Docs

Elastic Compute Service

Elastic Bare Metal Server

Big Data EBM

Overview

By referring to this topic, you can select an instance type that meets your business needs. This helps you ensure that the selected type meets the minimum CPU and memory requirements for your business.

warning

For information about instance types, estimated instance fees, instance type changes, and instance naming rules, please refer to Instance types and specifications.

For information on Elastic Bare Metal (EBM) instances, please refer to EBM overview.

Instance type	Applicable scenarios
Big Data EBM ebmd3s	Storage-intensive big data scenarios such as Hadoop MapReduce, HDFS, Hive, and HBase Machine learning scenarios such as Spark in-memory computing and MLlib Searching and log data processing such as Elasticsearch and Kafka

Limits

The instance type cannot be changed.

The specifications and lifecycle of local volumes are tied to the ECS instances and cannot be purchased separately. Local volumes cannot be separately mounted, unmounted, or scaled up.

When you delete an instance, all the data stored in the local volume is automatically cleared. Please back up data before you delete the instance.

A physical server failure or local volume damage may cause data loss. Therefore, do not store the data for long-term use on a local volume and back up data timely.

We recommend that you use a high-availability architecture, deploy your services across multiple instances, or use cloud volumes for data backup. These measures ensure high availability of your services and data reliability.

If the data architecture of your application is not reliable, we recommend that you use cloud volumes for data backup.

Big Data EBM ebmd3s

tip

This instance is currently in Beta in the China (Beijing) region. If you need a trial, please contact your account manager to apply.

Currently, you can use the following public images only:

Ubuntu 22.04

veLinux 1.0/2.0

Features

Instance type	Performance
Computing	Processor: the 4th-generation Intel® Xeon® Scalable processor (Sapphire Rapids) with a processor base frequency of 2.6 GHz and an all-core turbo frequency of 3.1 GHz. The processor-to-memory ratio is 1:5.3.
Storage	You can use ESSD volumes. A maximum of 16 volumes (including 1 system volume) can be mounted. Local HDDs with high capacity and throughput are configured, providing local storage of up to 24,000 GB * 24.
Network	Maximum network bandwidth: 96 Gbit/s Maximum throughput: 18,000,000 PPS

Instance types and metrics

tip

The maximum throughput (PPS) varies significantly in different business scenarios. Therefore, we recommend that you conduct a business stress test to learn the instance performance. After that, you can select an appropriate instance type.

Instance type	vCPU	Memory (GiB)	Local storage (GB)	Maximum network bandwidth (outbound + inbound, Gbit/s)	Maximum throughput (outbound + inbound) (10,000 PPS)	Connections (10,000)	Maximum queues per network interface	Elastic network interfaces (including 1 primary network interface)	Private IP addresses for one network interface	Volume IOPS (unit: 10,000)	Volume bandwidth (Gbit/s)	Volumes (including 1 system volume)	Maximum queues per volume
ecs.ebmd3s.24xlarge	96	512	24,000 * 24	96	1,800	400	32	15	30	24	16	16	4

Last updated: 2026.06.18 16:22:25