work, dirty do not have to do.
The flexible scheduling of software can solve the problem of hardware fault tolerance at the IaaS layer, and the standard components of the PaaS layer can eliminate the clustering, dynamic control node scale, backup, disaster tolerance and so on, which need to be considered and maintained by DevOps to build middleware.
Many people are already using cloud computing services, and many people do not choose to go to the cloud.
The difference is whether you believe that the Iaas/paas service provided by the cloud service can meet the requirements of the enterprise, whether it has stability, security, elasticity, enough performance and so on.
Let's say the resource orchestration.
What level of resources are organized? We generally say that the orchestration, including resource orchestration, service orchestration, the latter is the business level, the application level of the orchestration, based on the former.
Resource orchestration is the build layer above the IaaS layer, called Resource Orchestration. It builds systems on top of the IaaS platform and manages the lifecycle of the system.
Resource orchestration requires an abstract modeling capability: Each type of cloud resource has different devices, such as hosts, routers, load balancers, big data, and so on, and you need to model it abstractly-what resources are needed and how the topological relationships between resources are;
Then you can manage its lifecycle, how to create it from scratch, and then update, increase, reduce, and manipulate the resources.
In particular, there are several capabilities for resource orchestration:
The first is whether the existing resources can be abstracted, in accordance with our defined template specification model;
Second, based on the template to instantiate resources, become a set of independent resources, and can be the attributes of the resources, the original data to delete, search, recovery and other operations.
What can be done with resource orchestration?
The concept is abstract, so let's see what we can do about it.
The cloud service of Tsing Qingcloud is a concept that wants to build an ecosystem, with IaaS and Technical PaaS on the bottom.
When the offline communication was asked, could it be possible to provide Hadoop services on the Tsing Yun Platform for their users to use? Of course, you can make use of the existing IaaS and PaaS API interfaces.
To accomplish this, the traditional way to do this, you need to rely on the entire Hadoop cluster resources, such as routers, private networks, hosts, mirrors, hard disks, public IP, and so on, there is a topology concept and configuration planning, and then decided to order, call the API to complete the task.
If you use resource orchestration, you can create a one-time description of the configuration of the task and generate a template.
This is very convenient for our partners or for independent developers who provide services on this Tsing Qingcloud.
As long as the resource combination required to provide the service is configured as a template, the API is invoked once to generate the resource.
If not as an application provider, we list some usage scenarios.
The first scenario is the use of the cloud Platform for resource application approval management.
We have a number of traditional customers, his main account is as a resource approver, is the management department, sub-account as a resource application and user, is the business unit.
Each time a resource is needed, the business unit should apply to the management first. The application is based on business type and size.
For example, to do internal development systems, WEB systems and so on. Each business system actually has similar resource configurations and combinations, but the size is different.
In this scenario, the manager can book several sets of resources to orchestrate the configuration template, or write a real-time logic to generate business templates, receive scale, configuration, etc. as parameters. This requires the host, database, Hadoop, the need for resource networking, VPN login, these scattered resources can be combined to organize, one-time generation.
The second scenario is the replication of multiple sets of environments.
For example, there is only one small development environment at the beginning of development, testing, deploying Staging environment and online environment after going online.
For small-scale systems, each set of environments is relatively simple to rebuild. What if the resources are large in size? OPS engineers will be very tired to do very complicated work.
Resource orchestration has abstract ability, can be the existing resources, topological relationship and background all extracted, using the template to quickly generate resources, originally needed a large half a day, now only more than 10 minutes to be able to get it done.
The third scenario, no resources are used on the cloud, but you want to plan the current IT system.
This scenario usually knows how many host resources and database resources are required for a service to be on-line, but does not know how they should connect through the network, how to mount additional resources, what the price is, and how to calculate costs when scaling out.
You can use the Resource Orchestration feature to architect your system to see if it meets your requirements, including physical performance and configuration, and then see if the price budget is within your reach.
Fourth scenario, cross-region backup and deployment.
If you use our resources, you may encounter a situation: we have different areas of the computer room, physical resources are in the specific room.
If there is a system in Guangdong province 1, want to deploy in two places, Beijing 2 District to replicate a such structure, the same resource combination must be generated and connected in the Beijing 2 District, together.
Using resource orchestration, the resources of Guangdong 1 can be extracted into a globally unique abstract template, and then we apply this template in a new region.
Try to arrange the resources of the classmate, may think it is just a resource allocation, in fact, the things inside is quite complicated. There is a lot of work to be done about the relationship between resources, whether the constraints can be truly instantiated and managed.
First of all, we will have an abstract process: what resources, how their relationship, how to define their constraints, and then the embodiment of the entity, including routers, databases, caches and so on, these entities have the configuration, association, such as hard disk, network, etc., as its property modeling Finally, the topological structure is instantiated and the resources are really formed.
An open API
Each time we publish the service will provide the corresponding API, the API directly exposed to provide to the user.
Because the users of the service on the cloud, some partial technology, the use of the cloud platform to do a lot of things, including our partners to build the system, need to use the API programming, with software to form a rapid deployment;
The other is the end user, through the console, prefer to use a graphical mode of operation.
Resource Orchestration API is divided into three categories, one is Topologytemplate related APIs, to complete the maintenance of the topology template, that is, adding and deleting changes, as well as configuration, definition and so on;
The second is topology instantiation, the real batch of resources created;
Third, check whether the configuration is legitimate, can meet the quota, can establish resources.
How does resource orchestration work?
Having said so much, how do you use resource orchestration? On the console, resource orchestration can use a graphical operation to create a template.
We provide 4 ways to create a template: The system recommendation template is created as a sample, created from the current resource extraction topology, created manually from scratch, and created as a sample of the template you have created.
If you are not familiar with the Tsing-yun platform, or do not know much about the resource requirements of your business, you can start by creating a template that is recommended by your system.
Here the system recommends several common business needs of the Tsing-platform resource allocation and the relationship between resources.
such as simple Linux host cluster, public Network load Balancer portal cluster, database and cache cluster, big data cluster and so on.
By clicking on a different template, you can preview its topological relationships and see the basic configuration of each resource, as well as see the resource price for that combination. The recommended template here is our initial offer, based on user feedback and actual usage, we will give more common and getting started configurations.
If you want to extract the properties of the current resource as a configuration and further replicate the resource, you can extract the template from the current resource.
The extracted objects are entity resources under the user's name, such as hosts, routers/vpcs, load balancers, shared storage, databases, and big data platforms.
Hard disk, public IP, firewall these subordinate resources cannot be selected individually, but are extracted as the associated resources of the entity resources.
After extraction, you can continue to edit the template without affecting the extracted resources.
The template for resource orchestration is shared across regions, where users can extract a template from a region's resources and apply it to other regions to create a cross-region replication resource structure.
If the template relies on resources such as the own image that does not exist in the new zone, you can migrate the image first.
After you are familiar with resource orchestration features, creating templates manually is the most common way to create them.
In this case, the relationship between resources and resources can be generated from scratch based on the conceived resource combination.
The combination of resources follows the actual dependencies, such as the private network must be mounted in a router/VPC, the hard drive must be mounted on the host or Virtual San/vnas, and the public IP can only be bound to the host, router/VPC, or load balancer.
You can modify the basic properties and delete actions for the created template. You can also continue to create new templates based on the template.
You can find the button in the right-hand corner of the list of templates, or in the top right of the information area on the left side of the details page. Clicking this button will go to the template Editing page and import the template's configuration into the currently edited template.
The template editing interface is divided into 3 regions: Selection area, plotting area, configuration area, respectively, providing the function of resource selection, topology editing and configuration editing.
Resources support almost all IAAS/PAAS resources and components, which can be dragged to the drawing area by dragging.
Plot area realistic current template topology structure, you can add, delete, copy resources, zoom drag-and-drop paint, refresh, empty, view prices, etc.;
The configuration area can see the configuration of the selected resource and modify it. The relationship between template resources, such as port forwarding, load balancer backend, etc., can be directly displayed in the diagram by extracting the configuration.
Create a good template that you can see in the list of orchestration templates.
You can view the structure of a template, create a new profile based on a template, or apply a template to generate a complete set of resources with the same configuration
After the template is applied, the system begins to create resources by dependency. The process that you create can be seen in the real-time task in the upper-right corner of the console.
If you have more resources, or hierarchies and dependencies, it may take a few minutes or more, please be patient.
When you create a resource that fails with insufficient quotas, insufficient balances, or other causes, the entire creation process is rolled back, and the resources that are created periodically are recycled.
The above is an introduction to the definition, role, interface and console functionality of the resource orchestration.
Start with the functionality of the service orchestration, which we will provide based on resource orchestration.
In addition, the orchestration of resources, for users can be a business view, these views will provide new presentation and richer graphical interaction, so that different users easy to manage their own resources and business.
Technical Training | Resource orchestration, everyone can become an architect