What do you think of the world's first container computing service ACS released by Alibaba Cloud?
Contents of this article:
4. Topic discussion in this issue
4.1. What do you think of the release of container computing service ACS?
4.2. Do you think ACS’s product design can reduce the cost of using K8s for enterprises?
4.3. What questions do you have about this new product?
Preface
At the main forum of the 2023 Yunqi Conference, Alibaba Cloud CTO Zhou Jingren announced the official release of the container computing product ACS at the 2023 Yunqi Conference in Hangzhou to help customers lower the threshold for using the cloud.
Container computing service ACS takes the lead in integrating containers and resources, redefining container computing power, and upgrading the computing power delivery model to a serverless form. The load types supported by ACS are also more abundant, which greatly reduces the cost and cost of using containers and K8s for enterprises.
So, can the complexity issues of Kubernetes, microservices and other architectures that the industry has been shouting about for so many years really be resolved? What is the background behind the emergence of such a “new form, breaking the rules” product? Can computing power control be more flexible while ensuring stability?
The ACS technology introduction part of this article is reproduced on the ACS official website, and the address is in the appendix.
1. What is ACS
Containerization on the cloud has become the new normal, and Kubernetes has become the new interface for using the cloud.
According to a research report released by Dynatrace this year, the number of Kubernetes clusters hosted in the cloud is growing at an annual growth rate of 127%, which is approximately 5 times that of locally hosted clusters. In the near future, the number of K8s clusters deployed on the cloud will exceed local deployments. At the same time, 73% of K8s clusters on the cloud use managed distribution versions provided by cloud vendors.
In addition, as more and more application architectures are microservice-oriented, the complexity of service governance is constantly rising. When microservices encounter K8s, it brings more complexity. For example, K8s configuration is complex, there are many computing resource models and generations, cluster and node component management is complex, and it is difficult to achieve on-demand elasticity during business peaks.
Container Computing Service ACS
A cloud computing service that uses K8s as the user interface to provide container computing resources, and provides computing resources that comply with container standards.
Container Computing Service (ACS for short) is a cloud computing service that uses K8s as the user interface to provide computing power resources for containers. It provides computing power resources that comply with container specifications, supports flexible payment models for resource reservation and pay-as-you-go, and serverless computing power. In the delivery mode, users do not need to pay attention to the operation and maintenance management of underlying nodes and clusters. ACS computing resources can support users' container application loads and also support the loads of Alibaba Cloud products.
2. Core features of ACS
Container computing service ACS has four core features:
▶Extremely easy to use and get started quickly
ACS realizes the integration of containers and resources, and customers do not need to pay attention to the underlying nodes and clusters. The product design is minimalist, application configuration is simpler, and ease of use is improved by about 50%, lowering the threshold for customers to use containers and K8s.
ACS simplifies the two forms of ClusterLess and NodeLess. Customers do not need to pay attention to the operation, maintenance and management of the K8s cluster, nor do they need to care about the various information and operation and maintenance of the nodes. Different types of users can choose to use ACS in a way they are more familiar with.
For novice users: Application deployment can be achieved through simple white screen operations through the ACS console.
For developers: ACS supports Yaml mode for instance configuration, and then uses K8s API to quickly deploy applications.
For enterprise customers: ACS supports K8s Yaml and OpenAPI to adapt to its own CI/CD system.
▶On-demand flexibility, pay-as-you-go
ACS is the industry's first ultra-flexible Pod configuration mode, providing resource allocation ratios of 1:1~1:8, supporting second-level elasticity to calmly cope with traffic fluctuations, supporting second-level billing, and providing pay-as-you-go, resource reservation, and savings plans. 3 billing modes to achieve on-demand flexibility.
▶On-demand flexibility, pay-as-you-go
ACS supports ultra-flexible configuration of Pod resource specifications, starting from 0.25c/0.25GB, and provides CPU and memory ratio selections of 1:1~1:8, helping customers more flexibly match business needs, and supports second-level elasticity and second-level elasticity. Level billing to calmly cope with peak traffic fluctuations.
Enterprises usually ensure the certainty of computing resources through resource reservation. Appropriate resource reservation can not only ensure the long-term stable operation of the business, but also avoid the waste caused by excessive resource reservation. Therefore, based on ACS, customers set appropriate resource reservation water levels. Beyond the resource reservation water levels, when encountering business traffic peaks, customers can use second-level elasticity capabilities to cope with rapid business expansion. In this way of use, business resource occupation can be as close as possible to the traffic curve, reducing resource waste.
In addition, ACS supports savings plans to provide better product discounts.
▶Rich scenes and flexible deployment
ACS has been deeply integrated with other Alibaba Cloud products to support the seamless migration of open source ecosystem and self-developed products to the cloud.
ACS supports user self-development and open source applications, such as microservices, web applications, Spark applications, etc., and is deeply integrated with Alibaba Cloud's cloud products, including artificial intelligence platform PAI, etc.
ACS provides customers with a computing power cluster. In this computing power cluster, it supports the above-mentioned several workloads to run in a large cluster. It is the first computing power allocation function in the industry to support customers to use computing power resources among self-built loads. Flexibly allocate resources between spaces, between cloud product loads, and between self-built loads and cloud product loads to achieve resource sharing and improve idle resource utilization.
ACS supports the seamless migration of open source ecosystem and self-built products to the cloud. The cloud-native scheduling system Koordinator is compatible with the K8s ecosystem. Open source projects can be seamlessly migrated to ACS. User-developed products can be quickly migrated to ACS through K8s Yaml. It also supports self-built products. K8s clusters can be migrated to the cloud at low cost.
▶Free combination, high cost performance
ACS focuses on Pod definition, shielding model generations and differences. It also supports three instance types: general-purpose, task-based and standard, making computing resources more universal.
ACS provides the following 3 instance types:
▶Universal examples:
Suitable for most containerized workloads, such as microservices, web applications, network and video applications, etc.
▶Task type example:
Suitable for cost-sensitive and fault-tolerant workloads, such as big data computing, audio and video transcoding, etc.
▶Exclusive instance:
Suitable for extremely delay-sensitive workloads, such as high-performance network servers, gateway systems, middleware systems, etc.
Among these three types, ACS implements the function of shielding models and generations. This product capability can help customers automatically achieve optimal cost performance when faced with a variety of different scenarios and loads, and when multiple types are required to be combined and applied, and can save up to 20% of resource costs.
3. Key technologies of ACS
ACS seamlessly integrates Alibaba Cloud's product system, including application real-time monitoring service ARMS, log service SLS, load balancing SLB, elastic public IP, and supports storage services such as EBS and NAS, as well as access control services such as RAM and KMS, as well as support K8s ecosystem and extensions support Extension Webhook, Controller and Operator.
ACS adopts an ultra-large-scale multi-load differentiated SLO-enhanced scheduling system. This scheduler is derived from Alibaba’s large-scale practical experience during Double Eleven. In addition, it is deeply optimized for microservices, Java applications, AI big data, etc. . This scheduling system provides an open source version, namely Koordinator, so it supports the community ecosystem to be seamlessly connected to ACS products, achieving compatibility with various differentiated hardware, including GPU, NPU, etc.
This scheduling system has the following three major advantages:
Standardization: Built based on the community standard scheduleframework, fully compatible with the kubernetes scheduling ecosystem.
High efficiency: It has rich resource tuning capabilities and is deeply optimized in terms of topology awareness, load awareness, QoS awareness, etc., achieving 30,000 times/second scheduling.
Large-scale: It is used on a large scale in external customers and internal business scenarios, and supports ultra-large-scale clusters (a single cluster exceeds one million cores).
K8s is a single-tenant system. Self-built K8s faces design requirements for multi-tenancy isolation and security, including K8s Master on the control plane, multi-tenancy transformation of a full set of management and control links, and security isolation requirements on the data plane.
Relying on Alibaba Cloud's secure container technology, ACS can effectively avoid interference caused by multiple different tenants' workloads running on the same physical node, as well as mutual security visibility issues. For storage, ACS uses Alibaba Cloud's high-performance block storage EBS to meet the storage isolation and performance multi-tenancy requirements of containers. On the network, ACS connects to Alibaba Cloud Network Service VPC to provide a high-performance container network.
4. Topic discussion in this issue
4.1. What do you think of the release of container computing service ACS?
Container technology is one of the hottest technologies in the IT industry today. Most end users will use K8s and related container technologies in production. As containerization and cloud migration become the new normal, how to keep up with business needs and avoid resource waste has become a new challenge.
At the 2023 Yunqi Conference, Alibaba Cloud released the world's first container computing service ACS. ACS is the world's first container computing service (Alibaba Cloud Container Compute Service, referred to as ACS).
ACS is a container service based on Kubernetes, which provides users with a platform for rapid creation, easy operation and maintenance, and high availability of container clusters, supporting a variety of application scenarios. ACS uses scheduling technology to automate computing resources into containers, improving resource utilization and application scalability.
ACS uses K8s API as the computing power usage interface and adopts a serverless computing power delivery model. Users do not need to pay attention to the operation and maintenance management of the underlying nodes and clusters, and it also supports resource reservation and on-demand elasticity models. In addition to supporting users' application loads, computing resources also support users' flexible allocation of loads to Alibaba Cloud products.
In addition, ACS also has the advantages of automated scaling, comprehensive monitoring, and rapid deployment, which can significantly reduce the operation and maintenance burden of enterprise IT personnel and improve their efficiency and competitiveness.
Developers can more flexibly control computing power through the ACS service. This integration of containers and resources can be said to have greatly lowered the threshold for Kubernetes and cloud use.
The release of container computing service ACS is exciting news and a major milestone. It integrates containers and resources, redefines container computing power, and upgrades the computing power delivery model to a serverless form. This means that customers have more flexibility in their computing power and can adjust it at any time according to their needs. This new form factor and out-of-the-box product has the potential to reduce the cost of using containers and K8s for enterprises. It makes cloud computing and containerization technology more popular and convenient. Containers can deliver computing power in a serverless form and can be adjusted as needed.
Personally, I feel that if operations and post-technology can keep up, ACS will further promote the development and popularization of container technology, provide users with more efficient and reliable cloud computing services, and help promote the digital transformation of enterprises.
4.2. Do you think ACS’s product design can reduce the cost of using K8s for enterprises?
From the perspective of ACS product design, the release of ACS is expected to reduce the cost of using K8s for enterprises. It integrates containers and resources, redefines container computing power, and upgrades the computing power delivery model to a serverless form. This means that customers have more flexibility in their computing power and can adjust it at any time according to their needs. This new form factor and out-of-the-box product has the potential to reduce the cost of using containers and K8s for enterprises.
① ACS supports more load types, which means that enterprises can use ACS to run different types of applications, not just traditional stateless applications.
② ACS integrates containers and resources to make resource management and scheduling more efficient, thus reducing costs.
③. The Serverless form of ACS means that users only need to pay for the actual computing power used without having to purchase or configure resources in advance, further reducing costs.
④. ACS supports ultra-flexible configuration of Pod resource specifications, starting from 0.25c/0.25GB, and provides a CPU and memory ratio selection of 1:1~1:8, which helps to flexibly match business needs.
⑤. ACS supports second-level elasticity and second-level billing, and can realize real-time calculation of fees as the traffic peak fluctuates.
⑥. ACS supports resource reservation, which can ensure the certainty of computing resources and stable business operation, while avoiding resource waste.
⑦. ACS supports second-level elasticity, which can cope with rapid business expansion. Business resource occupation is close to the traffic curve, reducing resource waste.
⑧. ACS supports savings plans and can enjoy product discounts.
⑨. ACS provides 3 instance types (general instances, task instances, and exclusive instances). Among these three types, ACS implements the function of shielding models and generations, which can help customers face multiple instances. When multiple types are combined and applied in different scenarios and loads, the optimal cost performance can be automatically achieved, saving up to 20% of resource costs.
4.3. What questions do you have about this new product?
①. How does ACS ensure stability? Will new stability issues be introduced in the process of integrating containers and resources?
②. Does ACS support cross-regional deployment and expansion? This is important for businesses that need to run applications in different regions.
③. ACS provides monitoring and logging functions, which are crucial for troubleshooting and performance optimization. Are these for single tenants or multi-tenants? If a user tenant rents several container resources, is there a system-wide monitoring portal (you can see the running status of all rented resources below, instead of viewing them one by one)
④. Is ACS integrated with other Alibaba Cloud products? Can it be integrated seamlessly with other services?
⑤. Because it is the world's first container computing service ACS, the ecosystem of ACS should still be relatively weak. How can community support and developer resources be improved as soon as possible?
⑥. ACS essentially transforms Kubernetes from a modular service into a product with integrated container computing power. It transforms a single-tenant system like Kubernetes into a multi-tenant system. If in certain scenarios, developers It may be necessary to directly call the Kubernetes API. Can ACS meet this demand?
⑦. Can ACS customize the host specifications according to user needs?
⑧. For new products, we generally adopt a wait-and-see attitude. Unless there are convincing successful cases, we will decide to adopt them. How does ACS open up the situation?
⑨ How is the safety, isolation and stability of ACS? Going to the cloud brings convenient operation and maintenance methods and lower costs, but it also introduces some risks. How to ensure security and risk avoidance?
5. Appendix
https://www.aliyun.com/product/acs?spm=a2c6h.13066369.question.4.42ab6ec8dxjVsU
Recommended reading:
|
|
|
|
| tomcat11, tomcat10 installation configuration (Windows environment) (detailed pictures and texts) |
Tomcat startup crash problem solution set (eight major categories in detail) |
|
