At the beginning of May 2018, I participated in the KubeCon2018 European Conference held in Copenhagen, Denmark. The number of participants exceeded 4000 for this year’s event – a three-fold increase over last year. Representatives from industry giants like Microsoft, Amazon AWS, and Huawei were present as were IT professionals from such industries as finance, retail, and manufacturing who have learned to apply Kubernetes to the key services at their organizations. The era of running containers at-scale on Kubernetes has come.
Kubernetes as an Infrastructure Standard
A multi-container orchestration platform came into view a few years ago when Kubernetes first came out. In just two years, Kubernetes had won the battle of container orchestration tools and has since become the de facto standard in container management. However, the impact of Kubernetes goes far beyond that of container management. All global public cloud suppliers provide Kubernetes services. This means that for the first time in history, R&D personnel can use the same computing platform, regardless of which cloud platform they are using for compute resource provisioning. R&D personnel can seamlessly migrate the working environment on their laptops to the public cloud for development, testing, and large-scale production. It is no wonder why everyone in R&D wants to use it. More and more application development teams are resolving to use Kubernetes all the way from the programming phase. All these trends indicate that Kubernetes will become the standard in cloud platform infrastructure.
Huawei adopted Kubernetes technology early on, becoming a leader and top innovator in the field. Many of Huawei's engineers have made great contributions to the core code of Kubernetes. Naturally, it also made sense that Huawei would become the first Chinese cloud computing provider to provide Kubernetes services. You can use HUAWEI CLOUD to obtain the Kubernetes-based experience.
Kubernetes as a Distributed Application Management Platform
R&D personnel were confronted with the challenges in using different API interfaces from various cloud computing service providers. Each seemed to speak its own language, making it difficult to deploy their applications on different IaaS clouds. Kubernetes runs the same APIs for different IaaS clouds, making it the compelling choice in distributed application platforms. As long as the Kubernetes API is used to develop applications, developers need not worry about how to migrate their applications from one cloud platform to another.
More and more distributed applications are supporting Kubernetes. Spark 2.3 released in March 2018 supports Kubernetes, which means that users do not need to create a YARN or Mesos cluster to run Spark. At the KubeCon European Conference this year, the KubeFlow project garnered lots of attention. The biggest highlight of KubeFlow was that it could run TensorFlow, PyTorch, and many other AI frameworks on Kubernetes.
In the past, users often felt that they would be locked in by suppliers of cloud computing services like AWS. Once you use a platform-specific system like RedShift on AWS, migrating your applications becomes extremely difficult. With Kubernetes, developers can use all platform services that are built on Kubernetes no matter which cloud platform they choose to develop on.
In 2017, Rancher Labs cooperated with the HUAWEI CLOUD container team to create a service catalog for Kubernetes so that many distributed applications could run more smoothly. Today, this project has been successfully launched. Users of HUAWEI CLOUD Service Engine (CSE) can deploy such services as MySQL HA databases and ELK clusters in a single click. And, this is only the beginning! The long-term goal of the cooperation is to implement a powerful Kubernetes-based distributed application platform powered on the HUAWEI CLOUD. The value of such a platform is going to be immense.
The Future of Cloud Computing
The development in cloud computing platforms has left many surprised. Before Kubernetes, no one had really even envisioned a distributed application platform able to run on all cloud vendor platforms. Kubernetes greatly improves application portability and system availability for this generation of cloud computing platforms.
What is the future of cloud computing? In the past decade, cloud computing platforms used various open-source standards, such as Xen, KVM, OpenStack, and Kubernetes. It is highly likely that next-gen cloud computing platforms will still be built on open source technology. The most successful cloud computing services, including database, big data analytics, and artificial intelligence, are built on vetted open source technologies. In building next-gen cloud platforms, cloud service providers must adopt open source technologies and have the capability to innovate.
The competition in cloud services is a marathon, not a sprint. With its robust technological innovation capabilities, large pool of developers, and long-standing commitment to investing in open source, HUAWEI CLOUD is positioned as a leader in next-gen cloud platforms. I am looking forward to seeing the next batch of Kubernetes-enabled HUAWEI CLOUD services come to market and the tangible benefits these services will bring to users.
Dr. Liang Sheng, Co-founder and CEO of Rancher Labs, graduated from Yale University with a doctoral degree in Computer Science. He is well respected in Java and JVM design and development. In 2008, he created Cloud.com, which became known as the father of CloudStack. In 2011, Cloud.com was acquired by Citrix, and Liang Sheng became the first Chinese CTO of the company. In 2014, Liang Sheng founded Rancher Labs, which has remained the world's leading container management company.