Video analytics require sufficient hardware resources to perform, and with large installations, you might need several analytic servers which all have to be managed. But manually overseeing these servers increases the risk of low accuracy, and when multiple analytics run on the same server, it can get overloaded, all while additional computing resources remain unused.
By optimizing hardware utilization, you ensure the highest possible video analytics accuracy while minimizing the total cost of ownership.
Analyzer management enhances video analytics operations
To provide a single visualization of all your video analytics servers, Analyzer management – the new KiwiVision™ video analytics feature – provides an overview of each server’s load, including both their CPU and GPU components.
It also allows you to visualize the video analytics scenarios running on each server, so you can quickly identify bottlenecks.
Benefits of automatic load balancing
Deciding which and how many video analytics scenarios to run on your servers can be a challenge. Choose too few and you won’t use your available hardware resources optimally; too many and your servers get overloaded, resulting in inaccurate analytics.
That’s where automatic load balancing comes in. Automatic load balancing automatically distributes video analytics scenarios between all of your servers. It continuously monitors each servers’ load and configures analytics scenarios automatically, helping to maximize overall resource utilization.
For example, if one of your servers fails or a new one is added, automatic load balancing will reconfigure their analytics scenarios. The configuration is done seamlessly with zero interruptions or user interactions.
Flexible load balancing functions
With a flexible load balancing function, you can assign specific analytics scenarios or cameras to specific servers. If you have a large system geographically dispersed over several locations, analytics should be analyzed as close to the cameras as possible to minimize network load. Streaming video across locations also introduces additional network load and a potential point of failure.
With flexible load balancing and the ability to control server loads manually, you can assign cameras and scenarios to specific servers by “locking” a server.
This essentially takes it out of automatic load balancing and allows you to decide which cameras should be analyzed, helping you optimize your analytics deployment for minimum network traffic.
With Analyzer management, you can ensure optimal analytics performance with less hardware, less network traffic, and the most accurate video analytics results possible.