Traditional method - security cameras streaming video to a central recording computer (DVR, NVR). This design philosophy was introduced with the earliest CCTV systems over 50 years ago. Here the cameras are connected via data cable to a central recording device, which has evolved from recording at a VCR to recording to a DVR/NVR. Most systems are installed this way.
Two new methods have been introduced to lower bandwidth and increase system redundancy:
Distributed Architecture
This is a method of distributing multiple recording computers to smaller clusters of cameras located on the network. While it offers greater redundancy and lower bandwidth costs than the traditional method, this is easily the most expensive of all the options as the cost of hardware and software is increased. A central computer is still required for database synchronization, although some management software allows for local workstations to perform database synch. Either way, distributed architecture carries a hefty price tag.
Edge Recording
Primarily this is a system where the recording is either in or at the camera. Typically it’s about dedicating a single recording device, such as an SD/SDXC card or attached storage (i.e. Ethernet HDD, or RAID) to each camera. The advantage to edge recording is that it offers the highest redundancy and lowest bandwidth utilisation of all three storage platforms. It's also the least inexpensive as it only requires standard storage media for recording.
