Instance specification
Based on different service scenarios, Baidu Cloud Compute (BCC) is divided into different instance families, which consist of varying combinations of CPU, memory, storage, heterogeneous hardware, and network bandwidth. Within the same instance family, different specifications are classified according to different CPU-to-memory ratios. Each specification offers distinct compute, memory, and storage functions. You may select an appropriate specification based on the scale of applications to be deployed.
Note: Certain specifications have been discontinued. For more details, refer to Discontinued Specifications.
Instance family
| Instance type | Instance family | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Economy Instance Family | Economy e1 | |||||||||||
| Economy e2 | ||||||||||||
| General-purpose instance family | ||||||||||||
| General-purpose g7 | ||||||||||||
| General-purpose ga3 | RDMA network-enhanced g5re | |||||||||||
| General-purpose g6 | ||||||||||||
| General-purpose g5 | ||||||||||||
| General-purpose gr1 | ||||||||||||
| General-purpose g4 | ||||||||||||
| General-purpose ga2 | ||||||||||||
| General-purpose ga1 | ||||||||||||
| Compute-optimized instance family | ||||||||||||
| Compute-optimized c7 | ||||||||||||
| Compute-optimized ca3 | Compute-optimized c6 | |||||||||||
| Compute-optimized c5 | ||||||||||||
| Compute intensive ic5 | ||||||||||||
| Compute-optimized cr1 | ||||||||||||
| Compute-optimized c4 | ||||||||||||
| Compute intensive ic4 | ||||||||||||
| Compute-optimized ca2 | ||||||||||||
| Compute-optimized ca1 | ||||||||||||
| Compute intensive ica2 | ||||||||||||
| Compute intensive ica1 | Memory-optimized instance family | |||||||||||
| Memory-optimized m7 | ||||||||||||
| Memory-optimized ma3 | Memory-optimized m6 | |||||||||||
| Memory-optimized m5 | ||||||||||||
| Memory-optimized m4 | ||||||||||||
| Memory-optimized ma2 | ||||||||||||
| Memory-optimized ma1 | ||||||||||||
| High-frequency Instance Family | high-frequency memory-optimized hcm7 | |||||||||||
| Big Data Instance Family | Big data d1s | |||||||||||
| Big data d1 | ||||||||||||
| Local SSD Instance Family | Local SSD Type l7 (Public Beta) | |||||||||||
| Local SSD type l5d (Public Beta) | ||||||||||||
| Local SSD type l5 | ||||||||||||
| Local SSD type l3d | ||||||||||||
| Local SSD type la2 | ||||||||||||
| GPU Instance Family | Compute-optimized gn5 | |||||||||||
| Compute-optimized gn3 | ||||||||||||
| Elastic Baremetal Instance Family | Elastic Baremetal Local SSD Type l8 | |||||||||||
| Elastic Baremetal Local SSD Type la4 | ||||||||||||
| Elastic Baremetal Local SSD Type hcl7 | ||||||||||||
| Elastic Baremetal Local SSD Type l7 | ||||||||||||
| Elastic Baremetal General-purpose g7 | ||||||||||||
| Elastic Baremetal Local SSD Type la3 | ||||||||||||
| Elastic Baremetal Local SSD Type l5 | ||||||||||||
| Elastic Baremetal General-purpose g5 | ||||||||||||
| Elastic Baremetal Local SSD Type la2ne | ||||||||||||
| Elastic Baremetal Compute GPU type gn7s | ||||||||||||
| Elastic High-Performance Computing Cluster | Elastic high-performance computing cluster GPU type GN6T | |||||||||||
| Elastic high-performance computing cluster GPU type GN5 |
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.e1.c2m2 | 2 cores | 2 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.e1.c2m4 | 2 cores | 4 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.e1.c2m8 | 2 cores | 8 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.e1.c4m8 | 4 cores | 8 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.e1.c4m16 | 4 cores | 16 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.e1.c8m16 | 8 cores | 16 GiB | 3 Gbps | 1,500,000 | 4 |
| bcc.e1.c8m32 | 8 cores | 32 GiB | 3 Gbps | 1,500,000 | 4 |
Economy e2
Instance Family Characteristics
- Processor: AMD EPYC series
- Processor-to-memory ratios include 1:1, 1:2 and 1:4
- Costs reduced by over 50% compared to industry-standard instances
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Construction of small and medium-sized websites
- Development and testing
- Lightweight applications
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.e2.c2m2 | 2 cores | 2 GiB | Up to 1.5 Gbps | Up to 100,000 | 2 |
| bcc.e2.c2m4 | 2 cores | 4 GiB | Up to 1.5 Gbps | Up to 100,000 | 2 |
| bcc.e2.c2m8 | 2 cores | 8 GiB | Up to 1.5 Gbps | Up to 100,000 | 2 |
| bcc.e2.c4m4 | 4 cores | 4 GiB | Up to 3 Gbps | Up to 500,000 | 4 |
| bcc.e2.c4m8 | 4 cores | 8 GiB | Up to 3 Gbps | Up to 500,000 | 4 |
| bcc.e2.c4m16 | 4 cores | 16 GiB | Up to 3 Gbps | Up to 500,000 | 4 |
| bcc.e2.c8m8 | 8 cores | 8 GiB | Up to 6 Gbps | Up to 800,000 | 4 |
| bcc.e2.c8m16 | 8 cores | 16 GiB | Up to 6 Gbps | Up to 800,000 | 4 |
| bcc.e2.c8m32 | 8 cores | 32 GiB | Up to 6 Gbps | Up to 800,000 | 4 |
General-purpose instance family
General-purpose g7
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Support AMX (Advanced Matrix Extensions) to enhance CPU performance for deep learning training and inference
- Supports up to 160 Gbps internal bandwidth, with small-specification instances offering burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- Online inference, machine learning
- High network packet processing scenarios
Support resize only to g7, c7, and m7 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Specification
| Instance specification. | vCPU | Memory | Intranet bandwidth base/burst | Network packet sent/received PPS | Connections | Count of elastic network interfaces | Queue count | Count of private IPv4 addresses per network interface card | Count of private IPv6 addresses per network interface card | Count of cloud disk server (excluding system disk) | Cloud disk server storage base IOPS | Cloud disk server storage base bandwidth |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.g7.c2m8 | 2 cores | 8 GiB | 2.5/Maximum 15 Gbps | 1,000,000 | Up to 300,000 | 1 | 2 | 2 | 2 | 1 | 10,000 | 1 Gbps |
| bcc.g7.c4m16 | 4 cores | 16 GiB | 4/Maximum 15 Gbps | 1,200,000 | Up to 300,000 | 3 | 4 | 8 | 8 | 1 | 15,000 | 1.5 Gbps |
| bcc.g7.c8m32 | 8 cores | 32 GiB | 6/Maximum 15 Gbps | 1,600,000 | Up to 300,000 | 3 | 8 | 16 | 16 | 4 | 20,000 | 2 Gbps |
| bcc.g7.c16m64 | 16 cores | 64 GiB | 12/Maximum 25 Gbps | 3,000,000 | 350,000 | 7 | 16 | 16 | 16 | 14 | 25,000 | 3 Gbps |
| bcc.g7.c32m128 | 32 cores | 128 GiB | 20/Maximum 25 Gbps | 5,000,000 | 600,000 | 7 | Maximum 32 | 16 | 16 | 16 | 30,000 | 3.5 Gbps |
| bcc.g7.c48m192 | 48 cores | 192 GiB | 30 Gbps | 8,000,000 | 750,000 | 7 | Maximum 32 | 30 | 30 | 16 | 40,000 | 4 Gbps |
| bcc.g7.c64m256 | 64 cores | 256 GiB | 50 Gbps | 11,000,000 | 1,000,000 | 7 | Maximum 32 | 30 | 30 | 16 | 50,000 | 5 Gbps |
| bcc.g7.c96m384 | 96 cores | 384 GiB | 70 Gbps | 16,000,000 | 1,500,000 | 15 | Maximum 32 | 40 | 40 | 16 | 140,000 | 14.5 Gbps |
| bcc.g7.c104m512 | 104 cores | 384 GiB | 80 Gbps | 18,000,000 | 1,500,000 | 15 | Maximum 32 | 40 | 40 | 16 | 150,000 | 16 Gbps |
General-purpose ga3 (Public Beta)
Instance Family Characteristics
- Processor: AMD EPYC 9W24, base frequency 2.6GHz, turbo frequency 3.7GHz
- Processor-to-memory ratio is 1:4
- Supports up to 80 Gbps internal bandwidth, with small-specification instances offering burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- High network packet processing scenarios (e.g., gaming, live streaming, etc.)
Support resize only to ga3, ca3, and ma3 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Product Specifications
| Instance specification. | vCPU | Memory (GiB) | Network bandwidth base/burst (Gbit/s) | Network packet sent/received PPS | Connections | Queue count | Count of elastic network interfaces | Count of private network IPv4 addresses per network interface card | Count of private network IPv6 addresses per network interface card | Total number of disks (excluding system disk) | Cloud disk server base IOPS (10,000) | Cloud disk server base bandwidth (Gbit/s) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.ga3.c2m8 | 2 | 8 | 1.5/up to 12.5 | 800,000 | Maximum 250,000 | 2 | 1 | 2 | 2 | 1 | 1 | 1 |
| bcc.ga3.c4m16 | 4 | 16 | 2.5/Maximum 12.5 | 1,000,000 | Maximum 250,000 | 4 | 3 | 8 | 8 | 1 | 1.5 | 1.5 |
| bcc.ga3.c8m32 | 8 | 32 | 4/Maximum 12.5 | 1,600,000 | Maximum 250,000 | 8 | 3 | 16 | 16 | 4 | 2 | 2 |
| bcc.ga3.c16m64 | 16 | 64 | 10/Maximum 12.5 | 2,000,000 | 300,000 | 16 | 3 | 16 | 16 | 14 | 2.5 | 3 |
| bcc.ga3.c32m128 | 32 | 128 | 15/Maximum 25 | 3,000,000 | 600,000 | Maximum 32 | 3 | 16 | 16 | 16 | 3 | 3.5 |
| bcc.ga3.c48m192 | 48 | 192 | 20/Maximum 25 | 4,500,000 | 750,000 | Maximum 32 | 7 | 30 | 30 | 16 | 4 | 4 |
| bcc.ga3.c64m256 | 64 | 256 | 25 | 6,000,000 | 1,000,000 | Maximum 32 | 7 | 30 | 30 | 16 | 5 | 5 |
| bcc.ga3.c96m384 | 96 | 384 | 40/none | 9,000,000 | 1,500,000 | Maximum 32 | 14 | 40 | 40 | 16 | 8 | 8 |
| bcc.ga3.c128m512 | 128 | 512 | 50/none | 12,000,000 | 2,000,000 | Maximum 32 | 14 | 40 | 40 | 16 | 10 | 10 |
| bcc.ga3.c192m768 | 192 | 768 | 80/none | 18,000,000 | 3,000,000 | Maximum 32 | 14 | 40 | 40 | 16 | 15 | 16 |
RDMA network-enhanced g5re
Instance Family Characteristics
This instance is currently whitelisted; submit a Ticket if needed
RDMA network-enhanced instances can mount elastic RDMA interface (ERI) to communicate via RoCE v2 protocol in Virtual Private Cloud (VPC), featuring low latency, high bandwidth, and cost efficiency, facilitating large-scale instance deployment on cloud. For detailed ERI usage, refer to Elastic RDMA Interface.
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Support up to 200 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
- Support mounting elastic RDMA interface
-
Applicable scenarios
- Database systems, caching, search clusters
- High-performance computing applications (e.g., fluid dynamics, simulation computer)
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Mountable ERI count |
|---|---|---|---|---|---|---|
| bcc.g5re.c112m476 | 112 cores | 476 GiB | 200 Gbps | 50,000,000 | 32 | 1 |
General-purpose g6
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8463B, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Support up to 90 Gbps internal bandwidth
- Performance improved by 70% compared to the previous generation of Baidu Cloud Compute
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- Online inference, machine learning
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Connections |
|---|---|---|---|---|---|---|
| bcc.g6.c2m8 | 2 cores | 8 GiB | 2 Gbps | 800,000 | 2 | Up to 300,000 |
| bcc.g6.c4m16 | 4 cores | 16 GiB | 4 Gbps | 1,000,000 | 4 | Up to 300,000 |
| bcc.g6.c8m32 | 8 cores | 32 GiB | 6 Gbps | 1,400,000 | 8 | Up to 300,000 |
| bcc.g6.c12m48 | 12 cores | 48 GiB | 10 Gbps | 1,800,000 | 12 | 400,000 |
| bcc.g6.c16m64 | 16 cores | 64 GiB | 12 Gbps | 2,000,000 | 16 | 500,000 |
| bcc.g6.c24m96 | 24 cores | 96 GiB | 15 Gbps | 2,500,000 | 24 | 800,000 |
| bcc.g6.c32m128 | 32 cores | 128 GiB | 20 Gbps | 3,000,000 | 32 | 1,000,000 |
| bcc.g6.c48m192 | 48 cores | 192 GiB | 25 Gbps | 4,500,000 | 48 | 1,500,000 |
| bcc.g6.c64m256 | 64 cores | 256 GiB | 32 Gbps | 8,000,000 | 64 | 2,000,000 |
| bcc.g6.c96m384 | 96 cores | 384 GiB | 50 Gbps | 12,000,000 | 64 | 3,000,000 |
| bcc.g6.c128m512 | 128 cores | 512 GiB | 70 Gbps | 1,300,000 | 64 | 3,500,000 |
| bcc.g6.c172m966 | 172 cores | 966 GiB | 90 Gbps | 1,500,000 | 64 | 4,000,000 |
General-purpose g5
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- Compute clusters and memory-dependent data processing
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.g5.c2m8 | 2 cores | 8 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.g5.c4m16 | 4 cores | 16 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.g5.c8m32 | 8 cores | 32 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.g5.c16m64 | 16 cores | 64 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.g5.c24m96 | 24 cores | 96 GiB | 8 Gbps | 4,000,000 | 16 |
| bcc.g5.c32m128 | 32 cores | 128 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.g5.c48m192 | 48 cores | 192 GiB | 16 Gbps | 6,000,000 | 32 |
| bcc.g5.c56m238 | 56 cores | 238 GiB | 22 Gbps | 7,000,000 | 32 |
| bcc.g5.c64m256 | 64 cores | 256 GiB | 22 Gbps | 7,000,000 | 32 |
| bcc.g5.c112m476 | 112 cores | 476 GiB | 35 Gbps | 12,000,000 | 32 |
General-purpose gr1
Instance Family Characteristics
- Processor: Ampere Altra Q80-30, base frequency 3.0GHz
- Processor-to-memory ratio is 1:4
- Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Cloud terminal emulation, Baidu ARM Cloud, cloud gaming
- Web application services
- Containers, microservices
- AI inference (data governance), data analysis, and computing
- DevOps
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.gr1.c1m4 | 1 core | 4 GiB | 1 Gbps | 750,000 | 1 |
| bcc.gr1.c2m8 | 2 cores | 8 GiB | 1 Gbps | 900,000 | 2 |
| bcc.gr1.c4m16 | 4 cores | 16 GiB | 2 Gbps | 1,000,000 | 4 |
| bcc.gr1.c8m32 | 8 cores | 32 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.gr1.c16m64 | 16 cores | 64 GiB | 6 Gbps | 1,800,000 | 8 |
| bcc.gr1.c24m96 | 24 cores | 96 GiB | 8 Gbps | 2,250,000 | 12 |
| bcc.gr1.c32m128 | 32 cores | 128 GiB | 10 Gbps | 2,750,000 | 16 |
| bcc.gr1.c48m192 | 48 cores | 192 GiB | 16 Gbps | 3,750,000 | 16 |
| bcc.gr1.c64m238 | 64 cores | 238 GiB | 25 Gbps | 5,000,000 | 32 |
General-purpose g4
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:4
- Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Data analysis and computing
- Enterprise applications of all types and scales
- Small/medium database systems, caching, search clusters
- Compute clusters and memory-dependent data processing
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.g4.c2m8 | 2 cores | 8 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.g4.c4m16 | 4 cores | 16 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.g4.c8m32 | 8 cores | 32 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.g4.c16m64 | 16 cores | 64 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.g4.c24m96 | 24 cores | 96 GiB | 8 Gbps | 4,000,000 | 12 |
| bcc.g4.c32m128 | 32 cores | 128 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.g4.c48m192 | 48 cores | 192 GiB | 15 Gbps | 6,000,000 | 16 |
| bcc.g4.c64m256 | 64 cores | 256 GiB | 22 Gbps | 7,000,000 | 16 |
| bcc.g4.c84m352 | 84 cores | 352 GiB | 25 Gbps | 8,000,000 | 16 |
General-purpose ga2
Instance Family Characteristics
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:4
- Support Enhanced Cloud Disk Server
- Maximum network packet sent/received capacity: 1,200w
-
Applicable scenarios
- Container
- Offline computing frameworks
- High-performance HPC computing
- Web frontend services, distributed caching
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.ga2.c2m8 | 2 cores | 8 GiB | 1.5 Gbps | 100,000 | 2 |
| bcc.ga2.c4m16 | 4 cores | 16 GiB | 3 Gbps | 500,000 | 4 |
| bcc.ga2.c8m32 | 8 cores | 32 GiB | 6 Gbps | 800,000 | 8 |
| bcc.ga2.c16m64 | 16 cores | 64 GiB | 12 Gbps | 1,500,000 | 8 |
| bcc.ga2.c32m128 | 32 cores | 128 GiB | 15 Gbps | 2,000,000 | 16 |
| bcc.ga2.c48m192 | 48 cores | 192 GiB | 20 Gbps | 3,000,000 | 32 |
| bcc.ga2.c64m256 | 64 cores | 256 GiB | 30 Gbps | 4,000,000 | 32 |
| bcc.ga2.c80m320 | 80 cores | 320 GiB | 35 Gbps | 5,000,000 | 32 |
| bcc.ga2.c116m488 | 116 cores | 488 GiB | 45 Gbps | 6,000,000 | 32 |
| bcc.ga2.c232m978 | 232 cores | 978 GiB | 90 Gbps | 12,000,000 | 32 |
General-purpose ga1
Instance Family Characteristics
- Processor: AMD EPYC 7502, base frequency 2.5GHz, turbo frequency 3.35GHz
- Processor-to-memory ratio is 1:4
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Data analysis and computing
- Enterprise applications of all types and scales
- Small/medium database systems, caching, search clusters
- Compute clusters and memory-dependent data processing
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS |
|---|---|---|---|---|
| bcc.ga1.c2m8 | 2 cores | 8 GiB | 1.5 Gbps | 150,000 |
| bcc.ga1.c4m16 | 4 cores | 16 GiB | 1.5 Gbps | 300,000 |
| bcc.ga1.c8m32 | 8 cores | 32 GiB | 3 Gbps | 600,000 |
| bcc.ga1.c16m64 | 16 cores | 64 GiB | 6 Gbps | 1,200,000 |
| bcc.ga1.c24m96 | 24 cores | 96 GiB | 8 Gbps | 1,800,000 |
| bcc.ga1.c32m128 | 32 cores | 128 GiB | 12 Gbps | 2,500,000 |
| bcc.ga1.c48m192 | 48 cores | 192 GiB | 14 Gbps | 3,000,000 |
| bcc.ga1.c64m256 | 64 cores | 256 GiB | 18 Gbps | 4,500,000 |
Compute-optimized instance family
Compute-optimized c7
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:2
- Support AMX (Advanced Matrix Extensions) to enhance CPU performance for deep learning training and inference
- Supports up to 160 Gbps internal bandwidth, with small-specification instances offering burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- Online inference, machine learning
- High network packet processing scenarios
Support resize only to g7, c7, and m7 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Specification
| Instance specification. | vCPU | Memory | Intranet bandwidth base/burst | Network packet sent/received PPS | Connections | Count of elastic network interfaces | Queue count | Count of private IPv4 addresses per network interface card | Count of private IPv6 addresses per network interface card | Count of cloud disk server (excluding system disk) | Cloud disk server storage base IOPS | Cloud disk server storage base bandwidth |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.c7.c2m4 | 2 cores | 4 GiB | 2.5/Maximum 15 Gbps | 1,000,000 | Up to 300,000 | 1 | 2 | 2 | 2 | 1 | 10,000 | 1 Gbps |
| bcc.c7.c4m8 | 4 cores | 8 GiB | 4/Maximum 15 Gbps | 1,200,000 | Up to 300,000 | 3 | 4 | 8 | 8 | 1 | 15,000 | 1.5 Gbps |
| bcc.c7.c8m16 | 8 cores | 16 GiB | 6/Maximum 15 Gbps | 1,600,000 | Up to 300,000 | 3 | 8 | 16 | 16 | 4 | 20,000 | 2 Gbps |
| bcc.c7.c16m32 | 16 cores | 32GiB | 12/Maximum 25 Gbps | 3,000,000 | 350,000 | 7 | 16 | 16 | 16 | 14 | 25,000 | 3 Gbps |
| bcc.c7.c32m64 | 32 cores | 64 GiB | 20/Maximum 25 Gbps | 5,000,000 | 600,000 | 7 | Maximum 32 | 16 | 16 | 16 | 30,000 | 3.5 Gbps |
| bcc.c7.c48m96 | 48 cores | 96 GiB | 30 Gbps | 8,000,000 | 750,000 | 7 | Maximum 32 | 30 | 30 | 16 | 40,000 | 4 Gbps |
| bcc.c7.c64m128 | 64 cores | 128 GiB | 50 Gbps | 11,000,000 | 1,000,000 | 7 | Maximum 32 | 30 | 30 | 16 | 50,000 | 5 Gbps |
| bcc.c7.c96m192 | 96 cores | 192 GiB | 70 Gbps | 16,000,000 | 1,500,000 | 15 | Maximum 32 | 40 | 40 | 16 | 140,000 | 14.5 Gbps |
Compute-optimized ca3
Instance Family Characteristics
- Processor: AMD EPYC 9W24, base frequency 2.6GHz, turbo frequency 3.7GHz
- Processor-to-memory ratio is 1:2
- Supports up to 80 Gbps internal bandwidth, with small-specification instances offering burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- High network packet processing scenarios (e.g., gaming, live streaming, etc.)
Support resize only to ga3, ca3, and ma3 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Product Specifications
| Instance specification. | vCPU | Memory (GiB) | Network bandwidth base/burst (Gbit/s) | Network packet sent/received PPS | Connections | Queue count | Count of elastic network interfaces | Count of private network IPv4 addresses per network interface card | Count of private network IPv6 addresses per network interface card | Total number of disks (excluding system disk) | Cloud disk server base IOPS (10,000) | Cloud disk server base bandwidth (Gbit/s) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.ca3.c2m4 | 2 | 4 | 1.5/up to 12.5 | 800,000 | Maximum 250,000 | 2 | 1 | 2 | 2 | 1 | 1 | 1 |
| bcc.ca3.c4m8 | 4 | 8 | 2.5/Maximum 12.5 | 1,000,000 | Maximum 250,000 | 4 | 3 | 8 | 8 | 1 | 1.5 | 1.5 |
| bcc.ca3.c8m16 | 8 | 16 | 4/Maximum 12.5 | 1,600,000 | Maximum 250,000 | 8 | 3 | 16 | 16 | 4 | 2 | 2 |
| bcc.ca3.c16m32 | 16 | 32 | 10/Maximum 12.5 | 2,000,000 | 300,000 | 16 | 3 | 16 | 16 | 14 | 2.5 | 3 |
| bcc.ca3.c32m64 | 32 | 64 | 15/Maximum 25 | 3,000,000 | 600,000 | Maximum 32 | 3 | 16 | 16 | 16 | 3 | 3.5 |
| bcc.ca3.c48m96 | 48 | 96 | 20/Maximum 25 | 4,500,000 | 750,000 | Maximum 32 | 7 | 30 | 30 | 16 | 4 | 4 |
| bcc.ca3.c64m128 | 64 | 128 | 25 | 6,000,000 | 1,000,000 | Maximum 32 | 7 | 30 | 30 | 16 | 5 | 5 |
| bcc.ca3.c96m192 | 96 | 192 | 40/none | 9,000,000 | 1,500,000 | Maximum 32 | 14 | 40 | 40 | 16 | 8 | 8 |
| bcc.ca3.c128m256 | 128 | 256 | 50/none | 12,000,000 | 2,000,000 | Maximum 32 | 14 | 40 | 40 | 16 | 10 | 10 |
Compute-optimized c6
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8463B, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:2
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
- High-performance scientific and engineering applications
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Connections |
|---|---|---|---|---|---|---|
| bcc.c6.c2m4 | 2 cores | 4 GiB | 2 Gbps | 800,000 | 2 | Up to 300,000 |
| bcc.c6.c4m8 | 4 cores | 8 GiB | 4 Gbps | 1,000,000 | 4 | Up to 300,000 |
| bcc.c6.c8m16 | 8 cores | 16 GiB | 6 Gbps | 1,400,000 | 8 | Up to 300,000 |
| bcc.c6.c12m24 | 12 cores | 24 GiB | 10 Gbps | 1,800,000 | 12 | 400,000 |
| bcc.c6.c16m32 | 16 cores | 32 GiB | 12 Gbps | 2,000,000 | 16 | 500,000 |
| bcc.c6.c24m48 | 24 cores | 48 GiB | 15 Gbps | 2,500,000 | 24 | 800,000 |
| bcc.c6.c32m64 | 32 cores | 64 GiB | 20 Gbps | 3,000,000 | 32 | 1,000,000 |
| bcc.c6.c48m96 | 48 cores | 96 GiB | 25 Gbps | 4,500,000 | 48 | 1,500,000 |
| bcc.c6.c64m128 | 64 cores | 128 GiB | 32 Gbps | 8,000,000 | 64 | 2,000,000 |
| bcc.c6.c96m128 | 96 cores | 192 GiB | 50 Gbps | 12,000,000 | 64 | 3,000,000 |
| bcc.c6.c128m256 | 128 cores | 256 GiB | 70 Gbps | 1,300,000 | 64 | 3,500,000 |
Compute-optimized c5
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:2
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
- High-performance scientific and engineering applications
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.c5.c2m4 | 2 cores | 4 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.c5.c4m8 | 4 cores | 8 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.c5.c8m16 | 8 cores | 16 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.c5.c16m32 | 16 cores | 32 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.c5.c24m48 | 24 cores | 48 GiB | 8 Gbps | 4,000,000 | 16 |
| bcc.c5.c32m64 | 32 cores | 64 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.c5.c48m96 | 48 cores | 96 GiB | 16 Gbps | 6,000,000 | 32 |
| bcc.c5.c64m128 | 64 cores | 128 GiB | 22 Gbps | 7,000,000 | 32 |
Compute intensive ic5
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:1
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.ic5.c2m2 | 2 cores | 2 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.ic5.c4m4 | 4 cores | 4 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.ic5.c8m8 | 8 cores | 8 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.ic5.c16m16 | 16 cores | 16 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.ic5.c24m24 | 24 cores | 24 GiB | 8 Gbps | 4,000,000 | 16 |
| bcc.ic5.c32m32 | 32 cores | 32 GiB | 11 Gbps | 5,000,000 | 16 |
Compute-optimized cr1
Instance Family Characteristics
- Processor: Ampere Altra Q80-30, base frequency 3.0GHz
- Processor-to-memory ratio is 1:2
- Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Cloud terminal emulation, Baidu ARM Cloud, cloud gaming
- Web application services
- Containers, microservices
- DevOps
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.cr1.c1m2 | 1 core | 2 GiB | 1 Gbps | 750,000 | 1 |
| bcc.cr1.c2m4 | 2 cores | 4 GiB | 1 Gbps | 900,000 | 2 |
| bcc.cr1.c4m8 | 4 cores | 8 GiB | 2 Gbps | 1,000,000 | 4 |
| bcc.cr1.c8m16 | 8 cores | 16 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.cr1.c16m32 | 16 cores | 32 GiB | 6 Gbps | 1,800,000 | 8 |
| bcc.cr1.c24m48 | 24 cores | 48 GiB | 8 Gbps | 2,250,000 | 12 |
| bcc.cr1.c32m64 | 32 cores | 64 GiB | 10 Gbps | 2,750,000 | 16 |
Compute-optimized c4
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:2
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
- High-performance scientific and engineering applications
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.c4.c2m4 | 2 cores | 4 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.c4.c4m8 | 4 cores | 8 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.c4.c8m16 | 8 cores | 16 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.c4.c16m32 | 16 cores | 32 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.c4.c24m48 | 24 cores | 48 GiB | 8 Gbps | 4,000,000 | 12 |
| bcc.c4.c32m64 | 32 cores | 64 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.c4.c48m96 | 48 cores | 96 GiB | 15 Gbps | 6,000,000 | 16 |
| bcc.c4.c64m128 | 64 cores | 128 GiB | 22 Gbps | 7,000,000 | 16 |
Intensive computing ic4
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:1
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.ic4.c2m2 | 2 cores | 2 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.ic4.c4m4 | 4 cores | 4 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.ic4.c8m8 | 8 cores | 8 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.ic4.c16m16 | 16 cores | 16 GiB | 6 Gbps | 2,800,000 | 8 |
Compute-optimized ca2
Instance Family Characteristics
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:2
- Support Enhanced Cloud Disk Server
- Ultra-high network PPS packet sent/received capability
-
Applicable scenarios
- Container
- Offline computing frameworks
- High-performance HPC computing
- Web frontend services, distributed caching
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.ca2.c2m4 | 2 cores | 4 GiB | 1.5 Gbps | 100,000 | 2 |
| bcc.ca2.c4m8 | 4 cores | 8 GiB | 3 Gbps | 500,000 | 4 |
| bcc.ca2.c8m16 | 8 cores | 16 GiB | 6 Gbps | 800,000 | 8 |
| bcc.ca2.c16m32 | 16 cores | 32 GiB | 12 Gbps | 1,500,000 | 8 |
| bcc.ca2.c32m64 | 32 cores | 64 GiB | 15 Gbps | 2,000,000 | 16 |
| bcc.ca2.c64m128 | 64 cores | 128 GiB | 30 Gbps | 4,000,000 | 32 |
| bcc.ca2.c80m160 | 80 cores | 160 GiB | 35 Gbps | 5,000,000 | 32 |
Compute-optimized ca1
Instance Family Characteristics
- Processor: AMD EPYC 7502, base frequency 2.5GHz, turbo frequency 3.35GHz
- Processor-to-memory ratio is 1:2
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
- High-performance scientific and engineering applications
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS |
|---|---|---|---|---|
| bcc.ca1.c2m4 | 2 cores | 4 GiB | 1.5 Gbps | 150,000 |
| bcc.ca1.c4m8 | 4 cores | 8 GiB | 1.5 Gbps | 300,000 |
| bcc.ca1.c8m16 | 8 cores | 16 GiB | 3 Gbps | 600,000 |
| bcc.ca1.c16m32 | 16 cores | 32 GiB | 6 Gbps | 1,200,000 |
| bcc.ca1.c24m48 | 24 cores | 48 GiB | 8 Gbps | 1,800,000 |
| bcc.ca1.c32m64 | 32 cores | 64 GiB | 12 Gbps | 2,500,000 |
| bcc.ca1.c48m96 | 48 cores | 96 GiB | 14 Gbps | 3,000,000 |
| bcc.ca1.c64m128 | 64 cores | 128 GiB | 18 Gbps | 4,500,000 |
Compute intensive ica2
Instance Family Characteristics
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:1
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS |
|---|---|---|---|---|
| bcc.ica2.c2m2 | 2 cores | 2 GiB | 1.5 Gbps | 100,000 |
| bcc.ica2.c4m4 | 4 cores | 4 GiB | 3 Gbps | 500,000 |
| bcc.ica2.c8m8 | 8 cores | 8 GiB | 6 Gbps | 800,000 |
| bcc.ica2.c16m16 | 16 cores | 16 GiB | 12 Gbps | 1,500,000 |
| bcc.ica2.c32m32 | 32 cores | 32 GiB | 15 Gbps | 2,000,000 |
| bcc.ica2.c48m48 | 48 cores | 48 GiB | 20 Gbps | 3,000,000 |
| bcc.ica2.c64m64 | 64 cores | 64 GiB | 30 Gbps | 4,000,000 |
Compute intensive ica1
Instance Family Characteristics
- Processor: AMD EPYC 7502, base frequency 2.5GHz, turbo frequency 3.35GHz
- Processor-to-memory ratio is 1:1
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Web frontend servers
- Large-scale MMO game frontends
- Data analysis, batch computing, video encoding
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth |
|---|---|---|---|
| bcc.ica1.c2m2 | 2 cores | 2 GiB | 1.5 Gbps |
| bcc.ica1.c4m4 | 4 cores | 4 GiB | 1.5 Gbps |
| bcc.ica1.c8m8 | 8 cores | 8 GiB | 2 Gbps |
| bcc.ica1.c16m16 | 16 cores | 16 GiB | 3 Gbps |
Memory-optimized instance family
Memory-optimized m7
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:8
- Support AMX (Advanced Matrix Extensions) to enhance CPU performance for deep learning training and inference
- Supports up to 160 Gbps internal bandwidth, with small-specification instances offering burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- Online inference, machine learning
- High network packet processing scenarios
Support resize only to g7, c7, and m7 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Specification
| Instance specification. | vCPU | Memory | Intranet bandwidth base/burst | Network packet sent/received PPS | Connections | Count of elastic network interfaces | Queue count | Count of private IPv4 addresses per network interface card | Count of private IPv6 addresses per network interface card | Count of cloud disk server (excluding system disk) | Cloud disk server storage base IOPS | Cloud disk server storage base bandwidth |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.m7.c2m16 | 2 cores | 16 GiB | 2.5/Maximum 15 Gbps | 1,000,000 | Up to 300,000 | 1 | 2 | 2 | 2 | 1 | 10,000 | 1 Gbps |
| bcc.m7.c4m32 | 4 cores | 32GiB | 4/Maximum 15 Gbps | 1,200,000 | Up to 300,000 | 3 | 4 | 8 | 8 | 1 | 15,000 | 1.5 Gbps |
| bcc.m7.c8m64 | 8 cores | 64 GiB | 6/Maximum 15 Gbps | 1,600,000 | Up to 300,000 | 3 | 8 | 16 | 16 | 4 | 20,000 | 2 Gbps |
| bcc.m7.c16m128 | 16 cores | 128GiB | 12/Maximum 25 Gbps | 3,000,000 | 350,000 | 7 | 16 | 16 | 16 | 14 | 25,000 | 3 Gbps |
| bcc.m7.c32m256 | 32 cores | 256 GiB | 20/Maximum 25 Gbps | 5,000,000 | 600,000 | 7 | Maximum 32 | 16 | 16 | 16 | 30,000 | 3.5 Gbps |
| bcc.m7.c48m384 | 48 cores | 384GiB | 30 Gbps | 8,000,000 | 750,000 | 7 | Maximum 32 | 30 | 30 | 16 | 40,000 | 4 Gbps |
Memory-optimized ma3
Instance Family Characteristics
- Processor: AMD EPYC 9W24, base frequency 2.6GHz, turbo frequency 3.7GHz
- Processor-to-memory ratio is 1:8
- Small-specification instances offer burstable network performance. Network performance aligns with compute specifications—higher specifications provide better performance.
-
Applicable scenarios
- Enterprise applications of all types and scales
- Data analysis and computing
- Small/medium database systems, caching, search clusters
- High network packet processing scenarios (e.g., gaming, live streaming, etc.)
Support resize only to ga3, ca3, and ma3 specifications
Windows image is currently unsupported
Do not support Linux images of low version, including CentOS6 or earlier, Ubuntu14 or earlier, Debian8 or earlier, etc.
Product Specifications
| Instance specification. | vCPU | Memory (GiB) | Network bandwidth base/burst (Gbit/s) | Network packet sent/received PPS | Connections | Queue count | Count of elastic network interfaces | Count of private network IPv4 addresses per network interface card | Count of private network IPv6 addresses per network interface card | Total number of disks (excluding system disk) | Cloud disk server base IOPS (10,000) | Cloud disk server base bandwidth (Gbit/s) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bcc.ma3.c2m16 | 2 | 16 | 1.5/up to 12.5 | 800,000 | Maximum 250,000 | 2 | 1 | 2 | 2 | 1 | 1 | 1 |
| bcc.ma3.c4m32 | 4 | 32 | 2.5/Maximum 12.5 | 1,000,000 | Maximum 250,000 | 4 | 3 | 8 | 8 | 1 | 1.5 | 1.5 |
| bcc.ma3.c8m64 | 8 | 64 | 4/Maximum 12.5 | 1,600,000 | Maximum 250,000 | 8 | 3 | 16 | 16 | 4 | 2 | 2 |
| bcc.ma3.c16m128 | 16 | 128 | 10/Maximum 12.5 | 2,000,000 | 300,000 | 16 | 3 | 16 | 16 | 14 | 2.5 | 3 |
| bcc.ma3.c32m256 | 32 | 256 | 15/Maximum 25 | 3,000,000 | 600,000 | Maximum 32 | 3 | 16 | 16 | 16 | 3 | 3.5 |
| bcc.ma3.c48m384 | 48 | 384 | 20/Maximum 25 | 4,500,000 | 750,000 | Maximum 32 | 7 | 30 | 30 | 16 | 4 | 4 |
| bcc.ma3.c64m512 | 64 | 512 | 25 | 6,000,000 | 1,000,000 | Maximum 32 | 7 | 30 | 30 | 16 | 5 | 5 |
Memory-optimized m6
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8463B, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:8
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- High-performance databases, memory databases
- Data analysis & mining, distributed memory caching
- Hadoop/Spark clusters and other enterprise applications with large memory requirements
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Connections |
|---|---|---|---|---|---|---|
| bcc.m6.c2m16 | 2 cores | 16 GiB | 2 Gbps | 800,000 | 2 | Up to 300,000 |
| bcc.m6.c4m32 | 4 cores | 32 GiB | 4 Gbps | 1,000,000 | 4 | Up to 300,000 |
| bcc.m6.c8m64 | 8 cores | 64 GiB | 6 Gbps | 1,400,000 | 8 | Up to 300,000 |
| bcc.m6.c12m96 | 12 cores | 96 GiB | 10 Gbps | 1,800,000 | 12 | 400,000 |
| bcc.m6.c16m128 | 16 cores | 128 GiB | 12 Gbps | 2,000,000 | 16 | 500,000 |
| bcc.m6.c24m192 | 24 cores | 192 GiB | 15 Gbps | 2,500,000 | 24 | 800,000 |
| bcc.m6.c32m256 | 32 cores | 256 GiB | 20 Gbps | 3,000,000 | 32 | 1,000,000 |
| bcc.m6.c64m512 | 64 cores | 512 GiB | 32 Gbps | 8,000,000 | 64 | 2,000,000 |
Memory-optimized m5
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:8
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- High-performance databases, memory databases
- Data analysis & mining, distributed memory caching
- Hadoop/Spark clusters and other enterprise applications with large memory requirements
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.m5.c2m16 | 2 cores | 16 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.m5.c4m32 | 4 cores | 32 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.m5.c8m64 | 8 cores | 64 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.m5.c16m128 | 16 cores | 128 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.m5.c24m192 | 24 cores | 192 GiB | 8 Gbps | 4,000,000 | 16 |
| bcc.m5.c32m256 | 32 cores | 256 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.m5.c48m384 | 48 cores | 384 GiB | 16 Gbps | 6,000,000 | 32 |
| bcc.m5.c64m512 | 64 cores | 512 GiB | 22 Gbps | 7,000,000 | 32 |
Memory-optimized m4
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:8
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- High-performance databases, memory databases
- Data analysis & mining, distributed memory caching
- Hadoop/Spark clusters and other enterprise applications with large memory requirements
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.m4.c2m16 | 2 cores | 16 GiB | 1.5 Gbps | 400,000 | 2 |
| bcc.m4.c4m32 | 4 cores | 32 GiB | 1.5 Gbps | 800,000 | 4 |
| bcc.m4.c8m64 | 8 cores | 64 GiB | 3 Gbps | 1,500,000 | 8 |
| bcc.m4.c16m128 | 16 cores | 128 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.m4.c24m192 | 24 cores | 192 GiB | 8 Gbps | 4,000,000 | 12 |
| bcc.m4.c32m256 | 32 cores | 256 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.m4.c48m384 | 48 cores | 384 GiB | 15 Gbps | 6,000,000 | 16 |
| bcc.m4.c64m512 | 64 cores | 512 GiB | 22 Gbps | 7,000,000 | 16 |
Memory-optimized ma2
Instance Family Characteristics
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:8
- Support Enhanced Cloud Disk Server
- Ultra-high network PPS packet sent/received capability
-
Applicable scenarios
- Container
- Offline computing frameworks
- High-performance HPC computing
- Web frontend services, distributed caching
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| bcc.ma2.c2m16 | 2 cores | 16 GiB | 1.5 Gbps | 100,000 | 2 |
| bcc.ma2.c4m32 | 4 cores | 32 GiB | 3 Gbps | 500,000 | 4 |
| bcc.ma2.c8m64 | 8 cores | 64 GiB | 6 Gbps | 800,000 | 8 |
| bcc.ma2.c16m128 | 16 cores | 128 GiB | 12 Gbps | 1,500,000 | 8 |
| bcc.ma2.c32m256 | 32 cores | 256 GiB | 15 Gbps | 2,000,000 | 16 |
Memory-optimized ma1
Instance Family Characteristics
- Processor: AMD EPYC 7502, base frequency 2.5GHz, turbo frequency 3.35GHz
- Processor-to-memory ratio is 1:8
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- High-performance databases, memory databases
- Data analysis & mining, distributed memory caching
- Hadoop/Spark clusters and other enterprise applications with large memory requirements
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS |
|---|---|---|---|---|
| bcc.ma1.c2m16 | 2 cores | 16 GiB | 1.5 Gbps | 150,000 |
| bcc.ma1.c4m32 | 4 cores | 32 GiB | 1.5 Gbps | 300,000 |
| bcc.ma1.c8m64 | 8 cores | 64 GiB | 3 Gbps | 600,000 |
| bcc.ma1.c16m128 | 16 cores | 128 GiB | 6 Gbps | 1,200,000 |
| bcc.ma1.c24m192 | 24 cores | 192 GiB | 8 Gbps | 1,800,000 |
| bcc.ma1.c32m256 | 32 cores | 256 GiB | 12 Gbps | 2,500,000 |
High-frequency instance family
high-frequency memory-optimized hcm7
Instance Family Characteristics
- Processor: IIntel Xeon Platinum series, with a base frequency of 3.2GHz and a turbo frequency of 3.3GHz
- Processor-to-memory ratio is 1:8
- Supports up to 160Gbps of internal network bandwidth. Smaller instance types offer burstable network performance. Network performance is directly related to the compute specifications; the higher the spec, the stronger the performance.
-
Applicable scenarios - Data analysis and video encoding. - High-performance and in-memory databases. - High network throughput scenarios, such as video "bullet comments" and live streaming. - Hadoop, Spark clusters, and other enterprise applications with large memory requirements. .
Older Linux images are not supported, including CentOS 6 and earlier, Ubuntu 14 and earlier, and Debian 8 and earlier.
Specification | Instance specification. | vCPU | Memory | Intranet bandwidth base/burst | Network packet sent/received PPS | Connections | Count of elastic network interfaces | Queue count | Count of private IPv4 addresses per network interface card | Count of private IPv6 addresses per network interface card | Count of cloud disk server (excluding system disk) | Cloud disk server storage base IOPS | Cloud disk server storage base bandwidth | |:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----| | bcc.hcm7.c2m16 | 2 cores | 16 GiB | 2.5/Maximum 15 Gbps | 1,000,000 | Up to 300,000 | 2 | 2 | 2 | 2 | 8 | 10,000 | 1 Gbps | | bcc.hcm7.c4m32 | 4 cores | 32GiB | 5/Maximum 15 Gbps | 1,400,000 | Up to 300,000 | 4 | 4 | 8 | 8 | 8 | 15,000 | 1.5 Gbps | | bcc.hcm7.c8m64 | 8 cores | 64 GiB | 10/Maximum 15 Gbps | 1,800,000 | Up to 300,000 | 8 | 8 | 16 | 16 | 16 | 20,000 | 2 Gbps | | bcc.hcm7.c16m128 | 16 cores | 128GiB | 20/Maximum 40 Gbps | 3,750,000 | 500,000 | 8 | 16 | 16 | 16 | 16 | 30,000 | 4 Gbps | | bcc.hcm7.c32m256 | 32 cores | 256 GiB | 40 Gbps | 7,500,000 | 750,000 | 8 | 32 | 16 | 16 | 16 | 70,000 | 8 Gbps | | bcc.hcm7.c64m512 | 64 cores | 512 GiB | 80 Gbps | 15,000,000 | 2,000,000 | 8 | 64 | 30 | 30 | 16 | 150,000 | 16 Gbps | | bcc.hcm7.c96m768 | 96 cores | 768 GiB | 120 Gbps | 18,000,000 | 3,000,000 | 15 | 64 | 40 | 40 | 16 | 225,000 | 24 Gbps | | bcc.hcm7.c128m1024 | 128 cores | 1024 GiB | 160 Gbps | 30,000,000 | 4,000,000 | 15 | 64 | 40 | 40 | 16 | 300,000 | 32 Gbps |
Big data instance family
Big data d1s
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:4
- Instances are equipped with high-capacity, high-throughput SATA HDD local disks
- Supports internal bandwidth of up to 25 Gbps. Instance network performance increases with specifications—higher specifications provide stronger forwarding capacity and greater intranet bandwidth.
-
Applicable scenarios
- Hadoop MapReduce, HDFS, Hive, HBase, etc.
- Spark memory computing, MLlib, etc.
- ElasticSearch, log, etc.
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.d1s.c8m32 | 8 cores | 32 GiB | 1 block * 11174 GiB | 6 Gbps | 1,500,000 | 8 |
| bcc.d1s.c16m64 | 16 cores | 64 GiB | 2 blocks * 11174 GiB | 12 Gbps | 2,800,000 | 8 |
| bcc.d1s.c24m96 | 24 cores | 96 GiB | 4 blocks * 11174 GiB | 16 Gbps | 4,000,000 | 12 |
| bcc.d1s.c32m128 | 32 cores | 128 GiB | 6 blocks * 11174 GiB | 20 Gbps | 5,000,000 | 16 |
| bcc.d1s.c64m256 | 64 cores | 256 GiB | 10 blocks * 11174 GiB | 25 Gbps | 7,000,000 | 16 |
| bcc.d1s.c80m320 | 80 cores | 320 GiB | 12 blocks * 11174 GiB | 25 Gbps | 8,000,000 | 16 |
Big data d1
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:4
- Instances feature high-throughput SATA HDD local disks, supporting up to 11 Gbps internal bandwidth
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
-
Applicable scenarios
- Hadoop MapReduce, HDFS, Hive, HBase, etc.
- Spark memory computing, MLlib, etc.
- Industry clients such as internet and financial sectors with big data computation, storage, and analytics requirements, operating in service scenarios involving massive-scale data storage and processing
- ElasticSearch, log, etc.
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS |
|---|---|---|---|---|---|
| bcc.d1.c8m32 | 8 cores | 32 GiB | 1 block * 7190 GiB | 6 Gbps | 1,500,000 |
| bcc.d1.c12m48 | 12 cores | 48 GiB | 2 blocks * 7190 GiB | 6 Gbps | 1,600,000 |
| bcc.d1.c16m64 | 16 cores | 64 GiB | 4 blocks * 7190 GiB | 11 Gbps | 2,500,000 |
Local SSD instance family
Local SSD type l7 (Public Beta)
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Support Enhanced Cloud Disk Server
-
Applicable scenarios
- OLTP, high-performance relational database
- NoSQL databases (e.g. Cassandra, MongoDB, etc.)
- Search scenarios such as Elasticsearch
Specification | Instance specification. | vCPU | Memory(GiB) | Local data disk | Intranet bandwidth base/burst(Gbit/s) | Network packet sent/received PPS | Connections | NIC multi-queue | Count of elastic network interfaces | Count of private IPv4 addresses per network interface card | Count of private IPv6 addresses per network interface card | Count of cloud disk server (excluding system disk) | Cloud disk server storage base IOPS | Cloud disk server storage base bandwidth(Gbit/s) | |:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----|:----| | bcc.l7.c26m128.1d | 26 | 128 |3570 GiB 1| 20/Maximum 25 | 5,000,000 | 600,000 | 26 | 8 | 16 | 16 | 16 | 30,000 | 3.5 | | bcc.l7.c52m256.2d | 52 | 256 |3570 GiB 2| 50/None | 10,000,000 | 1,000,000 | 52 | 8 | 30 | 30 | 16 | 50,000 | 5 | | bcc.l7.c104m512.4d | 104 | 512 |3570 GiB * 4| 80/None | 18,000,000 | 2,000,000 | 64| 15 | 40| 40 | 16 | 150,000 | 16 |
Local SSD type l5d (Public Beta)
To request testing, please Submit a Ticket
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Compared to previous generation, same core count supports smaller local disk capacity
- Single-instance random read performance up to 140w IOPS, with doubling improvement in per-storage IOPS
- Single instance throughput up to 7GB/s, with 17% improvement in per-storage throughput
- Support Enhanced Cloud Disk Server
-
Applicable scenarios
- OLTP, high-performance relational database
- NoSQL databases (e.g. Cassandra, MongoDB, etc.)
- Search scenarios such as Elasticsearch
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Maximum read IOPS | Maximum read throughput | Maximum write throughput |
|---|---|---|---|---|---|---|---|---|---|
| bcc.l5d.c8m32.1d | 8 cores | 32 GiB | 1*894GiB | 3 Gbps | 1,400,000 | 8 | 350,000 | 1750MB/s | 1200MB/s |
| bcc.l5d.c16m64.1d | 16 cores | 64 GiB | 1*1788GiB | 6 Gbps | 2,800,000 | 8 | 700,000 | 3500MB/s | 2400MB/s |
| bcc.l5d.c24m96.2d | 24 cores | 96 GiB | 2*1788GiB | 9 Gbps | 4,200,000 | 12 | 1,400,000 | 7000MB/s | 4800MB/s |
| bcc.l5d.c32m128.2d | 32 cores | 128 GiB | 2*1788GiB | 11 Gbps | 5,000,000 | 16 | 1,400,000 | 7000MB/s | 4800MB/s |
Local SSD l5
Instance Family Characteristics
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
- Single local disk random read IOPS up to 70w
- Support Enhanced Cloud Disk Server
-
Applicable scenarios
- OLTP, high-performance relational database
- NoSQL databases (e.g. Cassandra, MongoDB, etc.)
- Search scenarios such as Elasticsearch
Specification
| Instance specification. | vCPU | Memory | Local (count of blocks * GiB) disk | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Maximum read IOPS | Maximum read throughput | Maximum write throughput |
|---|---|---|---|---|---|---|---|---|---|
| bcc.l5.c16m64.1d | 16 cores | 64 GiB | 1 block * 3570 GiB | 6 Gbps | 2,800,000 | 8 | 700,000 | 6GB/s | 3GB/s |
| bcc.l5.c32m128.1d | 32 cores | 128 GiB | 1 block * 3570 GiB | 11 Gbps | 5,000,000 | 16 | 700,000 | 6GB/s | 3GB/s |
| bcc.l5.c56m238.2d | 56 cores | 238 GiB | 2 blocks * 3570 GiB | 20 Gbps | 7,500,000 | 32 | 1,400,000 | 12GB/s | 6GB/s |
| bcc.l5.c112m476.4d | 112 cores | 476 GiB | 4 blocks * 3570 GiB | 35 Gbps | 12,000,000 | 32 | 2,800,000 | 24GB/s | 12GB/s |
Local SSD l3d
Instance Family Characteristics
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- Processor-to-memory ratio is 1:4
- Instance network performance scales with specification—higher specifications deliver stronger forwarding capacity and greater intranet bandwidth
- Data disk: NVMe SSD local disk
-
Applicable scenarios
- OLTP, high-performance relational database
- NoSQL databases (e.g. Cassandra, MongoDB, etc.)
- Search scenarios such as Elasticsearch
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.l3d.c16m64.1d | 16 cores | 64 GiB | 1 block * 3570 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.l3d.c32m64.1d | 32 cores | 64 GiB | 1 block * 3570 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.l3d.c32m128.2d | 32 cores | 128 GiB | 2 blocks * 3570 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.l3d.c40m174.2d | 40 cores | 174GiB | 2 blocks * 3570 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.l3d.c64m256.4d | 64 cores | 256 GiB | 4 blocks * 3570 GiB | 22 Gbps | 7,000,000 | 16 |
| bcc.l3d.c80m348.4d | 80 cores | 348GiB | 4 blocks * 3570 GiB | 22 Gbps | 7,000,000 | 16 |
Local SSD la2
Instance Family Characteristics
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:4
- Support Enhanced Cloud Disk Server
- Single-disk random read IOPS up to 65w
- Maximum network packet sent/received capacity: 1,200w
-
Applicable scenarios
- OLTP, high-performance relational database
- NoSQL databases (e.g. Cassandra, MongoDB, etc.)
- Search scenarios such as Elasticsearch
Specification
| Instance specification. | vCPU | Memory | Local disk | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.la2.c16m64.1d | 16 cores | 64 GiB | 1 block * 3570 GiB | 14 Gbps | 4,000,000 | 8 |
| bcc.la2.c32m128.2d | 32 cores | 128 GiB | 2 blocks * 3570 GiB | 30 Gbps | 6,000,000 | 16 |
| bcc.la2.c64m256.4d | 64 cores | 256 GiB | 4 blocks * 3570 GiB | 60 Gbps | 12,000,000 | 32 |
GPU instance family
For GPU instance details, visit the GPU Cloud Server Official Website page
Compute-optimized gn5
Equipped with Intel Xeon Icelake and NVIDIA Ampere series GPU
Ampere A100-80G series
-
Applicable scenarios:
- Large-scale AI training: Image recognition, object detection, autonomous driving, semantic analysis, ad recommendation, etc.
- High-performance computing: Life sciences, industrial manufacturing simulation, weather forecasting, computational physics, etc.
- High-performance data analysis: dataset visualization, big data analytics, etc.
-
Features:
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- GPU: NVIDIA Tesla A100, 312TFLOPS FP16 performance, 80GB HBM2 memory per GPU, supporting 600GB bidirectional Nvlink interconnection
- Memory: Up to 960 GiB memory
- Storage: Support CDS cloud disk server
- Network: Support up to 200 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
Instance specification.
| Instance | A100 GPU count | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.gn5.c14m120.1A100-80g | 1 unit | 14 cores | 120 GiB | 25 Gbps | 6,250,000 | 8 |
| bcc.gn5.c28m240.2A100-80g | 2 units | 28 cores | 240 GiB | 50 Gbps | 12,500,000 | 16 |
| bcc.gn5.c56m480.4A100-80g | 4 units | 56 cores | 480 GiB | 100 Gbps | 25,000,000 | 32 |
| bcc.gn5.c112m960.8A100-80g | 8 units | 112 cores | 960 GiB | 200 Gbps | 50,000,000 | 32 |
Ampere A10 series
-
Applicable scenarios:
- AI inference: Image recognition, object detection, autonomous driving, semantic analysis, ad recommendation, etc.
- High-performance data analysis: dataset visualization, big data analytics, etc.
- Graphics rendering: 3D graphics rendering, graphic design, cloud gaming, etc.
-
Features:
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- GPU: NVIDIA Tesla A10, support RTX ray-tracing algorithm acceleration, 125TFLOPS FP16 performance, 24GB memory per GPU
- Memory: Up to 476 GiB memory, 1:4 processor-to-memory ratio
- Storage: Support CDS cloud disk server
- Network: Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
Instance specification.
| Instance | A10 GPU count | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.gn5.c16m64.1a10 | 1 unit | 16 cores | 64 GiB | 6 Gbps | 1,200,000 | 8 |
| bcc.gn5.c32m128.2a10 | 2 units | 32 cores | 128 GiB | 12 Gbps | 2,500,000 | 16 |
| bcc.gn5.c64m256.4a10 | 4 units | 64 cores | 256 GiB | 18 Gbps | 4,500,000 | 32 |
| bcc.gn5.c28m112.1a10 | 1 unit | 28 cores | 112 GiB | 8 Gbps | 1,800,000 | 8 |
| bcc.gn5.c56m224.2a10 | 2 units | 56 cores | 224 GiB | 14 Gbps | 3,000,000 | 16 |
| bcc.gn5.c112m476.4a10 | 4 units | 112 cores | 476 GiB | 25 Gbps | 8,000,000 | 32 |
Ampere A100-40G series
-
Applicable scenarios:
- Large-scale AI training: Image recognition, object detection, autonomous driving, semantic analysis, ad recommendation, etc.
- High-performance computing: Life sciences, industrial manufacturing simulation, weather forecasting, computational physics, etc.
- High-performance data analysis: dataset visualization, big data analytics, etc.
-
Features:
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- GPU: NVIDIA Tesla A100, 312TFLOPS FP16 performance, 40GB HBM2 memory per GPU, supporting 600GB bidirectional Nvlink interconnection
- Memory: Up to 896 GiB memory, 1:8 processor-to-memory ratio
- Storage: Support CDS cloud disk server
- Network: Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
Instance specification.
| Instance | A100 GPU count | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.gn5.c112m896.8A100-40g | 8 units | 112 cores | 896 GiB | 25 Gbps | 7,000,000 | 16 |
Compute-optimized gn3
Equipped with Intel Xeon Cascade series and mainstream NVIDIA datacenter GPU
Volta V100 series
-
Applicable scenarios:
- AI training and large model inference: Image recognition, object detection, autonomous driving, semantic analysis, ad recommendation, etc.
- High-performance HPC computing: Life sciences, CAE, weather forecasting, collision, etc.
- High-performance data analysis: dataset visualization, big data analytics, etc.
- Encoding/decoding: video/image stream processing, etc.
-
Features:
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- GPU: NVIDIA Tesla V100 SXM2 with 125 TFLOPS FP16 computing power, 32GB HBM2 memory per GPU, and 300GB/s NVLink GPU interconnect bandwidth
- Memory: Up to 640 GiB memory
- Storage: Support CDS cloud disk server
- Network: Support up to 25 Gbps internal bandwidth with ultra-high network packet sent/received capabilities, meeting extremely high intranet transmission requirements
Instance specification.
| Instance | V100-32GB GPU count | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.gn3.c10m80.1v100-32g | 1 unit | 10 cores | 80 GiB | 4 Gbps | 1,500,000 | 8 |
| bcc.gn3.c20m160.2v100-32g | 2 units | 20 cores | 160 GiB | 7 Gbps | 2,800,000 | 8 |
| bcc.gn3.c40m320.4v100-32g | 4 units | 40 cores | 320 GiB | 13 Gbps | 5,000,000 | 16 |
| bcc.gn3.c80m640.8v100-32g | 8 units | 80 cores | 640 GiB | 25 Gbps | 7,000,000 | 16 |
Turing T4 series
-
Applicable scenarios:
- AI inference: Image recognition, object detection, autonomous driving, semantic analysis, ad recommendation, etc.
- High-performance data analysis: dataset visualization, big data analytics, etc.
- Encoding/decoding: video/image stream processing, etc.
-
Features:
- Processor: Intel Xeon Gold 6271C, base frequency 2.6 GHz, turbo frequency 3.1 GHz
- GPU: NVIDIA Tesla T4 with 130 TFLOPS INT8 computing power and 16GB memory per GPU
- Memory: Up to 320 GiB memory
- Storage: Support CDS cloud disk server
- Network: Ultra-high packet sent/received capabilities, meeting extremely high intranet transmission requirements
Instance specification.
| Instance | T4 GPU count | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| bcc.gn3.c20m80.1t4 | 1 unit | 20 cores | 80 GiB | 6 Gbps | 2,800,000 | 8 |
| bcc.gn3.c40m160.2t4 | 2 units | 40 cores | 160 GiB | 11 Gbps | 5,000,000 | 16 |
| bcc.gn3.c80m320.4t4 | 4 units | 80 cores | 320 GiB | 22 Gbps | 7,000,000 | 16 |
Elastic baremetal instance family
Elastic baremetal local SSD type l8
Instance Family Characteristics
- Processor: Intel® Xeon® 6 Granite Rapids-AP, base frequency 3.3GHz, all-core turbo frequency 3.7GHz, with strong single-core performance
- Processor-to-memory ratio is 1:4
- Comprehensive IO capability improvement, with significant performance enhancement in local disks compared to previous generations
- Support eRDMA capability
-
Applicable scenarios
- Gaming servers, high-frequency interaction scenarios (e.g., live video barrage)
- Data analysis, batch computing
- High-performance scientific and engineering applications
Specification | Instance specification. | vCPU | Memory | Local storage (count of blocks capacity) | Baseline intranet bandwidth | Network packet sent/received PPS | Primary network interface card queue count | ---|---|---|---|---|---|---| | ebc.l8.c192m768.2d | 192 cores | 768 GiB | 2 3570GiB | 160 Gbps | 36,000,000 | 64 |
Elastic baremetal local SSD type la4
Instance Family Characteristics
- Processor: Equipped with AMD EPYC™ Turin Classic Processor, base frequency 2.7GHz, maximum turbo frequency 4.1GHz
- Processor-to-memory ratio is 1:4
- Comprehensive IO capability improvement, with significant performance enhancement in local disks compared to previous generations
- Support eRDMA capability
-
Applicable scenarios
- Web application real servers
- Data analysis, batch computing
- Engineering applications, simulation processing
- Video encoding/decoding
Specification
| Instance specification. | vCPU | Memory | Local storage | Baseline intranet bandwidth | Network packet sent/received PPS | Primary network interface card queue count | Cloud disk bandwidth |
|---|---|---|---|---|---|---|---|
| ebc.la4.c256m1152.2d | 256 cores | 1152 GiB | 2* 3570GiB | 160 Gbps | 36,000,000 | 64 | 32 Gbps |
| ebc.la4.c256m1152.16d | 256 cores | 1152 GiB | 16* 3570GiB | 160 Gbps | 36,000,000 | 64 | 32 Gbps |
Elastic baremetal local SSD type hcl7
Instance Family Characteristics
- Processor: Intel® 5th Xeon® Emerald Rapids, base frequency 3.3GHz, all-core turbo frequency 3.4GHz
- Processor-to-memory ratio is 1:8
- Deliver peak-performance instance configurations with customizable hyper-threading and NUMA support, optimized for multiple high-clock-speed instance scenarios.
-
Applicable scenarios
- Gaming servers, high-frequency interaction scenarios (e.g., live video barrage)
- Data analysis, batch computing
- High-performance scientific and engineering applications
Specification
| Instance specification. | vCPU | Memory | Local storage | Baseline intranet bandwidth | Network packet sent/received PPS | Primary network interface card queue count |
|---|---|---|---|---|---|---|
| ebc.hcl7.c128m1024.4d | 128 cores | 1024 GiB | 4 *3570 GiB | 160 Gbps | 30,000,000 | 104 |
Elastic baremetal local SSD type l7
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4.92
- Provides local NVMe SSD storage with physical data isolation for enhanced efficiency and security.
-
Applicable scenarios
- Database business
- Distributed caching for online computing scenarios
Specification
| Instance specification. | vCPU | Memory | Local storage (count of blocks * capacity) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count | Cloud disk bandwidth |
|---|---|---|---|---|---|---|---|
| ebc.l7.c208m1024.2d | 208 cores | 1024 GiB | 2 *3570 GiB | 160 Gbps | 30,000,000 | 64 | 32 Gbps |
| ebc.l7.c208m1024.4d | 208 cores | 1024 GiB | 4 *3570 GiB | 160 Gbps | 30,000,000 | 64 | 32 Gbps |
Elastic Baremetal Compute general-purpose g7
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4.9
-
Applicable scenarios
- Web and application servers
- Distributed caching systems
- Search cluster
- Conventional data analysis and computational tasks
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| ebc.g7.c208m1024 | 208 cores | 1024 GiB | 160 Gbps | 30,000,000 | 64 |
Elastic Baremetal Compute local SSD type la3
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: AMD EPYC 9W24, base frequency 2.6GHz, turbo frequency 3.7GHz
- Processor-to-memory ratio is 1:4
- Provides local NVMe SSD storage with physical data isolation for enhanced efficiency and security.
- Delivers ultra-high network performance for an exceptional data interaction experience.
-
Applicable scenarios
- Database business
- Distributed caching for online computing scenarios
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| ebc.la3.c384m1536.2d | 384 cores | 1536 GiB | 2 blocks *3570 GiB | 160 Gbps | 30,000,000 | 64 |
Elastic Baremetal Compute local SSD type l5
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratios include 1:4 and 1:8
- Provides local NVMe SSD storage with physical data isolation for enhanced efficiency and security.
-
Applicable scenarios
- Database business
- Distributed caching for online computing scenarios
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| ebc.l5.c128m512.1d | 128 cores | 512 GiB | 1 block * 3570 GiB | 40 Gbps | 10,000,000 | 32 |
| ebc.l5.c128m512.4d | 128 cores | 512 GiB | 4 blocks * 3570 GiB | 40 Gbps | 10,000,000 | 32 |
| ebc.l5.c128m1024.1d | 128 cores | 1024 GiB | 1 block * 3570 GiB | 40 Gbps | 10,000,000 | 32 |
Elastic Baremetal Compute general-purpose g5
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- Processor-to-memory ratio is 1:4
-
Applicable scenarios
- Web and application servers
- Distributed caching systems
- Search cluster
- Conventional data analysis and computational tasks
Specification
| Instance specification. | vCPU | Memory | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|
| ebc.g5.c128m512 | 128 cores | 512 GiB | 40 Gbps | 10,000,000 | 32 |
Elastic Baremetal Compute local SSD type la2ne
Instance Family Characteristics
- Leverages Baidu Taihang Architecture to offer exclusive high-performance servers in the cloud, featuring physical resource isolation.
- Processor: AMD EPYC 7W83, base frequency 2.55GHz, turbo frequency 3.5GHz
- Processor-to-memory ratio is 1:4
- Provides local NVMe SSD storage with physical data isolation for enhanced efficiency and security.
- Delivers ultra-high network performance for an exceptional data interaction experience.
- Support eRDMA
-
Applicable scenarios
- Database business
- Distributed caching for online computing scenarios
Specification
| Instance specification. | vCPU | Memory | Local disk (count of blocks * GiB) | Baseline intranet bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|
| ebc.la2ne.c256m1024.2d | 256 cores | 1024 GiB | 2 blocks *3570 GiB | 180 Gbps | 30,000,000 | 64 |
| ebc.la2ne.c256m1024.12d | 256 cores | 1024 GiB | 12 blocks *3570 GiB | 180 Gbps | 30,000,000 | 64 |
Elastic Baremetal Compute GPU type gn7s
This instance family is currently Allow List-only. Submit a ticket if needed
Applicable scenarios:
- AI inference
- Graphics and video rendering
Features:
- Feature Baidu Taihang·Elastic Baremetal Compute capabilities
- Support selecting public images and compatible GPU driver versions when creating or reinstalling OS on instances
- Processor: Intel Xeon Platinum 8563C, base frequency 2.6GHz, turbo frequency 3.1GHz
- GPU: NVIDIA Tesla L20 with 120TFLOPS single GPU FP16 computing power and 40GB GDDR6 memory per GPU
- Network: Support 160G VPC network bandwidth
Specification
| Instance specification. | vCPU | Memory | GPU count | Total GPU memory | VPC bandwidth | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|---|
| ebc.gn7s.c208m1024.8l20 | 208 cores | 1024 GiB | 8 | 384G | 160 Gbps | 30,000,000 | 64 |
Elastic high-performance computing cluster
Elastic high-performance computing cluster GPU type GN6T
This instance family is currently Allow List-only. Submit a ticket if needed
Applicable scenarios:
- AI large-scale training: Autonomous driving, NLP models, ad recommendation, etc.
- High-performance computing: Life sciences, computational physics, etc.
Features:
- Feature Baidu Taihang·Elastic Baremetal Compute capabilities
- The GPU image includes essential components such as GPU drivers, CUDA, and OFED network interface card drivers.
- Processor: Intel Xeon Platinum 8463B, base frequency 2.6GHz
- GPU: NVIDIA Tesla H800 SXM, 989TFLOPS FP16 performance, 80GB HBM memory per GPU, supporting 400GB bidirectional Nvlink interconnection
- Storage: Support up to 16TB capacity all-flash local SSD storage with GPU Direct Storage support
- Network: Support 180G VPC bandwidth and 3200G RDMA private network, with GPU Direct RDMA capability
Specification
| Instance specification. | vCPU | Memory | GPU count | Total GPU memory | VPC bandwidth | Local disk (count of blocks * GiB) | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|---|---|
| ehc.lgn6t.c192m3072.8h800.8re.4d | 192 cores | 3072 GiB | 8 | 640 G | 180 Gbps | 4*3570GiB | 30,000,000 | 64 |
Elastic high-performance computing cluster GPU type GN5
This instance family is currently Allow List-only. Submit a ticket if needed
Applicable scenarios:
- AI large-scale training: Autonomous driving, NLP models, ad recommendation, etc.
- High-performance computing: Life sciences, computational physics, etc.
Instance family characteristics:
- Feature Baidu Taihang·Elastic Baremetal Compute capabilities
- The GPU image includes essential components such as GPU drivers, CUDA, and OFED network interface card drivers.
- Processor: Intel Xeon Platinum 8350C, base frequency 2.6GHz, turbo frequency 3.1GHz
- GPU: NVIDIA Tesla A800 SXM, 312TFLOPS FP16 performance, 80GB HBM2 memory per GPU, supporting 400GB bidirectional Nvlink interconnection
- Storage: Support up to 16TB capacity all-flash local SSD storage with GPU Direct Storage support
- Network: Support 180G VPC bandwidth and 800G RDMA private network, with GPU Direct RDMA capability
Specification
| Instance specification. | vCPU | Memory | GPU count | Total GPU memory | VPC bandwidth | Local disk (count of blocks * GiB) | Network packet sent/received PPS | Queue count |
|---|---|---|---|---|---|---|---|---|
| ehc.lgn5.c128m1024.8a800.8re.4d | 128 cores | 1024 GiB | 8 | 640 G | 180 Gbps | 4*3570GiB | 30,000,000 | 64 |
Instance families in public beta/invitation-only testing are ready for customer testing in terms of product or feature. Instance families in public beta are generally open, while invitation-only instance families require allow list access via Submit Ticket for creation application. These instances may have incomplete features and no Service Level Agreement (SLA). Issues can be reported via Submit Ticket for feedback.
Public beta or invitation-only testing usually lasts between 1 to 2 months on average.
