Alibaba Cloud Media and Entertainment China Translated from Chinese

People's Posts and Telecommunications Press (人民邮电出版社有限公司)

People's Posts and Telecommunications Press had already launched multiple AI scenarios across content production, reader services, and public-interest education, but it needed one governed way to deploy and call Qwen3-family models across departments. The publisher wanted stable large-model access, unified scheduling and permissions, full-path observability, and a repeatable architecture template so new AI use cases could move from discussion to production without ad hoc infrastructure work.

Read the original story Explore all use cases

Reported impact

Business outcome

People's Posts and Telecommunications Press gained a faster path from AI idea to production deployment. Alibaba Cloud gave the publisher a visual, repeatable architecture for standing up Qwen-based services, unified access control for model endpoints, and observability needed to keep multi-model experimentation stable. That lets the team align full AI architectures more quickly, validate multi-model combinations with less manual infrastructure work, and accelerate AI innovation across publishing and reader-facing scenarios.

Solution

The publisher adopted Alibaba Cloud's Platform for AI (PAI) as an enterprise LLM foundation rather than provisioning each scenario separately. PAI Lingjun supplies the AI compute pool, while Model Gallery provides one-click deployment for the full Qwen3 family so teams can stand up model endpoints without rebuilding serving stacks. Around the model layer, ARMS-Prometheus adds end-to-end observability across resource usage, dashboards, and alerts, giving the team a single telemetry surface for AI workloads. AI Gateway becomes the unified external access layer, concentrating traffic governance, authentication, and authorization at the API edge and even converting existing HTTP APIs into MCP servers so older business capabilities can be reused inside new AI workflows. CADT large-model templates are then used to visualize, validate, parameterize, and batch deploy the architecture, turning LLM infrastructure into a repeatable publishing platform instead of a one-off project.

Data flow

Publishing and reader-service applications call AI Gateway, which applies governance and routes traffic to Qwen3-family deployments running on PAI and PAI Lingjun. ARMS-Prometheus observes the full path, while CADT templates define and batch deploy the architecture so internal teams can reuse the same AI platform for new scenarios.

Solution architecture

6 components · 4 layers

Compute
- Platform for AI (PAI) Provides the enterprise foundation for model deployment, compute management, and permissions across AI workloads.
- PAI Lingjun Supplies the AI compute resources used to run Qwen3-family model deployments.
Serving
- AI Gateway Presents a unified model-access entry point with traffic governance, authentication, and HTTP-to-MCP conversion.
Orchestration
- Model Gallery Lets the team deploy supported Qwen3 models quickly without rebuilding each serving stack from scratch.
- Cloud Architecture Design Tool (CADT) Provides visual architecture validation, parameter adjustment, and batch deployment for the large-model stack.
Governance
- ARMS-Prometheus Adds end-to-end observability and multi-dimensional analysis for AI resource and service behavior.

Architecture clues

AI Gateway centralizes traffic governance, access control, and HTTP-to-MCP conversion at the API boundary.
ARMS-Prometheus provides one observability surface for AI resource usage and service behavior.
CADT templates are used to visualize, validate, parameterize, and batch deploy the large-model architecture.
PAI and Model Gallery form a reusable deployment layer for the Qwen3 model family rather than one-off endpoints.

Evidence from the source

Alibaba Cloud proposed an enterprise visual deployment plan on PAI that uses Lingjun compute, AI observability, AI Gateway, and CADT to support one-click deployment of the full Qwen3 family. (original: 采用基于人工智能平台 PAI 的企业级 AI 可视化部署方案。利用 PAI 灵骏智算资源，结合 AI 可观测、AI 网关、云速搭 CADT 等云产品，支持 Qwen3 全系列模型一键部署。)

PTPress said the PAI and CADT combination lets the enterprise visually build complete AI cloud architectures with upstream and downstream dependencies, improving deployment efficiency and accelerating innovation. (original: 基于阿里云 PAI 平台与云速搭 CADT，企业能够以可视化方式构建包含上下游依赖的完整 AI 云架构，实现架构方案的快速对齐与多模型的高效调试验证，从而提升部署效率和加速 AI 创新。)

The implementation uses AI Gateway to centralize traffic governance and authentication at the API layer while converting existing HTTP APIs into MCP servers for rapid business innovation. (original: 基于 AI 网关，将流量治理、认证鉴权等管控能力集成至 API 层，统一对外提供模型访问入口；并提供 HTTP to MCP 的转换能力，将现有 API 自动转换为 MCP Server，支撑业务快速迭代与创新。)

The publisher said it already had multiple AI scenarios internally and externally and wanted to use Qwen3-family models to support unified scheduling and stable growth of those businesses. (original: 在公司内外部已经落地多个应用场景。希望利用 Qwen3 等模型能力，满足不同业务场景对模型的调用需求，支撑 AI 业务统一调度管理、稳定发展等。)

Transparency Evidence & methodology 75/100 reference value · 93% confidence · 2 sources

Review how this reference was scored and trace each normalized claim back to its documented source.

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value75/100

Problem clarity20/20
Measures whether the customer's operational problem is specific enough to compare.
Implementation depth25/25
Rewards named components, data flow, integrations, and architecture detail.
Outcome specificity20/20
Measures whether the reported benefit is concrete rather than a generic claim.
Quantified impact0/25
0 metrics describe measurable impact.
Deployment scale10/10
Rewards evidence of a live or scaled deployment and meaningful integration scope.

Evidence confidence93%

Source authority20/25
The main evidence is an official vendor-published customer story.
Claim traceability25/25
12 of 12 key fields link to a named source.
Direct evidence20/20
4 direct source quotes retained.
Date provenance10/10
The publication date comes from the source page or publisher.
Source transparency10/10
2 source references document publisher, URL, authority, retrieval, and supported fields.
Implementation specificity8/10
Rewards source-backed components, data flow, quantified impact, and a specific outcome.

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

Alibaba Cloud
云速搭助力人民邮电出版社大模型业务快速上云_文化_阿里云客户案例

First-party
https://www.aliyun.com/customer-stories/culture-2025-ptpress

Source type
Vendor-published customer story
Supports
- Customer
- Region
- Business problem
- Solution
- Products
- Solution components
- Data flow
- Integrations
- technical area
- use case category
- Business outcome
- outcome category
- Metrics
- Architecture
- evidence quotes
人民邮电出版社
人民邮电出版社 - 人民邮电出版社

First-party
https://www.ptpress.com.cn/
The official site states the press is "隶属于中国工信出版传媒集团，由工业和信息化部主管", supporting classification as a public-sector organization.

Source type
Official company website
Supports
- Company size

Date evidence

Published on source page Dec 1, 2025 · Alibaba Cloud · First-party
Open date source

Not separately sourced Industry. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Similar use cases

See all like this

Alibaba Cloud Archived · Dec 8, 2025 and factory, inc Media and Entertainment Japan With domestic e-book growth slowing, Japan’s manga industry faces a pivotal shift toward global expansion. The success depends on transforming manga for a global audience, which requires sophisticated localization,… Manga production tasks that previously required 3 days were now reduced to just 1 day , freeing creators from… System Migration Auto Scaling · Container Registry · Elastic IP Address

Alibaba Cloud Nov 4, 2025 Morefun Entertainment Group (魔方文娱集团) Media and Entertainment China Morefun's ticketing and event platforms needed to hit a 1-minute detection, 5-minute diagnosis, and 10-minute recovery target, but its older monitoring stack had blind spots, alert storms, fragmented trace and log… 60% of common faults now support automatic detection, diagnosis, and repair IT Operations Automation Application Real-Time Monitoring Service (ARMS) · Cloud Dial Test · Grafana

Alibaba Cloud May 7, 2025 XD / TapTap (心动 / TapTap) Media and Entertainment Global TapTap's ad-driven recommendation, search, and content workflows depended on machine-learning models and NLP processing, but the older ECS-plus-Rundeck environment made deployments and service management too slow and… 38 games operated as of mid-2022 IT Operations AutomationMarketing Personalization Elastic Algorithm Service (EAS) · Function Compute (FC) · Kafka