JMR (JD MapReduce) is a hosted cluster platform that can run big data frameworks (such as Apache Hadoop and Apache Spark) on JD Cloud or run the open source projects (such as Apache Hive and Apache Pig) related to these frameworks. You may use JD MapReduce, Hive, Spark, Presto and other services conveniently to carry out large-scale distributed Data Compute and mass data analysis at low cost.

Try Now


Simple and User-friendly
The hardware model (CPU, memory and disk), software portfolio and version, and network environment are selected based on its business scale, and the whole process is automatically deployed. It usually takes a dozen minutes to create a cluster.
Cost Saving
The cluster can be created according to requirements. In other words, the cluster is released as soon as the off-line operation ends. You can add nodes dynamically as needed or purchase the service based on volume for long-term use.
Deep Integration
It is deeply integrated with other JD Cloud products, e.g. Cloud Storage, Monitoring and Big Data Analysis Platform, as the input source or output destination of Hadoop/Spark computing engine in JMR product.
Security and Reliability
Each cluster is isolated through VPC, and the firewall is automatically configured to manage network access; it supports the automatic configuration of network ACL and Security Group. The cluster has its own high availability scheme and provides the cluster monitoring management feature.


Cluster Management

Flexible Creation of Cluster

According to the hardware configuration, storage, and software of the user specified node, the cluster is automatically created in one-click mode, and the firewall settings and EIP association are automatically completed.

Flexible Expansion of Cluster

As the business volume increases, the cluster can be expanded at any time.

Flexible Release of Cluster

If you are temporarily computing a mass amount of data, you can release the associated cluster at any time.

All-round Monitoring

Main Indicators of Cluster

Provide monitoring for the availability and performance of cluster, to help users discover problems and solve such problems. It also meets user’s management requirements such as adjustment and release of cluster resources.

Cluster Node Indicator

Monitor the network status, hard disk status, etc. of each node in the cluster.

Cluster Service Process

Monitor based on the relevant components in your cluster, whether monitoring service nodes are working normally or not.

Job Management

Configure Job

Optional clusters, optional tasks run on different clusters. View the details of the operations and the history execution log of the tasks.

Deploy Job

The scheduling policy is divided into single scheduling and periodic scheduling during deployment. It can give the customer the flexibility to choose the time at which the job will be implemented automatically.


Off-line Analysis

Support processing of various logs generated by all kinds of application programs. Help the user dig and analyze unstructured data or semi-structured data. Support mass data ETL and data extraction, conversion and loading on large datasets. Support processing of customized data. Meet the application demands of mass data ETL.

Stream Compute

Data in the production environment is sent to the Message Queue. The Stream Compute engine will continuously process the data in the Message Queue in real time, and produce new data or computed results. The output of Stream Compute can be stored in the cluster or on the external cloud server.

One-stop Big Data Platform

JD MapReduce_BD-OS1.0 is a platform-level big data product. The platform provides multi-source heterogeneous data acquisition modules, real-time/offline computing framework, simple and easy-to-use development environment and platform interface, with big data management, development and computing capabilities to support the construction of data warehouse, user profiling, knowledge graph, deep learning, text analysis and other applications at enterprise-level.

Customer Case

  • Qiandama

    Specialized in fresh meat and vegetable market, the company is based on the core of meeting the customers’ demands for fresh and health food products.

    Based on JD MapReduce to build a big data platform, using JD artificial intelligence and big data technology to provide intelligent replenishment and sales forecast.


    It is a professional brand chain company, aiming to become the leading brand of the snack.

    Based on JD MapReduce to build a big data platform, using JD artificial intelligence and big data technology to provide intelligent replenishment and sales forecast.