Help Center/ Data Lake Insight/ Service Overview/ Advantages
Updated on 2025-12-19 GMT+08:00
View PDF

Advantages

Pure SQL Operations: Zero Learning Curve

  • DLI offers standard SQL APIs, enabling you to perform massive data query and analysis using only SQL. Its syntax is fully compatible with ANSI SQL 2003 standards.
  • This significantly lowers the barrier for data analysts and business professionals, enhancing overall efficiency in data analysis.

Decoupled Storage and Compute: Efficient Resource Utilization

  • DLI decouples storage and compute workloads through its decoupling architecture, allowing flexible configuration of resources based on demand. This improves resource utilization and reduces costs.
  • The elastic resource pool supports multiple engines like Flink, HetuEngine, and Spark, further optimizing resource allocation efficiency.

Serverless Architecture: Full-Scenario Adaptability

DLI is fully compatible with Apache Spark and Apache Flink ecosystems and APIs, providing a unified serverless big data computing service for real-time, offline, and interactive analytics.

  • On-premises Spark/Flink applications can migrate to the cloud effortlessly, minimizing migration efforts and ensuring smooth transitions.
  • A batch-stream fusion framework delivers scalable, high-performance processing for TB to EB-level data, meeting diverse big data needs.
  • Deep optimizations in product core and architecture result in performance over 100x faster than traditional MapReduce models, with 99.95% SLA.

Enterprise-Grade Multi-Tenancy: Secure and Controllable

Compute resources are isolated by tenant, with granular data permissions at the queue and job levels, facilitating secure inter-departmental data sharing and management.

Cross-Source Analysis: No Data Migration Required

  • Supports multiple data formats and sources, including cloud-based (e.g., OBS, RDS, DWS, CSS, MongoDB, Redis), ECS-hosted databases, and on-premises databases.
  • Enables unified cross-source analysis without data relocation, accelerating enterprise-wide data insights and innovation.

Advantages Over Traditional Self-Built Hadoop Clusters

Compared to self-built Hadoop clusters, serverless DLI offers distinct advantages:

Table 1 Advantages

Advantage

Dimension

DLI

Self-Built Hadoop System

Low cost

Capital expenditure

Pay-per-use billing based on actual data scanned or CUH. Up to 50% cost savings.

Fixed resource allocation leads to significant waste and higher costs.

Elastic scaling

Kubernetes-based containerization enables seamless scaling.

Resource configurations are fixed and inflexible.

No Ops

O&M costs

Ready-to-use with a serverless architecture, eliminating the need for dedicated operations teams.

Requires skilled professionals for setup, configuration, and maintenance.

High availability

Cross-AZ disaster recovery ensures stable service operation.

Availability depends on self-managed infrastructure.

Ease of use

Learning curve

Low learning curve with pre-tuned parameters from thousands of projects over 10 years, plus visual tuning tools.

High learning curve due to hundreds of manual tuning parameters.

Supported data sources
  • Cloud: OBS, RDS, DWS, CSS, MongoDB, Redis
  • On-premises: self-built databases, MongoDB, Redis
  • Cloud: OBS
  • On-premises: HDFS

Ecosystem compatibility

Supports DLV, Yonghong BI, and FineBI.

Limited to big data ecosystem tools.

Custom images

Allows custom images to meet diverse business needs.

Not supported.

Workflow scheduling

Integrated DataArts Studio-DLF for efficient process management.

Relies on self-built scheduling tools like Airflow.

Enterprise-grade multi-tenancy

Table-level permission management, down to column granularity.

File-level permission management only.

High performance

Processing speed

Deep vertical optimization leveraging hardware-software integration delivers faster processing speeds.

Open-source version performance without additional optimizations.