# Reference hardware and performance
# 1. In the case of Aliyun server configuration
| No. | Services Name | Instance Scale | OS | output Bandwidth |
|---|---|---|---|---|
| 1 | General Control（API function、Web、Task scheduling、Authentication） | ecs.g5ne.xlarge：4C/16GB/50GB | Ubuntu 20.04 X64 | 10Mbps |
| 2 | Transcoding（Transcoding、Word segmentation） | ecs.g5ne.xlarge：4C/16GB/100GB SSD | Ubuntu 20.04 X64 | 1Mbps |
| 3 | Compositing（Audio comparison、Compositing） | ecs.gn5i-c4g1.xlarge：4C/16GB/100GB SSD/1*Tesla P4 | Ubuntu 20.04 X64 | 1Mbps |
| 4 | Database | PolarDB：4C/16GB | MYSQL5.7 | 1Mbps |
| 5 | Storage | Object Storage | ||
| 6 | Domain name1，Certificate1 |
# 2. Performance
# 1.Transcoding Service Reference Performance：
Original material：100 seconds/MP4/1080P/1.75Mbps/H264；
processing time：25.6 seconds.（depends on input file's specification and output file's specification）
number of streams：10 . Configurable and queueable，Parallel processing can reduce the efficiency of individual task
processing.
# 2.Compositing Service Reference Performance：
# 1.Instance Scale：(ecs.gn5i-c4g1.xlarge：4C/16GB/100GB SSD/1*Tesla P4)
input file：MP4/1080P/1.75Mbps/H264/Multi-segment; timeline complexity: 2-track video + 1-track audio + 1-track
vocals subtitles + 1-track filters/100 seconds；
output file：100 seconds/MP4/1080P/1.75Mbps/H264；
processing time：72.2 seconds，（depends on the timeline complexity、input file's specification and output file's
specification）
number of streams：10 . Configurable and queueable，Parallel processing can reduce the efficiency of individual task
processing.
# 2.Instance Scale：(ecs.gn6i-c16g1.4xlarge：16C/62GB/40GB SSD/NVIDIA T4)
input file：MP4/1080P/2Mbps/H264/Multi-segment; timeline complexity: 2-track video + 1-track audio + 1-track vocals
subtitles + 1-track filters/600 seconds；
output file：600 seconds/MP4/1080P/2Mbps/H264；
processing time：112 seconds，（depends on the timeline complexity、input file's specification and output file's
specification）
number of streams：10 . Configurable and queueable，Parallel processing can reduce the efficiency of individual task
processing.
# 3. Platform concurrency
Estimated number of concurrent users of the platform (simultaneous use of cloud clips online): about 10,000，it depends on the General Control server's performance and database server's performance.
# 4. Client
Recommended configuration for clients：above intel 5th gen i5/8GB RAM/100GB free diskspace
Reference configuration and performance：intel i5-10210U/16GB RAM/256GB SSD，real-time 8-10 layers editing is
possible
Operating system: 32-bit is not be supported
Supported browsers:
- Chrome, Opera, 360 browser and other browsers with Chrome kernel , version 75 inclusive or above supported.
- Firefox: supported with kernel version greater than 58 inclusive (special case: version 72 is not be supported).