Interview with Tencent Cloud Architect Liu Yong on Enterprise Cloud Migration and High Availability

Introduction: In the cloud‑native era, moving to the cloud has become a primary demand, but many common problems and misconceptions exist. Ahead of the GIAC 2020 Global Internet Architecture Conference, High Availability Architecture interviewed Liu Yong from Tencent Cloud to explore key enterprise cloud migration issues.

High Availability Architecture: Liu Yong, could you introduce yourself to the readers?

Liu Yong: I have been in the IT industry for over 10 years, starting with network work. I participated in the design and construction of China Telecom's 163/CN2 backbone, later worked at Cisco on SP Routing products, supporting backbone construction for BAT and telecom operators. After leaving traditional network companies, I moved to the internet industry, working on cloud computing at Kingsoft Cloud, leading underlay network architecture and team management, and later transitioned to front‑end and cloud customer pre‑sales.

In 2018 I joined Tencent Cloud, responsible for architecture support for internet customers, focusing on technical transformation pre‑sales. I have also been selected as a Tencent Cloud WeExpert expert, aiming to promote cloud computing technology through practice and sharing.

High Availability Architecture: What are the main concerns for enterprises moving to the cloud, and what advantages does Tencent Cloud offer?

Liu Yong: The first issue is making the cloud "accessible" for enterprises. Just like people adapting to a new environment, business systems need an adaptation period. Tencent Cloud provides automated environment analysis, industry‑specific migration solutions, data and environment synchronization tools, and intelligent product specification selection, helping enterprises migrate smoothly. Products such as the Migration Service Platform, Data Transmission Service, and Cloud Data Migration, as well as on‑site expert support, are typical examples.

High Availability Architecture: How does cloud‑native development and deployment differ, and what common misconceptions do enterprises have?

Liu Yong: The change is mainly in mindset and approach. The concrete technologies are familiar to industry veterans; the shift is toward reusing and composing generic cloud capabilities rather than building custom wheels. Exposing requirements and problems drives service providers to iterate their products, creating a win‑win situation.

High Availability Architecture: For users migrating from private deployments to the cloud, what should they consider regarding availability, and how does Tencent Cloud support this?

Liu Yong: Focus on the rational combination of cloud products and features. Move from worrying about low‑level atomic high‑availability to understanding high‑availability principles of individual products and then ensuring the overall architecture’s availability through multi‑product composition.

Tencent Cloud offers highly available products at every architectural layer—access, logic, and data. Examples include seamless host migration, Anycast EIP based on Tencent Cloud backbone, and MySQL read/write separation with cross‑region disaster recovery, all typical cloud‑native high‑availability solutions.

High Availability Architecture: Could you share your outlook on future cloud platform trends?

Liu Yong: Many areas are evolving rapidly. Service Mesh is gaining a lot of attention, as are data governance and fine‑grained network traffic scheduling. With increasing usage of cloud‑native services, serverless combined with cloud products will also attract more focus.

High Availability Architecture: When an entire availability zone fails, how should customers respond? Do you recommend multi‑cloud or partial private deployments?

Liu Yong: Although the probability of an entire zone outage is decreasing due to continuous architectural and product improvements, the impact remains significant and must be considered in design. High availability involves a trade‑off with investment; as cloud products become more redundant, many now natively support cross‑zone disaster recovery, greatly enhancing resilience.

Architecturally, cross‑zone designs within the same region are essential, ensuring balanced service provision under normal conditions and automatic failover to healthy zones during incidents.

Whether to adopt multi‑cloud or a cloud‑plus‑IDC model depends on factors such as IT investment scale, business size, and in‑house capabilities. For most cases, we recommend selecting a primary cloud platform, deploying cross‑zone and cross‑region disaster recovery on it, and focusing resources on business logic rather than managing multiple heterogeneous environments.

High Availability Architecture: What will you share at the upcoming GIAC conference?

Liu Yong: My talk will focus on selecting appropriate cloud products for different deployment scenarios and locations, helping attendees understand how to build high‑availability architectures with the myriad cloud services available.

High Availability Architecture: Any message for the GIAC audience?

Liu Yong: I hope GIAC provides a platform for discussion and sharing, enabling everyone to progress together, and I wish the pandemic ends soon.

Technical original and architecture practice articles are welcome for submission via the public account menu "Contact Us".

Reference Reading:

How to evaluate the recent top‑tier open‑source deep learning framework OneFlow?

Programmers: How to objectively measure business ideas?

Spotify's three‑step technology upgrade strategy

Programmers' thinking tool: Reverse engineering

How programmers can train data thinking (with metric scorecards)

Follow the "High Availability Architecture" public account by scanning the QR code below.