Technological changes in the field of cloud computing

The essence of technology lies in the accumulation of accumulation. The virtualization, RESTFull API, and distributed concepts of the past have undergone earth-shaking changes over the past decade. This time, the author tries to give you a look at the technological changes that have taken place in the field of cloud computing in recent years, to see the evolution and evolution of cloud computing related technologies.

Software defines everything

Software-defined networks, software-defined storage, and software-defined data centers sound like a bit of a speculation. In fact, the technology that underpins these concepts has grown by leaps and bounds.

First, VxLan, ARP flood suppression, DVR, VPN and other technologies have been implemented. The domestic cloud computing manufacturer Qingyun QingCloud has overcome the technical difficulties with the industry leading level, and developed software-defined network 2.0 to support business VPC, load balancing, and large Scale virtual machine migration, physical network efficient connection, etc. provide excellent solutions in the industry.

Second, storage, especially the scalability of block storage, and the balance of performance are the most troublesome issues in the industry. The emergence of products like NeonSAN completely broke this spell.

On the other hand, cloud computing has gradually entered the core application of the enterprise, and the business is in a convergence trend.

The concept of cloud computing is born with the application of technologies such as the Internet, mobile Internet, and big data processing. Those traditional business applications, such as ERP, CRM, office suites, etc., have been eager to run in the cloud platform to solve the capacity. Planning, performance, operation and maintenance issues. Qingyun QingCloud, a domestic cloud computing manufacturer, has developed a unified platform integrating virtual machines, containers and physical machines to meet the needs of various applications with a consistent experience, and to promote cloud computing into enterprise core applications.

Microservices into application mainstream architecture

The emergence of Docker completely overturned the past understanding of PaaS, and replaced many virtualization-centric IaaS cloud services, and even once the life of all PaaS platforms, such as CloudFoundry, OpenShift and so on.

Docker is an implementation of a container under Linux. The Linux container, the kernel, allows you to run multiple instances of user-space at the same time, not just the common one, which is the operating system-level virtualization we often refer to. Compared to virtualization that simulates the entire hardware architecture, such as KVM, containers have many advantages such as flexibility and lightweight. Especially in the form of version control, the warehouse can quickly deploy and roll back existing services. For a while, containerized applications have become everyone's daily routine.

Since the lightweight container has emerged, the subversion is the resource scheduling, the understanding of the service, and the further loosely coupled distribution. Therefore, the container orchestration framework represented by Kubernetes further splits the business and micro-services, each The service is divided into multiple Pods, and each Pod runs multiple containers. It has been turned out to be self-healing and automatic service discovery, and is favored by the industry.

Open source is becoming a public cloud battlefield

Open source is the cornerstone of the entire Internet. In fact, the entire IaaS platform is based on open source. It is no exaggeration to say that without open source, there is no such thing as cloud computing.

Linux, KVM, Xen, Hadoop, Docker, Kubernetes... Numerous excellent open source projects have laid the foundation for major cloud computing platforms, but public cloud platforms sell resources in business models: computing, storage, networking, bandwidth, etc. Rather than the framework of its own software, we will find that the giants are open source to attract and recruit developers and build an ecological environment.

DevOps gradient to AIOps

DevOps has been going from landing to landing for many years, but the foundation is too deep, the barriers are too heavy, and there are very few that can cross the divide. Dev or Ops? Who is leading? So, the debate on technology culture, the choice of technology for a while, in short, no progress. This is good, the emergence of artificial intelligence may completely ruin the lives of all people, the entire operation and maintenance are done by artificial intelligence.

Through the learning and training of historical data, artificial intelligence can achieve advance prediction, automatic fault transfer, and automatic repair. Well, don't mention DevOps anymore. The change of culture takes several generations, and the change brought by technology is hard to stop.

Artificial intelligence "democratization"

As AlphaGo defeated the top Go players, artificial intelligence seemed to succeed overnight, followed by face recognition, autopilot, voice assistant, image recognition and other applications. But be aware that artificial intelligence, like humans, requires resources to learn and calculate.

All artificial intelligence requires a lot of computing resources, and GPUs known as floating-point operations can't even meet the demand. Major companies are actively developing TPUs. Of course, science is wonderful, and all the machine learning frameworks in the world are open source for everyone to use. Qingyun QingCloud is also actively exerting its own advantages and providing sufficient resources to fuel the “democratization” of artificial intelligence.