Introduction: Cloud native technology is not only technology, but more importantly, cloud native technology needs to be combined with cloud computing to help users build applications with cloud native architecture.

On November 26, 2021, ali Cloud User Group (AUG) phase 3 activity was successfully held in Guangzhou. Li Guoqiang, a senior expert of Aliyunyunnative with rich experience in container, micro-service and other fields, shared the trend direction of aliyunyunnative and the capacity layout of aliyunyunnative to dozens of Guangzhou enterprises. This article is based on the author’s speech.

Good afternoon, everyone! Welcome to this afternoon’s forum. As I said before, the protagonist of today is everyone here. We share with you in the hope of playing a role of casting off a brick to attract jade.

Definition of cloud native

Cloud native community definition

Pivotal was the first to come up with the concept of cloud native, which he defined as DevOps, CI/CD, microservices, and Container.

Technically speaking, it is basically correct, but it is the CNCF Foundation initiated by Google that really promotes cloud native. Today CNCF has more than 1,000 projects, and let’s take a look at what these projects are trying to do for users.

CNCF defines cloud native as helping users build resilient applications, and mentions a series of representative technologies: container services, grids, microservices, immutable infrastructure and declarative apis. I’m sure you’ve all heard these technical terms.

So what does all this technology do? A good summary is to help users build systems that are fault-tolerant, easy to manage, easy to observe, loosely-coupled. These are all key words, and that’s pretty much what users are looking for when they’re building applications or when they’re building systems.

In particular, new Internet applications will face a variety of technical and market challenges, such as the impact of large traffic, malicious attacks, fast online promotion activities, etc. Under these challenges, customers expect their software or systems to be highly fault tolerant, easy to manage, loosely coupled and easy to observe, and so on. At the same time, the business value of these features is that they help organizations make major and more frequent and predictable changes. Together, these are today’s definition of cloud native.

Cloud origination derived from clouds

Is it cloud native to use these technologies? In this year’s Cloud Computing Conference, a new definition of cloud native was extended. Cloud native is not only these technologies, but more importantly, cloud native technology needs to be combined with cloud computing to help users build applications of cloud native architecture.

As can be seen in the figure above, there are a series of cloud native technologies on the left. When combined with cloud computing, it will produce a series of cloud native products, including container services we mentioned before. As the first project of CNCF, K8s must be the core of the whole cloud native system today. It also includes middleware, database and security around cloud native. Today, they all operate in accordance with the mode of cloud native, which uses cloud native technology to help users build applications and systems in line with the definition of cloud native.

Today, ali Cloud has a large number of cloud native products. Enterprise actually in use and be born to a real scene, will send these products to form a a solution, such as live solutions, AI platform, cloud, a unified scheduling and so on a series of elasticity, the above products to build the whole of this is our for cloud native from technology to products to the solution of the system.

Cloud native trends

Cloud native opens the era of all-cloud development

There are a few trends to share with you, you must be curious, today’s cloud native in the industry and enterprise inside what stage? If I started using cloud native today, would I be the one to do it, or would there be a lot of people using cloud native today?

According to some industry analysis reports, containers are widely used today. Up to now, 68% of enterprises are using containers in production environment. Of course, not all of them are core systems, but two-thirds of enterprises are using containers in production, so containers are very mature.

More than 80% of users are using or planning to use microservices. This is also a very big trend, and its usage is higher than that of containers. But it is not to say that all businesses need containers and microservices, it is an application architecture choice, but it is more and more popular. Another is Serverless technology, which is used by 25% of developers. I will briefly introduce Serverless later, which is also growing in maturity.

Today I’m going to share a few key trends with you, and throw a brick in the door, so you can think about how these areas fit into your current business.

Distributed cloud has become a new trend

First, distributed cloud has become a new trend, and distributed cloud has been one of Gartner’s top ten technology trends for two consecutive years. Today, more and more enterprises including manufacturers are talking about distributed cloud, what is behind it?

In fact, it is the change of business that brings the change of cloud form and presents new challenges to technology. Today, various cloud vendors, such as Ali for example, in addition to public cloud, there are local cloud, edge cloud, including to help users build private cloud form inside IDC, so there are more and more cloud forms.

Aliyun puts forward the new concept of “one cloud and multiple forms”. Cloud not only refers to public cloud, but also includes multiple forms. So why does one cloud have many forms? Because more and more business scenarios today require such polymorphisms. Today, edge-side video technology is becoming more and more developed. Live broadcasting, VR and AI services require data and computing power to appear on the edge, which promotes the development of edge cloud.

Second, with the development of IDC and public cloud, many enterprises may hold more than one cloud, which is also a business demand. For example, enterprises want to build a multi-active high availability architecture across multiple computer rooms or clouds. Customers’ offline IDC hope to fully use the ability of public cloud, which gives rise to the emergence of one cloud and multiple forms. However, the emergence of a cloud with multiple forms will also bring great complexity. There are certain heterogeneity among these clouds, so how to manage the business availability system on the cloud is a common demand of enterprises.

For example, how can the customer establish an active/standby relationship or establish a active-active system between IDC and public cloud? Suppose my business was mainly in IDC before, but IDC might have problems. Can I establish an active/standby environment on public cloud? There is another situation, for example, I used to have a business in IDC, but it may not be able to move all of it to the cloud today. Can I play it to the cloud?

These are the more and more scenarios that can be considered in today’s multi-platform business integration under one cloud. This is also the focus of our discussion today, and we’ll talk about it in more detail later.

AI load cloud biogenesis

Another very important trend is ai-loaded cloud protogenics. Why does AI load become more and more primitive? There are several important reasons for this. First came big data, then data-based deep learning, and then the integration of AI loads with containers and K8s ecosystems.

In the field of big data, GPU is widely used, but the traditional scheduling system is weak, and K8s gradually assumes the management ability of more and more important AI load. New computing engines such as Spark, Flink and the K8s ecosystem are so well integrated that more and more enterprises are naturally putting AI-type loads on the K8s ecosystem, and the K8s ecosystem is being promoted by it, and this technology is developing very fast.

We see a trend. In Aliyun today, the fastest growing business of many customers is AI. On the one hand, it is the need of business itself, and data needs to be analyzed after accumulation.

However, AI cloud bioengineering also faces many challenges, such as how to pool and share GPU, scheduling ability and data access efficiency. These are all challenges that AI will face after being put into K8s. There will also be many corresponding solutions, some of which may be open source and some provided by Ali Cloud. Help users to complete the AI in the original biochemical process encountered a variety of problems.

Cost reduction and efficiency increase is the universal appeal of all enterprises –FinOps

The third topic is FinOps. Cost reduction and efficiency increase are common demands of all enterprises. Even if business growth is very important for an enterprise in its initial stage, cost reduction and efficiency increase are still very important for IT department or budget department.

There is a very simple schematic diagram of how to reduce cost and increase efficiency or understand the FinOps concept. There are three important steps: capacity cost planning, capacity cost visualization and capacity cost optimization. I believe many enterprises present here have done a lot in two aspects, namely capacity planning and optimization, but capacity cost visualization is also very critical.

This includes how observability tools can be used to see very accurately the utilization of the system, including whether there is redundancy for capacity optimization. There is also a very important direction, on the public cloud, whether the utilization and cost of cloud resources are directly displayed in a graphical way. This is actually a new direction in the field of FinOps today, cost visualization. Allows you to see the utilization of each department cluster and its relationship to billing.

This view is very much needed from both a financial person’s and a manager’s point of view. Each team looking at what’s going on from an expense and utilization perspective, combined with a cost trend analysis, can be very helpful for IT and finance to come up with effective cost optimization strategies.

Capacity cost optimization, there are so many in the field of cloud native technology, such as the ability of elasticity, before there is a saying, cloud is one of the most important characteristic of elasticity, all kinds of elastic, such as flexibility, elasticity, fine scheduling nodes, etc., all of these can go to explore and think about how to optimize your costs, the premise is you have visual ability to see these costs. This is also a hot topic for many enterprises this year, that is, how to achieve cost visualization and optimization.

Cloud native core technology

Containerization + microservice = good partner

There are already a lot of users. However, many enterprises encounter many problems when using microservices today. Container + microservices is a very good practice. Microservices provide a lot of value, but they also introduce complexity, including complexity of links, complexity of operations and maintenance, and how to quickly locate problems when they occur.

The container is designed to run online microservices from the beginning, so whether it is the definition of Service, high availability, scheduling capabilities, resilience capabilities are actually a good partner with microservices. At the beginning of K8s, the first wave of business type was micro-service. Micro-service system centered on K8s platform, which required a series of supporting capabilities.

Microservices are not free lunch. Besides business benefits, you also need to do a lot of extra things for microservices, such as service, registration, discovery and configuration of the center high availability, service governance, how to ensure the control of offline traffic, gateway layer design, log monitoring and so on. These are all things that enterprises need to consider when they decide to run their microservices on containers. In fact, there is a complete system in Aliyun today. Think about what you’ve already done and what you need to strengthen when you run microservices on containers today.

All on Serverless

The last trend to share is Serverless. The number of enterprises using Serverless is growing very fast. Today, many people equate Serverless with functional technology, which is not completely correct. Today’s Serverless is actually an architectural concept, not a product.

If we look at the details, Serverless has many forms, it has the Serverless facing function, on the most common, I write a piece of function code can run. The other one is application-oriented. Many enterprises cannot accept the new programming model. They hope that the existing applications can also run in the form of Serverless and enjoy the technical dividend brought by Serverless.

Today, in addition to function-oriented Serverless, there are application-oriented, K8S-oriented, container-oriented Serverless instances. More and more enterprises will choose different types of Serverless to do business. You can think about this. Is it possible to use Serverless for some business? The benefits are flexible cost savings, separation of development and operation and maintenance concerns, so that developers can only focus on development, and everything below is a Serverless architecture to help users solve the problem.

Yunyuansheng product family

The last picture contains the family of aliyun cloud native products. Today, this field focuses on several of them. One is container service, which is the base of the whole cloud native products. There is also cloud native application architecture, covering microservice architecture, message-based heterogeneous invocation, event-driven architecture, Serverless architecture, etc. Finally, there is the operation and maintenance of cloud native applications. When your application architecture and system architecture become cloud native, its observability and high availability also need some corresponding capabilities, including our pressure measurement ability and high availability ability.

The most comprehensive cloud native open source contribution

Ali’s contribution to open source is also very rich. I believe most of you have used our open source products, such as Nacos, Spring Cloud Alibaba, Seata, Sentinel, etc. All of these open source products come from our cloud native department.

The above is all the content I introduced, I hope to help you, welcome to discuss in the comments section!

The original link

This article is the original content of Aliyun and shall not be reproduced without permission.