Volcengine Kubernetes Engine
Volcengine Kubernetes Engine
Volcengine Kubernetes Engine
Document
Volcengine Kubernetes Engine
What's New
History feature release notes
Kubernetes version release notes
Containerd runtime release notes
Add-on release notes
Monitoring add-ons
Scheduling add-ons
DNS add-ons
Security add-ons
GPU add-ons
Sandbox add-on
Security announcements
Product Announcement
Historical announcements
Overview
Getting started
User guide
Clusters
Creating clusters
Connecting to clusters
Upgrading clusters
Node pool
Creating node pools
Scaling out/in node pools
Nodes
Authorization management
Image
Namespaces
Networking
Cluster network
Container Network
Service
LoadBalancer
CLB
Ingress
APIG Ingress
ALB Ingress
Customize ALB Ingress forwarding rules
Certificate and security
CLB Ingress
Nginx Ingress
Gateway API
Workloads
Creating workloads
Configure workloads
Pods
Configuration management
Image acceleration
Storage
FSX client
EBS Volumes
TOS Volumes
NAS volumes
FSX access methods
NFS access methods
EFS volumes
vePFS volumes
FSX access methods
vePFS Client access methods
NFS access methods
CloudFS volumes
Local volumes
Autoscale
Node autoscaler
Workload autoscaler
Horizontal Pod Autoscaler (HPA)
Cron Horizontal Pod Autoscaler (CronHPA)
Intelligent Horizontal Pod Autoscaler (IHPA)
Horizontal Pod Autoscaler based on event-driven (KEDA)
Add-on management
Network add-ons
Storage add-ons
Monitoring add-ons
Scheduling add-ons
DNS add-ons
GPU add-ons
Core add-ons
Sandbox add-on
Upgrading add-ons
Heterogeneous computing
NVIDIA GPU
mGPU scheduling
RDMA high-performance computing
Scheduling management
Topology-aware scheduling
QoS-aware scheduling
Observability
Log management
Basic observation
Setup & Configuration
Cluster observation
Storage observation
Other observation
Using metrics
Full stack observation
Alert management
Diagnosis and inspection
Node check and self-healing
Application center
Backup center
Stability
Large-scale cluster management
Best practices
Authorization
Clusters
Nodes and node pools
Networking
Storage
Scheduling
Observability
Solution
Security and compliance
API reference
Cluster management
NodePool management
Node management
Virtual nodes management
Add-on management
Tag management
Access policy management
Instance images management
Cloud native observability
Inspection and diagnosis
Task management
Volcengine Container Instance
Appendix
SDK reference
Development guide
Terraform
FAQs
General FAQs
Cluster FAQ
cluster management
cluster network
cluster upgrade
Node and Node pool FAQs
Authorization FAQs
Workload FAQs
Service and Ingress FAQs
Storage FAQs
Heterogeneous computing FAQs
Observability FAQs
Quota FAQs
Add-on FAQs
Cloud-native AI
ServingKit
LLM applications
ComfyUI application
Deploying the Al inference application through Helm
Deepseek practices
Qwen practices
Kimi practice
Seed practices
Best practices
DeepSeek model deployment
GLM model deployment
MiniMax model deployment
Qwen model deployment
Kimi model deployment
Volcengine Container Instance
What's new
Historical feature release notes
Product Announcement
Overview
Billing
User guide
Integrating with VCI
Virtual node
VKE cluster uses virtual nodes
Self-managed/third-party cluster uses virtual node
Creating an instance
General purpose instance
GPU-accelerated computing instance
VCI Agent Sandbox
Startup acceleration
E2B-based management
Upgrade guide
Networking
Monitoring
Scheduling
Container configuration
VCI Image Cache
Best practices
Security and compliance
FAQs
- Documentation
Volcengine Kubernetes EngineVolcengine Kubernetes EngineUser guideWorkloadsImage accelerationP2P accelerationReference performance of P2P acceleration
Copy page
Download PDFP2P acceleration
Reference performance of P2P acceleration
Copy pageDownload PDFReference performance of P2P acceleration
The P2P acceleration performance varies depending on the storage type used by nodes. To help you evaluate the acceleration performance for container image pulls, this topic provides test data for three storage scenarios: local NVMe SSD, cloud ESSD Flex, and cloud ESSD PL0.
warning
Actual performance is also affected by factors such as cluster size, instance type, network bandwidth, image size, and concurrency. Refer to the test results from your production environment.
Test environment
The tests described in this topic were conducted with the same cluster size, ECS instance type, network performance, and P2P Peer specifications. Only the storage resource type varied. Different storage types have different read/write performance characteristics, resulting in differences in image pull performance. The test environment configuration is shown in the following table.
Test environment | Local disk | Cloud ESSD Flex | Cloud ESSD PL0 |
Test environment | Local disk | Cloud ESSD Flex | Cloud ESSD PL0 |
Cluster size | 300 nodes | | |
ECS instances | Instance specification ecs.i4s.4xlarge | | |
Network performance | Baseline bandwidth 25 Gbit/s | | |
Storage resources | Local NVMe SSD | Cloud ESSD Flex | Cloud ESSD PL0 |
P2P Peer specifications | 1C2G | | |
CR bandwidth | 100Gbit/s | | |
Test result
Local NVMe SSD
When a local NVMe SSD is used as the cache disk, the end-to-end container image startup time (including the time for containerd to download and decompress the image) is shown in the following table.
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
1 GiB (256 MiB × 4 layers) | 3.73 | 4.46 | 5.03 | 9.21 | 20.64 | 1 |
2 GiB (256 MiB × 8 layers) | 6.52 | 8.55 | 9.96 | 11.68 | 12.63 | 1 |
4 GiB (512 MiB × 8 layers) | 11.59 | 13.62 | 15.24 | 17.11 | 18.28 | 1 |
10 GiB (1 GiB × 10 layers) | 25.27 | 28.38 | 29.79 | 33.83 | 37.51 | 1 |
Cloud ESSD Flex
When a cloud ESSD Flex is used as the cache disk, the end-to-end container image startup time (including the time for containerd to download and decompress the image) is shown in the following table.
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
1 GiB (256 MiB × 4 layers) | 9.14 | 9.83 | 10.13 | 13.67 | 13.84 | 1 |
2 GiB (256 MiB × 8 layers) | 16.61 | 17.48 | 17.84 | 21.44 | 21.78 | 1 |
4 GiB (512 MiB × 8 layers) | 34.08 | 36.82 | 37.26 | 39.81 | 47.05 | 1 |
10 GiB (1 GiB × 10 layers) | 115.99 | 125.31 | 127.11 | 131.16 | 137.10 | 1 |
Cloud ESSD PL0
When a cloud ESSD PL0 is used as the cache disk, the end-to-end container image startup time (including the time for containerd to download and decompress the image) is shown in the following table.
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
Image specification | P50 latency (s) | P90 latency (s) | P95 latency (s) | P99 latency (s) | Maximum latency (s) | Number of origin-pull nodes |
1 GiB (256 MiB × 4 layers) | 10.19 | 11.03 | 11.26 | 14.99 | 15.13 | 1 |
2 GiB (256 MiB × 8 layers) | 19.97 | 20.81 | 21.05 | 24.71 | 24.77 | 1 |
4 GiB (512 MiB × 8 layers) | 40.33 | 41.52 | 43.97 | 44.38 | 44.51 | 1 |
10 GiB (1 GiB × 10 layers) | 105.92 | 110.53 | 111.57 | 113.26 | 114.44 | 1 |
Last updated: 2026.09.23 20:01:11
All rights reserved © Beijing Volcano Engine Technology Co., Ltd2024
