What's MapReduce Service?
Spin-up and manage high availability big data clusters with scalable power to run open source analytics workloads
-
Fast Provisioning and Easy Maintenance
Rapidly provision big data clusters with automated software installation and configuration, integrated rolling patches and upgrades, and simplified operations and maintenance.
Rapidly provision big data clusters with automated software installation and configuration, integrated rolling patches and upgrades, and simplified operations and maintenance.
-
Decoupled Storage and Compute
Increase flexibility and reduce costs by using compute clusters for analysis, while storing data without limits on low-cost Object Storage Service (OBS) with 99.9999999999% durability.
Increase flexibility and reduce costs by using compute clusters for analysis, while storing data without limits on low-cost Object Storage Service (OBS) with 99.9999999999% durability.
-
End-to-End Security
Manage cluster project permissions with Kerberos authentication, and keep users and resources secure with Identity and Access Management (IAM) service. Data is encrypted in transmission and while in storage.
Manage cluster project permissions with Kerberos authentication, and keep users and resources secure with Identity and Access Management (IAM) service. Data is encrypted in transmission and while in storage.
-
High Availability and Performance
No single points of failure with high availability across availability zones. Quickly scale up instances and/or RAM and CPU in clusters to unleash massive computing power.
No single points of failure with high availability across availability zones. Quickly scale up instances and/or RAM and CPU in clusters to unleash massive computing power.
Major Releases
Apache Pulsar, a next-generation distributed messaging and streaming platform
Apache Pulsar is a next-generation messaging system with decoupled storage and compute. This high performance cloud native service has unified and streaming models. It is highly reliable, scales easily, and is easy to maintain. Huawei Cloud MRS integrates the enterprise-class Apache Pulsar messaging system to provide you with a fantastic option for message queue and stream processing platforms.
- Application scenarios
- Creation of message queues
- Stream data processing
- Technical advantages
MRS provides a cloud native architecture for hybrid cloud with real-time, offline, and logical data lakes; and a one-stop big data platform for public cloud that is inexpensive, flexible, secure, and reliable. Decoupled storage and compute allow for more flexible scalability, and you can take advantage of multi-tenant configuration, tiered storage, and a range of flexible subscription modes (exclusive, active/standby, and shared).
ClickHouse (enterprise edition)
ClickHouse is an excellent real-time analytical database with a stellar compression ratio and lightning fast query. With reliable security features from Huawei Cloud and multi-architecture computing (x86 and Arm), ClickHouse can respond to queries of tens of billions of records from hundreds of dimensions in milliseconds. It is your ideal choice for building a high-performance massive data analysis warehouse on the cloud.
- Application scenarios
- Real-time analytics of massive amounts of data in flat-wide tables
- Real-time BI report analytics
- User behavior analytics
- Gaming operations data analytics
- Technical advantages
Ultimate performance, multi-architecture computing, security and reliability, smooth and elastic scaling, flexible configuration, and expert support
Use Cases
Use a range of big data components to process, analyze, query, or mine huge amounts of data
Rapid Migration
Migrate over from other big data platforms to MRS in just a few steps. A set of migration tools for each component is provided to minimize errors and service interruption.
Advantages
Unified Data Storage
Eliminate data silos with a combination of real-time, offline, and logical data lakes. One copy of data can be shared and analyzed by multiple services.
Elastic Scaling
Optimize your instances and configure usage-based automatic scaling policies, greatly reducing the cost of migrating to the cloud.
Open Source Compatibility
MRS is fully compatible with open source APIs. During migration, your services are not affected and no service code needs to be modified.
Internet of Vehicles (IoV)
MRS leverages compatibility with open source APIs to provide a fast and efficient data processing computing engine that allows you to analyze huge amounts of car status, telemetry, and user experience data.
Advantages
Unified, Full-Stack, Scalable Data Platform
MRS is an enterprise-level, cloud native, big data platform with decoupled storage and compute, for more convenient, more flexible scalability.
Multi-engine Processing for Hybrid Workloads
MRS provides open source components that can be combined freely as needed, supporting real-time/offline complex service processing.
High Performance at a Low Cost
Storm can obtain real-time stream data from Kafka for real-time computing and analysis with high throughput and low latency.
Finance and Insurance
MRS meets the strict requirements of the insurance industry with regards to compliance, security, and reliability. A traditional architecture can be rapidly rebuilt and deployed for insurance enterprises that need to transform fast. Digital transformation makes innovating and evolving services easier and faster.
Advantages
Robust Security
Meets industry regulatory requirements and protects customers' sensitive data.
Dedicated Resources
Provides dedicated MRS clusters and exclusive resources, and decouples compute resources from storage.
Flexible Creation, Full-Stack, Easy O&M
Allows users to create a full-stack big data platform with just a few clicks and provides an enterprise-class platform management interface, simplifying O&M.
Smart Logistics
MRS is used for intelligent management of logistics and supply chain routes, improving service operation efficiency and greatly reducing costs.
Advantages
High Throughput and Low Latency
Dedicated high-throughput, high availability, and low latency MRS Kafka clusters facilitate real-time access for millions of messages.
Large-scale Data Analysis and Fast Processing
MRS Spark supports large-scale data computing. MRS HBase can load and update logistics data in milliseconds, and query and analyze petabytes of time series data.
More Intelligence with AI
MRS uses AI for big data mining, and provides precise and intelligent prediction and analysis capabilities for logistics organizations, marketing, and operation management.
Internet of Elevators (IoE)
MRS is an open one-stop big data processing platform for intelligent elevator management. The platform is flexible and customizable enough to adapt to nearly every scenario.
Advantages
Open and Flexible
MRS provides a wide variety of computing products and storage hardware to meet specific needs and help you build a big data platform that is unified and open. It offers powerful computing engines, an unlimited amount of storage, and enables flexible integration of service components.
High Performance and Large Capacity
MRS Kafka works with high-performance general network enhancement (C3ne) ECSs to provide real-time data access for millions of elevators.
AI Supported
MRS supports GPUs that provide real-time high-speed parallel processing and floating-point computing capabilities, which is useful for encoding and decoding, deep learning, and scientific computing.
Smart Water Management
MRS Hadoop provides reliable, high-performance big data storage and analysis for intelligent water management.
Advantages
Unified and Scalable Data Platform
MRS gives you an enterprise-level big data platform with open source components that can be flexibly stacked, supporting real-time or offline complex service processing.
High Throughput and Low Latency
Storm can obtain real-time stream data from Kafka for real-time computing and analysis with high throughput and low latency.
Integration of Various Types of Data
Structured, semi-structured, and unstructured data can be computed and processed, and traditional data warehouse data can be easily migrated, facilitating cross-source data exploration and analysis.
Real-time Gaming
Game log data can be accessed through Kafka and Flume in real time. Spark Streaming then processes and analyzes the data in real time and stores the analysis results to HBase or Hive for quick game advertisement analysis, data query and analysis, and revenue analysis.
Advantages
Unified and Scalable Data Platform
MRS offers an enterprise-level big data platform with open source components that can be flexibly stacked, to meet highly complex service processing needs.
Real Time and High Throughput
MRS Kafka and Flume collect real-time data and integrate with high-performance general network enhancement (C3ne) ECSs for real-time access of massive amounts of data.
Smart Power Management
MRS provides enterprise-level big data cloud services that enable power plant operators to use Hadoop, Spark, HBase, Storm, and other big data components for predictive device maintenance.
Advantages
Unified Big Data Platform
MRS gives you an enterprise-level big data platform with open source components that can be flexibly stacked, supporting real-time or offline complex service processing.
Massive Data Ingestion
MRS Kafka and Sqoop support multiple data ingestion methods, facilitating real-time ingestion of millions of messages.
Easy of Integration
SQL APIs can be used to query multi-dimensional data for easy data exploration and analysis.
Cluster Options
Quickly deploy from a pre-configured cluster template with pre-installed components
-
Hadoop Analysis Cluster
Spark allows for analysis and query of vast amounts of data and use Hive to analyze terabytes, or even petabytes, of data.
Spark allows for analysis and query of vast amounts of data and use Hive to analyze terabytes, or even petabytes, of data.
-
HBase Cluster
HBase is used to store massive amounts of data and query data in milliseconds.
HBase is used to store massive amounts of data and query data in milliseconds.
-
Kafka Cluster
Flume is used for real-time data ingestion, Kafka is used for real-time access of tens of thousands of data records, and Storm is used for reliable, fault-tolerant, and low-latency online service data processing.
Flume is used for real-time data ingestion, Kafka is used for real-time access of tens of thousands of data records, and Storm is used for reliable, fault-tolerant, and low-latency online service data processing.
-
ClickHouse Cluster
ClickHouse is used to query and analyze massive amounts of data in real time, accelerating data value extraction.
ClickHouse is used to query and analyze massive amounts of data in real time, accelerating data value extraction.
New Features
Success Stories
Meitu
FusionInsight uses decoupled storage and compute to enable flexible and independent expansion of compute and storage resources on demand. An EC algorithm replaced Hadoop big data 3-copy so that one data copy can support multiple types of analysis and computing, achieving 40% higher resource utilization and reducing costs by 30% overall. This helps Meitu's big data service efficiently serve 2+ billion users worldwide.
China Merchants Bank
"The Huawei Cloud FusionInsight big data platform significantly boosted innovation in CMB. Now we can provide better on-demand, real-time online services. In the past, we could only query historical data from the past 13 months. Now we can query data from the past seven years. We can also reach all our target users with 82% fewer SMS messages than before."
— Liu Jing, Manager of Next-Gen Cloud Computing Project, Information Technology Department, China Merchants Bank
T3Go
The lakehouse of T3Go features decoupled storage and compute. It is built based on the open source Hudi framework, and applies to both BI and AI. Currently, our lakehouse is hosted on Huawei Cloud FusionInsight.
— Yang Hua, big data platform owner of T3Go