One word is stable, the cloud native product family supports the nine business scenarios of the Winter Olympics and creates a new experience of the Olympics on the cloud

The Beijing Winter Olympics has become the most-watched Winter Olympics, setting new records in terms of broadcast duration, technology, and content production methods. The application of cloud technology is a major feature of the broadcast of the Beijing Winter Olympics.

As a new interface of cloud computing, how can cloud native support the multiple business systems of the Beijing Winter Olympics?

Nine business scenarios to create a "stable" experience of the Winter Olympics

Business Scenario 1: Olympic ticketing, practice ACK Anywhere, and integrate cloud and edge

The ticketing system is an Olympic replacement service provided by the Ali Youku Olympic team relying on Damai's original on-site service capabilities. It mainly carries the verification service of on-site ticket sales, improves the safety and stability of on-site service, and strives to provide silky smooth on-site service for the Winter Olympics audience.

In order to reduce the delay of ticket inspection and improve the admission rate of Olympic spectators, the ticketing system needs to be deployed in each Olympic venue respectively, to realize the closed loop of the ticket inspection process in the venue, and to give full play to the advantages of large connection and low delay of edge computing. Improve the ticketing experience of ticketing and reduce the waiting time of Olympic spectators. However, since the ticketing system is deployed in each Olympic venue, the distance between the Olympic venues is relatively long. For security reasons, the servers deployed in the Olympic venues are not allowed to have public IP addresses. Each Olympic venue has its own intranet environment. It brings great challenges to the development, testing, verification, upgrade, deployment, operation and maintenance of the ticketing system. If each venue arranges several application delivery personnel to upgrade and deploy applications, the cost is too high and the efficiency is too low.

Staff on-site ticket inspection

Based on the above considerations, the ticketing system team hopes that developers can develop, manage, operate and maintain these edge-side nodes and applications on the cloud, and realize the integrated operation and maintenance capabilities of the cloud and the edge. Therefore, a set of cloud-edge integrated architecture and platform to support.

Cloud Edge Container Service (ACK@Edge for short) is a cloud-native application platform that provides standard Kubernetes cluster cloud hosting, supports edge computing resources, fast business access, unified management, and unified operation and maintenance. It can help users easily realize cloud computing Side-by-side collaboration. users use ACK@Edge to extend cloud applications to the edge by hosting edge nodes, linking data between the edge and the cloud, so that the edge nodes have the same capabilities as the cloud. Provides unified Ops capabilities for edge devices and edge applications in the cloud to ensure less O&M and high availability for edge devices and edge intelligent applications.

Based on ACK@Edge cloud-edge integration, Kubernetes container orchestration and scheduling capabilities, and ACK@Edge superimposed on Kubernetes for edge scenarios, such as edge autonomy, edge unitization, unitized deployment, and tunnel channel capabilities, it effectively solves the ticketing system. The pain point of application operation and maintenance finally carried the unified management and operation and maintenance business of all venues for the Winter Olympics and Paralympics in Beijing, Yanqing, and Zhangjiakou, as well as on-site ticketing services for the opening and closing ceremonies of the Bird's Nest.

ACK@Edge provides unique enhancements for Kubernetes in edge computing scenarios:

Node Unitization-Edge Node Pool: creates a corresponding edge node pool for each Olympic venue, and the edge server of the Olympic venue can quickly access the corresponding edge node pool.

Application Unitization - Unitized Deployment: ticketing system deploys UnitedDeployment through the unitized deployment of ACK@Edge, deploying services to the edge servers of different Olympic venues, realizing the ability to uniformly distribute applications and uniform operation and maintenance on the cloud. Business personnel can quickly upgrade and deploy the ticketing system of all Olympic venues without on-site deployment. This function has played an important role in the online process of the ticketing system.

Cloud operation and maintenance, remote debugging: The Tunnel channel provided by ACK@Edge allows business personnel to quickly view container logs and enter container debugging.

Edge autonomy: The edge autonomy capability of ACK@Edge can ensure the normal operation of services on the local edge server in extreme cases such as cloud edge network disconnection and host restart. During this period, the ACK@Edge team and the ticketing system team did a lot of disconnection and restart operations, which finally proved that the ticketing system has always provided services normally.

Up to now, based on the cloud-native cloud-edge integrated architecture, ACK@Edge has been widely used in CDN, IoT, smart logistics, industrial brain, new retail and many other scenarios. The perfect guarantee of the ticketing system for the Winter Olympics further verifies the value of ACK@Edge in the field of cloud-edge integration and edge computing. At the same time, Alibaba Cloud contributed all the core codes of ACK@Edge and established the OpenYurt open source community. OpenYurt, as a CNCF sandbox project, adheres to an open, free and neutral attitude, attracting a large number of developers from VMware, China Telecom, Zhejiang University Lab, China Unicom, China Mobile and other companies and organizations to participate. In the "2021 China Open Source Annual Report", the OpenYurt project entered the CNCF China Top 10 project activity. In the future, Alibaba Cloud will expand more business scenarios in the field of edge computing, and better serve the society by continuously polishing product capabilities and expanding the open source community.

Business scenario 2: Olympic light application, fast delivery based on container service ASK

Due to the impact of the new crown epidemic on the scene, the requirements for on-site viewing are relatively strict. PinQuest has launched an Olympic-themed adventure mobile game, which allows users to start their own "Winter Olympic Adventure" journey on their mobile phones, allowing the public to enhance their interaction with the event through novel and interesting online methods.

Different events lead to real-time changes in the number of players online, so extreme flexibility is required. The game deploys key modules based on the extreme elasticity provided by ASK (Alibaba Cloud Container Service Serverless Edition). The background services can be expanded in real time, making the user experience silky smooth.

ASK cluster is a serverless Kubernetes container service launched by Alibaba Cloud. Users can directly deploy container applications without purchasing nodes, without node maintenance and capacity planning for the cluster, and pay on-demand based on the amount of CPU and memory resources configured by the application. ASK cluster provides complete Kubernetes compatibility while lowering the threshold for using Kubernetes, allowing users to focus more on applications rather than managing the underlying infrastructure.

Business Scenario 3: The core system of the Winter Olympics, calmly responding to large-scale sudden network traffic

The core system of the Winter Olympics is an important carrier of information services for the entire Winter Olympics and the Winter Paralympics. It integrates event-related information, provides service function modules, realizes the personalized service of "one app for the Winter Olympics", and also provides services for the organization and operation of the Games. Mobile communication collaboration and sharing service platform. The core system of the Winter Olympics is one of the most important core systems of the Olympic Games.

Among them, the Winter Olympics mobile service is provided by the Beijing Winter Olympics Organizing Committee, mainly providing event information, instant messaging, and mobile information service functions in the business areas of the Winter Olympics transportation, catering, accommodation, arrivals and departures, and venues. Urban mobile services mainly provide information on services such as eating, living, traveling, traveling, shopping, and entertainment outside the venues. For example, the health reporting function is closely related to athletes. Athletes arriving from overseas need to complete the health reporting on the Winter Olympics Pass before they can start the competition plan smoothly. This is just an example of the core system functions of the Winter Olympics. In general, stability is critical.

Winter Olympics core system implements front-end, proxy and back-end services based on the ACK cluster. In these various services, the unified page entry of the system is deployed on the ACK cluster, which means that the customer's traffic needs to go through the ACK to access the above-mentioned various background services, and the traffic link is very complicated.

During the stress test, in the scenario of high traffic and high concurrent access, the error rate of client requests continued to rise, which became an obstacle to going online. After receiving the problem request, the Alibaba Cloud Container Service ACK team and the customer immediately deployed performance observation points on all levels of the cluster. After investigation, the ACK team found that the service link is dominated by short connection requests, and the seven-layer SLB as the traffic entry will frequently reuse source ports in high concurrency scenarios, resulting in quintuple conflicts. After detailed verification, the pressure measurement traffic and reduced the error rate to 0% by optimizing the recovery time of TCP sessions in the Time-Wait state and other means, solving an important obstacle before the Winter Olympics core system was launched. , and finally successfully guaranteed the stability of the entire Winter Olympics.

The background application of the core system of the Winter Olympics is a Java-based microservice architecture, including nearly a thousand deployment application instances of Kubernetes. These application instances are co-located on a large number of nodes in the cluster. It is necessary to use resources efficiently and not allow these applications to interact with each other. They affect each other, which brings great challenges to resource sharing and isolation.

The container team optimizes the application configuration by means of node oversold evaluation, JVM memory parameter check, Pod resource tuning, Pod bottleneck and OOM real-time monitoring, etc., so as to balance the resource utilization of the cluster and business stability. Especially for a series of matching problems caused by the memory management of the JVM itself and the actual Pod memory limit, ACK introduces a means of automatically evaluating the rationality of the configuration to quickly find abnormal memory configurations and avoid OOM applied during the Winter Olympics. Generated to ensure the stability of the application during the event.

Business Scenario 4: Olympic Press Release Platform, One-Stop Container Intelligent Diagnosis Behind the Smooth Operation

InfoAV is the press conference platform of the Beijing Winter Olympics. It undertakes the core functions of live broadcast, directing, and on-demand of all the conferences of the Olympic Games, including pre-match interviews and post-match athlete group interviews. Initially, the customer deployed the InfoAV service based on self-built Kubernetes, but encountered problems such as inability to perform disaster recovery in the same city and difficulties in operation and maintenance.

adopted the container service ACK, it significantly accelerated the cloud migration and cloud migration of its business. The ACK cluster where InfoAV is located, whether it is a Kubernetes Master or a Kubernetes Worker, has achieved high availability and disaster tolerance at the level of multiple data centers in the same city; InfoAV uses the rich observability capabilities of ARMS to observe Kubernetes cluster resources and components with high precision. , traffic and other indicators, which has effectively improved the customer's protection ability.

In terms of architecture, the services deployed by InfoAV based on ACK include video and audio editing, media content management and other services. The business characteristics of the InfoAV system are computing-intensive and high-load. Therefore, the fast and comprehensive daily inspection capability provided by ACK significantly improves the reliability of assurance.

InfoAV interface, press conference scene

Business Scenario 5: Olympic Global Business, Best Practices for Safe Hosting and Efficient Distribution of Cloud-Native Application Products

The containerized business of the Beijing Winter Olympics, whether it is the APP and InfoAV business in the online scenario, the Olympic ticketing business in the edge scenario, and the PinQuest game business in the serverless scenario, are all critical to the security, stability, and scale of business application container image distribution. have different requirements. The Beijing Winter Olympics business has been expanded to 5 regions around the world, and 14 ACR instances have been activated to ensure the deployment, expansion, and fast recovery links of the Winter Olympics business containerized applications.

Alibaba Cloud Container Image Service (ACR for short) is one of the important infrastructures of cloud-native architecture, responsible for hosting and distributing cloud-native application products. The ACR Enterprise Edition provides enterprise-level cloud-native application products with secure hosting and efficient distribution capabilities for the Winter Olympics business.

The Winter Olympics business involves nearly 200+ application container images, which are deployed in edge, serverless, and online scenarios through ACK, ASK, ACK@Edge, EDAS, and ECS to ensure the security of container images and ensure complex scenarios. stable and efficient distribution.

addresses the security requirements of the Winter Olympics business, and ACR Enterprise Edition provides public network and VPC access control capabilities to control business application container image updates and distribution sources; at the same time, for the security requirements of the container image itself, ACR Enterprise Edition provides coverage. Container image vulnerability scanning and repair, container image signing and verification, distribution security blocking and other cloud-native applications deliver DevSecOps capabilities, further reducing potential risks after business applications are released and updated.

aimed at edge scenarios, and ACR Enterprise Edition supports the deployment of edge pre-instances in edge nodes, provides distribution acceleration capabilities such as distribution cache and edge P2P networking, and improves the distribution efficiency of container images in cloud-edge interconnection scenarios. In addition, for global business, ACR Enterprise Edition also provides global synchronization capabilities to support global multi-region deployment of container applications and disaster recovery services.

In addition to business link assurance, during the Winter Olympics assurance process, with the help of enterprise version instance panoramic monitoring and business exception analysis, ACR Enterprise version can quickly identify user business application containerization deployment exceptions and feedback them to customers, allowing customers to Solve the problem of application publishing failure caused by credentials, permissions, configuration errors, etc. as soon as possible. At the same time, relying on cluster events and monitoring to realize full-link container image distribution problem identification and diagnosis. For example, during the Winter Olympics, it was identified that the temporary DNS resolution abnormality after the node was restarted caused the node to fail to pull the image (that is, the application recovery failed). , which can allow customers to plan the application water level more reasonably in advance and improve the overall stability of the application.

Business Scenario 6: Data Exchange and Sharing Platform, "Data Middle Office" Based on Cloud Native Microservices

The 2022 Beijing Winter Olympics and Winter Paralympic Games data exchange and data sharing platform (hereinafter referred to as the exchange and sharing platform) is responsible for providing the organizing committee with the internal system, the organizing committee's external dedicated line data source and the data exchange and data sharing of the Internet data source. Service is a veritable "data center". This system is the data center and data brain of all core system groups on the cloud, and undertakes the functions of data aggregation, data filtering, data delivery, and data protection on the cloud.

The exchange and sharing platform is based on the Spring Cloud architecture design, and relies on the enterprise-level distributed application service EDAS to achieve application life cycle management, microservice governance capabilities and environment isolation management capabilities. EDAS enterprise service helps the exchange and sharing platform to complete the rapid cloud migration, iterative release and operation support.

Application life cycle management: EDAS manages the ECS cluster where the business is located, and provides rich release strategies such as batch release and grayscale release of applications, as well as functions such as change record and log query, which greatly improves the application release of the exchange and sharing platform and operational management capabilities.

governance: provides functions such as application monitoring, current limiting and downgrading, and full-link grayscale, ensuring the high availability of the exchange and sharing platform.

Environment isolation management: divides the production environment and the test environment through the microservice space to ensure data and traffic isolation between environments; and supports service east-west authentication to ensure the security of the exchange and sharing platform.

Adhering to the concept of green Olympics on the cloud, this Winter Olympics has built more than 20 events such as ADS personnel arrival and departure, MDV data screen, InfoAV data release, RHP car-hailing, etc. Digital management system. The business logic between these systems will generate a large amount of data exchange requirements.

Due to the large number of business subsystems and the complex business logic classification, traditional synchronous calls between systems will lead to a significant increase in the coupling of each subsystem and difficulties in system data link management. The technical architecture team of the Beijing Olympic Organizing Committee adopted a unified DES data exchange platform based on message queues to solve the problems of coupling and complexity of the system by standardizing the data format and communication method. Data sources and Internet data sources provide data exchange and data sharing services.

In the DES data exchange platform of the Beijing Winter Olympics, Alibaba Cloud RocketMQ is used to build the underlying asynchronous reliable communication channel:

1. The data source subsystems of all parties report data to the DES gateway through a unified API specification.
2. After the DES gateway performs data format and permission verification, it performs simple data encapsulation and format conversion, and writes it to the RocketMQ topic on the backend.
3. The downstream data consumer subsystem subscribes to the RocketMQ topic to obtain the data it needs to achieve complete decoupling from the upstream system.

During the Winter Olympics, there are two types of data transmitted based on the DES data distribution platform. One is task-synchronized data, such as the required personnel registration information. The core OMS system generates the registration information and pushes it to DES, which is then sent to the backend of the Winter Olympics core system. Subscribe to pull personnel information to achieve login authentication; another type of real-time notification data, such as event schedules, medal standings and weather data, generates real-time score data through the internal ODF system and publishes it to DES, and then pushes it to the downstream system through the message queue to expand screen display.

Business Scenario 7: RocketMQ carries the sudden traffic of ADS arrivals and departures of the Winter Olympics and RHP car-hailing trips

During the Winter Olympics, with the entry, exit and daily travel needs of many domestic and foreign personnel, the technical architecture team of the Beijing Olympic Organizing Committee designed and implemented the ADS personnel arrival and departure system and the RHP car-hailing travel system for personnel travel management. The ADS personnel arrival and departure system is responsible for receiving the arrival and departure itinerary information and entry and exit information of personnel from the Capital Airport and the Immigration Bureau, and writing it into the background management system to facilitate subsequent push processing. The RHP car-hailing travel system needs to manage a large number of online car-hailing order management.

Due to the great uncertainty in the flow of people during the Winter Olympics, there is an uncertain flow pressure in the entry and exit of travelers and in the processing of travel data. In the process of data processing, the technical architecture team of BOCOG used the Alibaba Cloud message queue RocketMQ for traffic buffering and reliable data transmission. The system links are as follows:

The gateway systems of ADS and RHP receive data calls from outside and verify the validity.

The gateway system has a built-in RocketMQ SDK, which writes traffic calls into RocketMQ messages for asynchronous buffering to avoid burst traffic from overwhelming systems such as background databases.

The background processing system consumes messages asynchronously to achieve smooth traffic writing to the database and other systems. If there is a writing problem, the RocketMQ reliable retry mechanism is used for retry processing.

With the help of RocketMQ's billion-level stacking capacity and massive concurrent writing capacity, the RHP travel system smoothly carries the travel management needs of 6000+ drivers and passengers during the Winter Olympics, as well as the entry and exit management and push needs of all personnel.

Business Scenario 8: Cloud integration, MQTT supports CCTV News App massive user comment forwarding

In the CCTV News APP, the MQTT version of the micro message queue is used in the live broadcast room to send users' comments and receive comments from other users. On February 4, the CCTV news app broadcasted the opening ceremony of the Winter Olympics live. After the opening ceremony, the average number of messages in the MQTT version of the micro-message queue was hundreds of thousands. The total amount of upstream and downstream messages in the whole process is billions.

Alibaba Cloud Micro Message Queue MQTT version is a message middleware built based on the open source MQTT 3.1.1 protocol for end-to-cloud communication. It is suitable for Internet of Things and mobile Internet business scenarios. Through the MQTT client-server message publish/subscribe transmission mode, it provides reliable message transmission services for devices in low bandwidth and unstable network environments. Moreover, if there is a performance bottleneck during the peak business period, through the load balancing message push capability of the MQTT version of the micro message queue, it is only necessary to quickly upgrade the specification to increase the number of application instances, and the user can achieve a horizontal and linear improvement in application consumption capacity without feeling. .

By using the MQTT version of the micro message queue, developers do not need to care about complex technical issues such as weak network environment, disconnection and reconnection, exception handling, massive concurrency, and high system availability at the application layer, which greatly reduces development costs and improves It improves the user experience and the rapid launch of the business.

Business Scenario 9: Generate Thumbnails Quickly, Serverless Supports Event Broadcasting

The video thumbnail is the thumbnail of a certain frame of animation in the video, which is equivalent to converting the cover of the video or the content of one of the frames into a picture file. The selected video thumbnail can highlight the bright spots in the video and quickly capture it. Keeping users’ attention and improving the click-to-play rate, the thumbnails of competition videos are somewhat different from those of ordinary videos, and it is necessary to capture a large number of videos more quickly and in real time.

Customers supported by Function Compute need to process the massive live video of the Beijing Winter Olympics in real time, and take screenshots at different time points in the live video (the technical term "video image capture"). The customer's requirements are as follows:

Need timed events to trigger screen capture operations

The customer only wants to concentrate on the specific logic development, the core video screenshot logic code is not large, and the development needs to be completed quickly

Customers don't want to manage and deploy virtual/physical machines, they want the simpler the better

Customers need this core business logic to be flexible and highly available, free of operation and maintenance

By using Alibaba Cloud Function Compute (FC), users can develop the core business code of video screenshots. With low development cost and one-click deployment, they can obtain a live video screenshot service that is flexible, highly available and free of operation and maintenance, helping to realize the highlights of the Winter Olympics video. Crawl in real time.

Comprehensive security system to escort the Winter Olympics

In the previous Tokyo Olympic Games, Alibaba Cloud Container Service ACK supported the smooth operation of the Olympic Games through integrated operation and maintenance, full-scenario diagnosis, and a complete guarantee system. For this 2022 Beijing Winter Olympics, the container service ACK has strengthened its guarantee capabilities, provided new data-based and refined operation and maintenance, and created a more three-dimensional escort guarantee system.

Data operation and maintenance

The unified operation and maintenance platform of Container Service ACK integrates the monitoring, alarm, event, log, inspection, metadata management and other functions of the entire network cluster, and can observe and manage tens of thousands of clusters in 24 regions of the entire network in real time. However, the data sources are scattered, heterogeneous and fragmented, which makes operation and maintenance under great pressure and cannot do correlation analysis. Therefore, the container service ACK builds a data warehouse platform according to the standard big data system, uniformly cleans and stores the data in a structured manner, and conducts analysis according to statistical requirements. Hierarchical modeling provides a real-time + offline data model rich in operational semantics.

The modeled data can be directly drawn and analyzed or used for alarming, such as quantifying the health of the cluster: calculate the component health score through component configuration, monitoring values, event/log anomalies, historical water levels, etc., and then combine nodes, work orders, quota specifications, etc. The weighted calculation of the overall health score of the cluster. Based on this model, a wide table is drawn and aggregated by multiple dimensions to display high-risk items and uniformly manage clusters with low health scores.

Anomaly Analysis

The Container Service ACK operation and maintenance platform cleans and stores cluster events and component logs in a unified format, and conducts alarms and trend analysis for abnormal content. Exceptions are classified for all types of clusters on the entire network, detected in advance, and resolved by special classification to ensure that important periods of core business are not affected.

At the same time, the observability system is driven by the SLO definition, and the normalization capability of performance stress measurement is built according to the capacity scale of Kubernetes, and the QPS number of the core Verb of the Kubernetes cluster has a clear understanding. At the same time, the SLO is sorted out in combination with different workload scenarios, and the request volume is continuously paid attention to. , latency, number of errors, saturation and other golden indicators.

Cluster-level grayscale and network blocking

For re-insurance activities, Container Service ACK has a complete and professional plan management system. During the guarantee period of large-scale activities, Alibaba Cloud's overall management and control plan is followed, and the network is blocked in a specific range and period. In this Beijing Winter Olympics, ACK will refine the network closure management and control to the cluster level. For the operation and maintenance of the re-protected cluster, you need to apply for specific permissions separately. Read and write operations are applied in two levels. The release of changes requires a special approval process. Revoke existing permissions and re-issue specific personnel.

By refining the grayscale management and control to the cluster dimension, the scope of risks can be more strictly controlled, while avoiding large-scale interception and affecting routine operation and maintenance.

Perfect security plan system

For the Olympic Games, based on the existing guarantee process, Container Service has formulated a full-course guarantee plan, including advance plans, emergency plans, fault drills, and on-duty schedules. Container Service has rich experience in guaranteeing, and the annual routine guarantee activities include the annual Double Eleven, 618, Spring Festival, etc. These large-scale guarantee activities are complex and comprehensive, and the container service has achieved nearly zero failures during these activities every year.

In addition to the above major support activities, there are normalized chaos-based fault drills and raids inside the container service. The chaos system randomly injects faults, and the container team on-duty will receive an alarm and deal with it immediately according to the plan in the plan system; after normalization The team's emergency handling ability has been tempered with mature and tacit understanding, and can well achieve the goal and methodology of 1-5-10 (send an alarm within 1 minute, locate faults within 5 minutes, and fix faults within 10 minutes). These guarantee systems, which have been repeatedly polished in actual combat, have been applied to the special Olympic Games guarantee, effectively guaranteeing and supporting the stable operation of the Olympic Games.

The advance plan for the Olympic cluster will be implemented for each Olympic re-insurance cluster before the start of the event, which can identify and eliminate risk points to the greatest extent. Inspection methods include Qianzhou platform and automated inspection scripts to check the water level and status of cluster nodes and components covered by the project, network traffic, whether Pods are broken up by availability zone, configuration and technical indicators of key components (CoreDNS, Ingress Controller, etc.), Kubernetes Multi-dimensional events such as events; the mechanism of performing daily inspections can continuously discover new problems and quickly feedback customer corrections.

Take the Apiserver-related plan of the ACK cluster as an example to explain. ACK has a refined flow control strategy, and can build protection capabilities against abnormalities found during stress testing and fault drills. In terms of specific implementation, Alibaba Cloud Container Service has built-in self-developed UserAgentLimiter to further refine current limiting and more comprehensively protect Kubernetes. The Apiserver emergency plan can be based on the above strategy to quickly issue the current limiting configuration, stabilize the Apiserver in seconds, and quickly restore the cluster.

A comprehensive and continuously iterative emergency plan is an essential weapon for quickly stopping bleeding and ensuring business after a fault occurs. For the Beijing Winter Olympics, the container service ACK and the container image service ACR have prepared dozens of emergency plans for the four categories of central management and control and operation and maintenance, management and control links, user plane, and data plane, and have practiced many times, covering high Risky failure scenarios. Ensure that if a dangerous situation occurs, the security personnel have rigorous steps and procedures to operate step by step, which significantly improves the speed and effect of emergency response.

ultimate flexibility

Business scenarios such as large-scale events and live broadcasts are often accompanied by sudden business traffic at a specific point in time, and the pressure on the system also changes in real time. Alibaba Cloud Message Queuing provides serverless on-demand service capabilities. The business scale can be flexible in real time within 5,000 TPS concurrently, and the business is completely insensitive. Ultra-large-scale concurrency can also be upgraded to the platinum version at any time, with smooth expansion in minutes, and a single instance supports up to 1 million TPS.

During the opening mode of the Winter Olympics, a large number of user live comments on the CCTV live broadcast client brought increased system pressure. The MQTT version of the micro-message queue can be expanded to a single instance of 1,000,000 TPS in time, and its strong elasticity ensures the stable operation of the system.

do not say goodbye

The Beijing Winter Olympics have come to an end

But cloud-native escort continues

at the upcoming Winter Paralympics

We will also do our best

Steady!