Introduction: An Integrated Solution to Address Challenges in Digital Transformation
SHENZHEN, China, April 28, 2025 /PRNewswire/ -- In the era where LLM driven intelligence prevails, enterprise data platforms face emerging challenges from LLM application and AI development. To name a few, efficiency loss owing to segmented AI/Data development process, data governance for heterogeneous data, traditional architecture that struggles to meet the needs for real-time data and rapid AI integration. Tencent Cloud WeData addresses these complexities through flexible architecture and product innovation, presenting an integrated data intelligence platform for enterprises to navigate these challenges and prosper in the age of AI.
4 Unique Aspects of the Integrated Data+AI Solution
-- Unified Storage Layer: A unified Lakehouse storage system with a
centralized catalog service that supports structured/unstructured data
and model assets, enablingcross-engine queries .
-- Unified Compute Layer: A serverless resource pool to integrate
frameworks like Spark, Flink, and PyTorch, improving CPU/GPU resource
utilization by 30%.
-- Integrated Streaming-Batch-Lakehouse Engine: The Setats engine
dynamically switches between streaming, batch, and incremental compute
modes, merging real-time states with lake storage to reduce storage
costs by 50%.
-- Unified Catalog: A centralized catalog to manage structured/unstructured
data, enabling seamless access across compute engines (Spark, Flink) and
AI frameworks (TensorFlow, PyTorch).
-- Global Asset View: Integration of metadata for tables, volumes, models,
and applications, supporting semantic search across all assets.
-- Federated Permissions: One single authorization to grant data access
across engines like Spark and Ti-One, reducing policy propagation time
from hours to minutes.
-- Intelligent Semantic Enhancement: Leverages LLMs to auto-generate data
description, perform intelligent asset classification and maintenance,
and improve documentation completeness by 80%.
Built on the OneOps philosophy, our solution combines DataOps and AIOps to deliver three key capabilities:
-- Full-Lifecycle Standardization: Automates enterprise CI/CD pipelines for
end-to-end management of data and models.
-- End-to-End Observability: Automatically captures lineage from raw inputs
to AI model outputs, enabling rapid root-cause analysis, change impact
assessment, real-time monitoring, and audit logging.
-- Governance-by-Design: Embeds standardized development practices,
AI-oriented data modeling, and enterprise semantic layers to
systematically accumulate digital assets.
Ensures security across the entire lifecycle--data preparation, model training, and LLM applications:
-- Data Security: Global access control, intelligent data
classification/categorization, sensitive data masking/encryption, and
audit trails.
-- Model Security: Detects and defends against adversarial sample attacks
during training; embeds watermarks to prevent model theft; blocks
discriminatory/harmful AI-generated content during inference via rule
engines.
-- Application Security: API protection (e.g., unauthorized access
prevention) and content moderation (e.g., forgery detection, sensitive
content filtering).
#Tencent Cloud Big Data #WeData
View original content:https://www.prnewswire.co.uk/news-releases/tencent-cloud-big-data-wedata-an-integrated-data--ai-intelligent-platform-302438247.html [https://www.prnewswire.co.uk/news-releases/tencent-cloud-big-data-wedata-an-integrated-data--ai-intelligent-platform-302438247.html]
Tencent CloudCONTACT: Lisa ling, xiao.ling@bluefocus.com
