The logical architecture of EMR is as shown below:
An EMR cluster consists mainly of three parts: open-source components, Tencent Cloud infrastructure, and cluster management.
Open-source components
EMR integrates dozens of cutting-edge open-source big data components from the Apache community, such as Hadoop, Hive, Spark, HBase, Presto, Flink, Alluxio, and Iceberg. For more information, see Product Releases and Component Versions.
Based on optimized Iceberg, Alluxio and other open-source components, EMR offers Iceberg Z-Order algorithm, Alluxio transparent URI and other enhanced features.
Tencent Cloud infrastructure
EMR can be deployed based on multiple types of underlying computing resources, including CVM and CBM. It supports containerized deployment.
Data can be stored in a local disk, cloud disk, COS bucket, or CHDFS instance.
VPCs, network ACLs, and security groups provide EMR with a securely isolated network environment.
Cluster Management
EMR supports smart cloud deployment management, including fast creation, flexible scaling, and auto scaling.
EMR offers a great variety of convenient Ops tools, such as service configuration management, batch node management, and visual service Ops.
EMR provides complete cluster monitoring and diagnosis capabilities ranging from multidimensional metric monitoring, event, and inspection to alarming and log search.
