The author | Hao Shuwei

Flexera’s RightScale2021 cloud status report states that 92% of large enterprises are adopting hybrid cloud strategies. Gartner also said in a report that 90% of medium and large enterprises will use hybrid cloud architecture to manage their infrastructure in the future.

Over the years, with the rapid development and landing of cloud computing technology, more and more enterprises choose to use cloud computing technology to help them quickly complete the digital transformation of their business, so as to better adapt to market changes and win more market space. Among them, a large number of enterprises choose to deploy their business in the cloud for the reasons of reducing the cost of technology development and operation and maintenance, and enjoying instant services anytime and anywhere. Some enterprises choose to build their own proprietary cloud platform in their own internal data center environment due to data sovereignty and security and privacy considerations. Enterprise users with requirements for both public and private clouds will choose to build hybrid cloud architectures.

Why a hybrid cloud architecture

Enterprise business security considerations

For enterprise users, especially large ones, there are risks in completely entrusting the company’s critical “lifeline” business to an external cloud vendor. Although public cloud vendors generally provide secure and reliable redundancy schemes to ensure uninterrupted service for enterprise users, it is not without incident. The hybrid cloud solution ensures that enterprise users can choose and switch between A and B solutions at the same time, maximizing service stability.

Regulatory requirements for data sovereignty and secure privacy

Some laws and regulations or the company’s own security strategy for the enterprise data stored or hosted sites have rigid requirements, such as the eu’s “universal” Data Protection Act (GDRP) for the data control and data processing of the digital supervision measures, such as the policy for enterprise data can reside in a designated place, the purpose is to protect the data privacy and security, and so on. Hybrid cloud architectures can help enterprise users meet these needs.

Enjoy the service features of cloud vendors

There are differences in service quality between local cloud and public cloud vendors. These differences are reflected in various aspects and depend on the actual needs and considerations of users. The differences of geographical coverage, for example, users usually in the local cloud services, cloud vendors in a specific area to provide better service in the access latency, enterprise users in the region have important client and access latency for cloud services have higher requirements, enterprise users can choose the area of business deployment in a public cloud, Other services continue to be deployed on the local cloud.

Cost optimization

Local cloud infrastructure lacks flexible capacity expansion and scaling. It cannot reasonably arrange basic computing resources during service peaks and downturns, resulting in resource waste and cost increase to a large extent. The flexible, agile and on-demand capacity expansion feature on the cloud can compensate for this defect.

Follow technological innovation

For similar to some artificial intelligence, machine learning, the Internet of things, the evolvement of the technology innovation, and advanced technologies such as cloud vendors usually can provide and at the same time of the first cloud services, enterprise users can use these cloud services at smaller cost, and promote the technological innovation and development of enterprise itself, a hybrid cloud architecture can make enterprises adopt the best cloud services anytime and anywhere.

How does cloud native help hybrid cloud architecture evolve

Public and local cloud itself is two different cloud, they have different infrastructure, the ability of different features and different API interface, to build a hybrid cloud architecture, on the one hand a cloud provider cost a lot of energy on the adapter and the ability to integrate cloud platform, on the other hand, users under this architecture cannot switch real on-demand cloud service providers, Instead, it’s another form of binding. The defects of traditional hybrid cloud lead to the failure of this cloud architecture to form a standardized ecosystem, which is also the reason why we have been unable to build unified management and unified delivery for this cloud architecture.

The emergence of Kubernetes makes hybrid cloud architecture enter the 2.0 era. Multiple features of Kubernetes and its related ecosystem provide possibilities for the standardization of hybrid cloud:

  • Cloud native technologies represented by Kubernetes shield the differences in infrastructure. At present, various cloud vendors and a large number of data centers have implemented these technologies, making it possible to apply “once defined, everywhere deployed”.

  • Kubernetes’ standardized, declarative API simplifies application deployment, makes application delivery more standardized and unified, and enables applications to be described and choreograph in the same way across different clouds

  • Grid service technology can span multiple Kubernetes clusters to achieve unified flow management and service governance, so that the application services under the hybrid cloud architecture can be unified to a control plane for management.

In the era of cloud native, cloud native technology represented by Kubernetes has promoted the arrival of application-centered hybrid cloud architecture, and Kubernetes has become the de facto basis of enterprise multi-cluster management.

Typical usage scenarios of cloud native hybrid cloud multi-cluster

Live in different places — Cross-region disaster recovery

From the perspective of infrastructure services and Kubernetes container platform, users can build a high availability application business architecture at low cost, but some services that require higher DISASTER recovery capability need to be realized by regional disaster recovery capability such as remote multi-live.

You can set up multiple clusters in different areas of a single cloud vendor or in different areas of offline IDCs and online cloud vendors to implement remote multi-active deployment of service applications. The following figure shows the active-active deployment of container clusters in the IDC and container clusters on the public cloud in the hybrid cloud scenario. In the remote active-active architecture, application loads are deployed on multiple clusters at the same time, and a global DNS service is used to forward requests to corresponding back-end clusters. When one of the clusters fails to process a request, the DNS service automatically processes the request and forwards it only to the healthy cluster.

 

Low latency – proximity access

For users carrying out global international business, the visitors of the service are widely distributed. If the server is deployed in a specific region, it will inevitably cause poor network experience in other parts of the region.

In this scenario, we can choose to deploy clusters in multiple regions and forward user requests to the nearest cluster for processing through intelligent DNS resolution, minimizing network latency. For example, in the figure below, an application service is deployed in Kubernetes cluster in Beijing, Chengdu and Hong Kong respectively. User requests from North China will be intelligently resolved to Kubernetes cluster in Beijing, and user requests from southwest China will be intelligently resolved to Kubernetes cluster in Chengdu. Requests from overseas users are intelligently parsed to the Kubernetes cluster in Hong Kong, which minimizes network latency due to geographical distance and provides a consistent service experience for users everywhere.

 

Reduced blast radius

In general, multiple small clusters are more easily isolated from failures than one large cluster. A cluster may fail to process requests due to disk or network faults. In this case, multiple clusters can be used to isolate faults in a cluster to avoid a chain reaction.

Business isolation

Different services usually need to do a good job of business isolation, although Kubernetes itself has a namespace mechanism to help users do security isolation, but this is only logical soft isolation, different namespace can still communicate with each other, and there are also resource preemption problems. You need to configure network isolation policies and resource quotas.

Different services can be deployed in different Kubernetes clusters to achieve complete physical service isolation, with higher security and reliability than using namespace isolation. For example, separate clusters are deployed for different departments within the enterprise, and multiple clusters are deployed for development/test/production environments.

summary

Cloud has become a general trend. Some enterprise customers will adopt hybrid cloud architecture based on data sovereignty and security and privacy considerations. In addition, some enterprise customers will choose the hybrid Cloud Plus multiple cluster architecture based on data sovereignty, cost optimization and geographical coverage. Hybrid cloud and multi-cluster architectures have become the new normal for clouds on the enterprise.

If you want to know more about enterprise IT solutions, application scenarios and usage modes under the fusion of cloud native and hybrid cloud, you are welcome to click to read the original text for understanding, or nail the qr code to join the “Ali Cloud Container Service ACK User Group” :

 

The selection of welfare

  • What do you think are the biggest challenges facing modernizing IT?
  • What are your concerns about cloud biotechnology?
  • What scenarios do you expect cloud native technology to apply to hybrid cloud architectures, and why?

If you want any of the above problems, published his views, before 18:00 on September 17, welcome you and everyone study published in the message area, alibaba cloud native public number will select the message before the thumb up three, each sent ali cloud container services technology expert Hao Shuwei book “cloudy and hybrid cloud: the cloud more than native cluster and application management. Looking forward to your communication!

Author’s brief introduction

Hao Shuwei, ali Cloud container service technical expert, core member of cloud native distributed cloud team, focuses on cloud native multi-cluster unified management and scheduling, hybrid cluster, application delivery and migration and other cloud native technologies. _ Click to read the original article to understand aliyun container service ACK Mixed Swarm User Guide

Click the link below to learn about Aliyun container service “ACK Mixed Swarm User Guide” help.aliyun.com/document_de…