Categories
Sales Materials Storage Calculator Documentation Categories Products exacqVision Hardware

How exacqVision Performance is Benchmarked

For help selecting the correct exacqVision Server hardware, please refer to our Configuration Calculator.

Exacq system performance is benchmarked under typical real-world conditions to provide meaningful context in selecting a system that meets the needs of your application. All performance guidelines are approximate, and we recommend having at least 20% extra data rate capacity for your installation. The following definitions and guidelines are used by our Configuration Calculator and should be kept in mind when configuring any system.

IP Stream

An IP stream is a video or audio connection transmitted over the network to the server. Many cameras are capable of providing more than one stream simultaneously.

Maximum IP Streams

The highest number of IP Streams the system is capable of recording. This is an upper limit on the number of streams that can be handled and may not be reachable based on the total amount of data from connected devices. Secondary streams from the same camera are included in this number. Datasheets for each system may be found on the Product pages of Exacq.com.

Video Throughput

The maximum amount of data that can be recorded while simultaneously performing reads for video search. Testing is performed so that users performing forensic searches should not expect significant delays in retrieving results while being able to record without interruption. Datasheets for each system may be found on the Product pages of Exacq.com.

The Maximum Throughput may be lower than the theoretical maximum disk throughput values reported by other’s systems but allows us to provide the expectation that users may perform functional searches in a reasonable time, whereas citing a higher write throughput to the disk would negatively impact a user performing searches in a real-world use case. Throughput values are also tested to ensure that storage drives, which are 100% full, are capable of removing old video files without negatively affecting the ability to write new files.

Our Configuration Calculator considers both Maximum IP Streams as well as the data throughput when recommending system hardware. Whichever limit is reached first is considered the cutoff before suggesting adding another system to the overall configuration.

While typically used to select an appropriate size and number of servers for new purchase, the Configuration Calculator may be a useful tool for existing systems when trying to determine if additional cameras may be added.

NOTE: X-Series systems manufactured through our partnership with Dell will list an increase in the number of streams over the prior generation Z-Series despite slightly lower disk throughput values. This is due to the higher prioritization the RAID management these systems place on writing data to disk versus read activity.

Network Interface

Tests are performed using default interface settings. With regards to the disk throughput, the network interface is a determining factor, e.g.- a disk that can write 3 Gbps will be limited to only 2.5 Gbps if that is the maximum throughput of the data crossing the network interface to be written to disk.

Active Search

When considering the ability to perform reasonable forensic searches, tests are performed to ensure that 2 hours of video from 4 simultaneous streams are returned within 2 minutes.

Operating System

Exacq systems are available with Windows or Ubuntu Linux operating systems. Each operating system is tested independently due to Linux ext4 file system’s ability to provide higher throughput in many server configurations than the Windows NTFS file system. Ext4 is capable of handling larger ranges of contiguous blocks of data, allowing it to work more efficiently with large files, reducing drive fragmentation, and has shown more efficient journaling and checksum processes.

Display Rate

The local client’s display frame rate represents a client running directly on the chosen system, rather than a networked client workstation, to decompress and display live or recorded video. This is tested using 1080p H.264 video streams with default quality settings. H.264 requires significantly more CPU power to decompress than MJPEG or MPEG; thus, performance is the same or better using MJPEG or MPEG compression. Testing determines the total frames capable of being displayed without any additional applications running.

<br>

Categories
Knowledge Support Support exacqVision Server Categories Products

How exacqVision Server Performance is Benchmarked (Legacy)

IMPORTANT: This is a legacy document which applies to older systems built primarily before mid-2024. Newer systems manufactured by Dell or Avnet for Exacq were benchmarked according to the latest version of this article at How exacqVision Performance is Benchmarked.


For help selecting the correct exacqVision server for your application, please use our configuration calculator.

exacqVision Server performance is benchmarked under typical real-world conditions to provide meaningful guidance in selecting the correct server for your application. All performance guidelines are approximate. Exacq recommends having at least 20% spare data rate capacity for your application.

Maximum Camera Data Rate represents the server’s ability to continuously record, search, and play back video. This value is measured during a 24-hour load test under the following conditions:


Maximum IP cameras: The server is connected to the maximum number of IP cameras. Reducing the number of IP cameras increases the continuous data rate by lowering the operating system’s file management overhead. 

Drives 100% full: This ensures the server is capable of deleting old video files without negatively affecting the ability to write new files.

Local client operational: The local client is operated at the rated frame rate. Remote clients (running on a seperate PC) typically have a minimal impact on server performance. The client application consumes significant CPU capacity for video decompression. 

Minimum disk quantity: Servers are tested with a typical disk quantity for each server model, as shown in the following list. Video storage is usually limited by disk input/output capacity; increasing the disk quantity increases the server’s video storage rate.

*EL & ELX series: 1 drive
*A series desktop: 1 drive
*A series 2U: 2 drives
*A series 4U: 4 drives (RAID 5)
*Z series 2U: 4 drives (RAID 5)
*Z series 3U: 6 drives (RAID 6)
*Z series 4U: 8 drives (RAID 6)

Search active: The server must perform a search for 30 minutes of video across eight cameras without negatively affecting local client or video storage operation.

Operating System: Each OS is tested independently because the Linux EXT4 file system has shown higher throughput in some server configurations than the Windows NTFS file system. EXT4 can handle larger ranges of contiguous physical blocks of data, allowing it to work more efficiently with large files and reduce drive fragmentation. EXT4 has also shown more efficient journaling and checksum process.

Local Client Display Frame Rate represents the client application’s ability to decompress and display live or recorded video with 80% of CPU load. This is tested using a D1-sized (.36MP) H.264 video stream with a quality of 9. H.264 requires significantly more CPU power to decompress than MJPEG or MPEG; thus, performance is the same or better using MJPEG or MPEG compression.

When using cameras with large image sizes, divide the camera image size in megapixels by .36 to determine the equivalent client display load. For example, a 720×1080 camera (.77MP) requires about 2.2 D1 frames worth of decompression capacity. A server rated for 500FPS at D1 can display 230FPS at 720×1080.

<br>