MapReduce Service (MRS)

MapReduce Service (MRS) provides enterprise-level big data clusters on the cloud, which are fully controlled by tenants and support the Hadoop, Spark, HBase, Kafka, and Storm components.

Enjoy the product for 1 year by paying for only 10 months. Learn more

Product Advantages

Enterprise-class

Supports Kerberos authentication to satisfy enterprises' demand for high security. Provides a level-based tenant model and allows tenants to be added dynamically.

Easy-to-use

Simplifies operation and maintenance (O&M) with a reliable, self-tuning, and easy-to-use O&M platform.

Stable

Supports automatic failover to provide 99.9% service availability. Stores multiple data copies, ensuring up to 99.9999% data reliability.

Cost-effective

Separates computing and storage resources. Computing clusters are created on-demand and can be released after jobs have been executed.

Open

Embraces the open source big data ecosystem and is compatible with open source interfaces. Provides REST API and JDBC for easy interconnection with other services.

Industry Architecture

Application Scenarios

  • Analysis and processing of mass data

  • Storage of mass data

  • Streaming processing of low-latency data

Analysis and processing of mass data

Usage

analysis and processing of massive sets of data, online and offline analysis, and business intelligence

Characteristics

processing of massive data sets, heavy computing workloads, long-term analysis, and data analysis and processing on a large number of computers

Application scenarios

log analysis, online and offline analysis, simulation calculations in scientific research, biometric analysis, and spatial-temporal data analysis

Storage of mass data

Usage

storage and retrieval of massive sets of data and data warehouse

Characteristics

storage, retrieval, and disaster recovery of massive sets of data with zero data loss

Application scenarios

log storage, file storage, simulation data storage in scientific research, biological characteristic information storage, genetic engineering data storage, and spatial-temporal data storage

Streaming processing of low-latency data

Usage

real-time data analysis, continuous computing, offline and online message consumption

Characteristics

massive amount of data, low latency, high throughput, high reliability, flexible scalability, and distributed real-time computing framework

Application scenarios

streaming data collection, active tracking on websites, data monitoring, IoT analysis, and risk control

Function Description

Isolation Between Data Storage and Computing

Based on OBS data storage, users can use MRS on demand. Computing resources can be released after the users finish data processing, improving resource utilization and saving costs.

Distributed Processing

Uses the Hadoop, Spark, and HBase processing frameworks. Provides PB- or higher-level data processing capability, which is far better than that of traditional databases.

Job Management

Users can submit and manage tasks on the management console or using application program interfaces (APIs). The tasks are automatically performed and monitored, and the operation is easy.

Load Balancing

Uses the high availability (HA) mechanism for management nodes. Services are automatically switched to the standby node if the active node becomes faulty. Uses the automatic load balancing mechanism for core nodes. If a core node becomes faulty, other nodes automatically take over services of the faulty node. Supports elastic capacity expansion.

Developer Resources

API

Users can send HTTP/HTTPS requests for invoking HUAWEI CLOUD APIs to manage ECSs. This enables users to operate and monitor their applications, resources...

Registration