In recent years, as more and more enterprises build their own core business platforms based on micro-service architecture, micro-service has been recognized by more and more technical personnel. At the same time, micro-service also carries the important task of enterprise digital transformation. However, the implementation of microservice architecture brings many challenges to enterprise o&M teams. The original o&M methods and tools can no longer meet the requirements of microservice architecture. Such as:

  • Application management difficulty in microservice architecture Compared with the previous smokestack system construction mode, microservice architecture has reached a new level in terms of the number of servers and the complexity of service interaction mode. The original full-process management mode of application creation, deployment, startup, stop and offline is no longer suitable for microservice architecture.
  • Different from the traditional physical machine and VIRTUAL machine cluster mode, o&M personnel pay attention to the wiring, network communication, hardware deployment, application architecture, and data disaster recovery of the equipment room. In the microservice architecture, to ensure the continuous stability of the platform, Attention also needs to be paid to key technologies in distributed domains such as limiting and degrading traffic, traffic scheduling, capacity compression and evaluation, and full-link compression.
  • After the construction of the digital operations difficult micro service architecture, the technical platform evolved into a complicated network of service interaction link, how to service invocation appear error occurs every day fast positioning, how to carry out the operation of the monitoring service, pay attention to business to business team standards can be used to provide real-time data, these are the micro service system after the landing of the new demands.

Ali Cloud recently released a new version of enterprise-level distributed application service EDAS, which is born to solve the above problems, one-stop solution to a series of operation and maintenance challenges and problems encountered in the process of micro-service construction. On this year’s Tmall Double 11, the number of orders created, the number of payments and the total amount of transactions all hit new highs. During this period, the smooth shopping experience enjoyed by users cannot be separated from the service capabilities provided by Alibaba Cloud EDAS.



Live to review

According to Dai Xin, middleware product expert of Ali Cloud, EDAS is alibaba’s self-developed PaaS platform centering on applications and micro-services. After 10 years of technological precipitation of Ali, EDAS has successfully passed the severe test of double 11 flood peak flow every year, and has been called more than 200 billion times per day at present. Users can deploy applications in various languages through Docker images, wars or JARS, and manage the life cycle of these applications. Compared to other microservice PaaS platforms on the market, EDAS offers a more complete microservice solution with extensive production experience.



1. Fully enhance the native support of mainstream micro-service frameworks, and realize the migration of Spring Cloud & Dubbo user code to EDAS with zero intrusion. You do not need to independently build and maintain components such as ZooKeeper, Eureka, and Consul, saving hardware and maintenance costs and lowering the access threshold for developers.

2. EDAS has been greatly optimized in terms of throughput and response time. Compared with the open source implementation of Spring Cloud under the same testing conditions, the throughput of EDAS is more than 12,000 times per second, more than one times higher than the performance of Spring Cloud. The average processing response time is only 50% of that of the Spring Cloud framework. Not only is EDAS powerful, but the framework uses fewer hardware resources, about 60% less CPU and Load than Spring Cloud.

3. Integrated with high availability management and control components incubated from Alibaba Double 11 scenario, such as traffic limit degradation, traffic monitoring, application physical examination and fault drill, it can effectively deal with sudden traffic peak and avalanche caused by service dependence, greatly improving the stability of the platform.

4, the first large-scale distributed call tracking system integration industry, summary, this article analyzes the distributed calls every request information, thus accurate discovery system performance bottlenecks and exception, at the same time provides three-dimensional monitoring and digital operation ability, thread diagnosis, log diagnostic, database, and a series of diagnostic ability of application of intelligence operations.



It is worth noting that EDAS launched Serverless public beta version at this press conference, becoming the first Serverless container based on Kubernetes and supporting microservices framework in China. With this service, users do not need to manage and maintain servers and clusters, saving 5-7 minutes for managing iaas&cluster initialization, and creating and expanding applications can be completed in about 20 minutes. At the same time, users can quickly upgrade from the original ECS resource utilization rate of 10% to on-demand application, charging by volume, no longer need to pay for idle resources, greatly improving o&M efficiency and resource utilization.



In addition, the new version of EDAS launched a special project to optimize the developer experience, enhance the entry ability, lower the access threshold, open up the cloud products and tool chain, and continue to solve the pain points of the whole development cycle of users using the micro-service architecture. The new version introduces Alibaba Cloud Toolkit developer tools, such as edas-Maven-Plugin, which can be deployed to EDAS with a single command after building with the help of Maven plug-in, reducing application deployment from 7 steps previously to just 1 step.

It is reported that EDAS has been serving well-known enterprises such as Sinopec, China Post and Bosideng, helping users to obtain one-stop micro-service construction capabilities. Zhan Yan, senior product expert of Ali Middleware, said that it is our goal to provide one-stop micro-service solution with lower cost, higher efficiency, higher stability and universal on public cloud. We will continue to evolve around intelligent scheduling, operation and maintenance automation and interface standardization. It is believed that in the future, more and more developers and enterprises will use EDAS to obtain powerful micro-service architecture and governance capabilities, as well as efficient business delivery capabilities, and accelerate the digital transformation of enterprises.



The original link

This article is the original content of the cloud habitat community, shall not be reproduced without permission.