Hi everyone, welcome to the second chapter in our Tencent Cloud Developer Associate course. Getting started with Tencent Cloud development. At the end of this chapter, you'll be able to understand Tencent Cloud and its global infrastructure. Prepare to develop applications and Tencent Cloud, and use Tencent Cloud API. In this chapter, we'll cover three sections; overview of Tencent Cloud, preparing to develop in Tencent Cloud, and overview of Tencent Cloud API. Let's get started with section 1; overview of Tencent Cloud. In this video, we'll cover an overview of Tencent Cloud, an overview of Tencent Cloud's management platforms and services, and an overview of Tencent Cloud's global infrastructure. First, let's take a look at an overview of the Tencent Cloud platform. From its inception in 1999-2010, Tencent accumulated a wealth of experience in Cloud service operations through the rapid growth of its leading products, QQ and Qzone. Then Tencent went on to launch WeChat and Tencent games, and gained technical experience from its social media and gaming services. Leveraging these experiences, Tencent Cloud has integrated and opened up various Cloud services, ranging from infrastructure as a service to software as a service. In a nutshell, Tencent Cloud provides basic Cloud infrastructure services such as Cloud servers, Cloud databases, Cloud storage, content delivery network, CDN, as well as industry solutions such as games, videos, finance, and big data. Besides its basic infrastructure, Tencent Cloud has established several research labs focusing on cutting-edge technologies, such as facial recognition, information security, Cloud native technology, 5G, and edge computing. Speaking of Tencent Cloud's strengths in infrastructure, Tencent Cloud offers 26 regions and 70 availability zones in all major continents, as well as more than 2,000 high bandwidth CDN cache nodes. Tencent Cloud's second strength is its management experience on a global scale. Currently Tencent operates over a million servers and has over two decades of operations and maintenance experience. Third, Tencent purchases a large number of servers, networks, and storage equipment every year, which helps it to significantly reduce the cost of its Cloud products. Fourth, Tencent Cloud has strong R&D capabilities in terms of industry-leading CVM performance and proprietary database performance. Fifth, it also provides business capabilities in a variety of fields, such as optical character recognition, OCR, image recognition, speech recognition, and facial recognition. Here are some examples of Tencent Cloud R&D capabilities. For processing massive amounts of data, Tencent has developed and customized the MySQL database at the kernel level, which greatly improves the replication performance of the MySQL database. This technology is also applied to Tencent Cloud database, CDB, making the multi-copy replication function of the Tencent Cloud database more subtle and efficient. For general computing use cases, Tencent Cloud has launched three new instances, standard, computing, and memory, which are based on the latest CPU generations. The overall computing performance is increased by 60 percent. Tencent Cloud provides up to 25 gigabytes of private network bandwidth. This is an ongoing development effort and we are working closely with AMD to launch the second generation of CPUs. For storage intensive scenarios, Tencent Cloud launched two new storage optimized instances, a big data model and an IO intensive model. In particular, the IO intensive instances use local NVMe SSD hard disks with random input and output operations up to one million per second. Tencent also operates three products with a large number of users, WeChat, QQ, and WeChat Pay. These three products generate huge amounts of data every day, which requires us to build mature processes to support the development of these products, along with the necessary computing architecture. Tencent has applied these experiences when it developed its Cloud platform. Compared with building local private data-centers, the construction cost is eliminated. The system and software procurement cost is reduced. The deployment cycle is greatly shortened and the time and labor costs required for operations are also reduced to a large extent, allowing customers to focus on adding more value to their business. Tencent Cloud's service provision capabilities include infrastructure as a service, IaaS, platform as a service, PaaS, and software as a service, SaaS. The IaaS layer provides products such as Cloud servers, load balancing, network infrastructure, and Cloud storage. The PaaS layer provides products such as Cloud database, Cloud cash, video services, and container management. Lastly, the SaaS layer provides products such as face recognition, intelligent voice, and communication services. Next, let's go over an overview of Tencent Cloud's underlying management stack. We use a Cloud management platform called V-Station, which is independently developed by Tencent. V-Station was created in December 2013 and went online quickly a month later. Previously, Tencent used OpenStack for internal operations, but found that OpenStack was not suitable as the company grew to a global scale. Therefore, the principals of design for V-Station, aim to make the system more scalable, more efficient, and easier to manage. Let's look at the underlying architecture of V-Station, starting from the dispatcher. A dispatcher is a module that decouples the virtual machine tasks and Q tasks. Also, V-Station has an API module that provides this HTTP based service externally. The API is responsible for external interfaces and tasks, step decomposition, while the entire services framework communicates through the messaging queue MQ, where the API receives messages and sends them to the MQ. In addition, there are also different modules that handle the different components within the Cloud platform, such as compute access, image access, network, scheduler, resource, error, and debug. For example, the scheduler handles the module scheduling function in the V-Station platform. While the resource module provides resource management functions for virtual and host machines. This table compares the OpenStack and V-Station Cloud management platforms. As you can see, V-Station not only greatly reduces the amount of code, but also supports larger cluster sizes, more effective disaster recovery, and better performance. First of all, a mature Cloud platform is inseparable from the underlying infrastructure architecture. In Tencent Cloud, the underlying Cloud computing infrastructure is backed by global cross-regional IDCs that are able to sustain the necessary workloads and services. Additionally, the network infrastructure provides a strong resource foundation for the upper layers that provide resources for computing, network, storage, database, and the deployment of various applications on top. Secondly, based on its large-scale Cloud computing infrastructure, Tencent Cloud can deliver computing products like CBMs. These resources also provide elastic computing, container services, batch compute, and load balancing functions. Thirdly, Tencent Cloud Storage Service supports object storage, file storage, block storage, as well as other storage-related services like log service. Fourthly, the globally distributed infrastructure allows for a large variety of network resources, such as private networks, dedicated line access networks, cross-regional access, IP infrastructure, NAT gateways, and VPN connections. Lastly, managed database services offer relational databases, document databases, columnar databases, distributed Cloud databases, as well as support for elastic caching and migration services through DTS. This figure depicts the Cloud product information that corresponds to each Cloud development stage in theory. In reality, however, the Cloud development stages are more complicated than this figure shows. Each stage may correspond to more than one Cloud product. Typically when customers move their businesses into the Cloud, they can use CBM for servers, CLB for load balancing, CBS for storage, and Tencent dB for MySQL for database management. However, when developing Cloud native applications, they can choose to wrap the application package into Container Services or serverless functions. As of March 2022, Tencent Cloud operates 70 availability zones AZs in 26 geographical regions, providing strong technical support to more customers and helping them rapidly expand their businesses. Let's move on to introduce the concept of regions and availability zones. A region refers to the geographic area of the physical datacenter. Datacenters are completely isolated from each other to ensure maximum stability and fault intolerance within the same region or across regions. Currently, Tencent Clouds' coverage in China can be classified into South China, East China, and North China. Tencent Clouds Hong Kong, and Singapore datacenters cover Southeast Asia, while the Toronto datacenter covers North America. The Silicon Valley node covers the Western United States. In order to reduce access latency and increased download speeds, we recommend that you choose the region closest to your customers. On the other hand, an AZ is a physical IDC of Tencent Cloud with independent power supply and network resources within a region. There are multiple AZs in the same region which are interconnected over the low latency private backbone. When creating an instance, you can select any AZ in the specified region. You're also allowed to migrate instances to another AZ. Since some resources only work within a specific AZ unless you manually replicate or migrate them to another one. What are the advantages of having multiple regions in AZs? First of all, having different regions and AZs allows for high availability and disaster recovery. Secondly, choosing the region closest to your users will reduce their access latency. Thirdly, the compliance requirements for data management and data protection vary across regions. Lastly, the pricing of Cloud services also varies across different regions. We recommend that you follow several best practices when using Cloud services. As we discussed earlier, when you purchase, CVM instances, we recommend that you select the region closest to your end-users to minimize the access latency and improve access speeds. When you choose other Tencent Cloud services, we recommend selecting the same region and AZ so that such services can communicate over the private network to reduce the access latency and improve access speeds. Even if there is only one virtual private Cloud, VPC, we still recommend that you deploy your business in different AZs to ensure cross-AZ fault isolation and implement cross-AZ disaster recovery. In addition, because there may be network communication delays and different available intervals, the actual needs of the business must be evaluated to strike a balance between high availability and low delays.
