- An increase of channels just means an increase of NVRs.
- Multiple stand-alone NVRs are usually managed by embedded software.
- As the number of channels increases, client-server architecture may help manage more effectively.
When the number of channels increases — say, to 64 or hundreds of channels — the principles remain the same, but the architecture and implementation have to evolve.
One way is through stacking standalone NVRs: an increase of channels just means an increase of NVRs. Individual NVRs, added as needed, have the advantage of being close to the familiar DVR or VCR model that many security practitioners are used to, said Jeff Whitney, VP of Marketing at Intransa. Stacking NVRs is also less expensive, as NVRs often use older components and technologies. Sometimes, all you need is the simplicity of NVRs.
There are, however, other more efficient options. “With the current generation of specialized video processing hardware, there is no need for multiple processors in multiple NVRs,” said Jumbi Edulbehram, VP of Business Development, Next Level Security Systems. “Streaming video directly to storage can be more efficient.”
System architecture, on the other hand, can be more complicated. Servers with the needed amount of storage can be purchased with third-party VMS preloaded, offering the flexibility to add channels incrementally and decrease the number of appliances in the surveillance system. This flexibility, however, also means that installation is far more complicated than installation of multiple NVRs, as the number of architectural options also increases.
Multiple stand-alone NVRs are usually managed by embedded software. “All the vendor-specific features are integrated into one program,” said Aaron Yeh, Sales and Marketing Director for Surveon Technology. Control is from either one of the NVRs or from the overlaying software (similar to CMS).
“Like the DVRs they followed, they are individually managed systems,” Jeff Whitney, VP of Marketing at Intransa, explained further. “That means an operator is required to manually operate and maintain them, or at best run some basic overlaying software that allows them to switch between them.”
The overlaying software, or CMS, allows for easier and quicker management of multiple NVRs, many channels and different clients. “But there are some disadvantages,” Yeh said. “All video streams must go through the CMS client, which puts it at higher risk for single point of failure.”
As the number of channels increases, client-server architecture may help manage more effectively. Client-server architecture is found in some NVRs, and also found in servers and PCs plus VMS installations. Recording is controlled from a PC with client software; NVRs with client-server architecture or servers with VMS allow for more centralized management of recording storage. Client-server architecture is also better at accommodating multiple clients. “You can have multiple clients, and one recording server streams the video to different clients,” Yeh said.
When video is stored on intake NVRs, they are separate islands of information as storage is not shared, and it may be difficult to recall and manage video as you would like. This is similar to when information is stored separately on multiple servers or when DAS is added to NVRs or servers.
“If another NVR in the system has unused capacity, it is not simple to move the recording capacity to that system without at least shutting everything down and recabling,” Whitney said. “As such, groups of NVRs often experience some over utilizing and some underutilizing over time, wasting equipment resources and costing excess energy and environmental needs.”
In practice, resellers end up over provisioning each NVR, since the performance and capacity are limited to each box, said Lee Caswell, founder and Chief Strategy Officer of Pivot3.
With higher camera counts, centralized storage becomes more efficient and a necessity. NAS, FC-SAN and IP-SAN all provide centralized, shared storage among different appliances on a network, like front-end servers or NVRs. “Storage can be in a centralized pool with multiple NVRs,” Yeh said. “The NVRs on the front end don't need any disks; all the NVRs will record their video to centralized storage. In this kind of architecture, the NVRs share the channel loading from all the cameras on the system.”
Any added storage is shared. “If a modular approach was used in place of standard NVRs, additional external storage modules could be plugged in as additional capacity was needed,” Whitney said. “That additional capacity could support multiple servers. So, less expensive servers, without internal storage, can be added for additional camera support.”
NAS has file-level transfer protocol, which results in slower, less efficient transfer of data. In video surveillance, this may cause a bottleneck in the system. There have been improvements made to NAS in recent years, but whether the NAS that users purchase makes use of these advancements is another issue.
SAN, with its more efficient block level transfer protocol, is more favored in recent years, especially with the budget-friendly alternative of IP-SAN that uses iSCSI. “In 2010, the use of IP- and FC-SANs began to impact video surveillance storage,” said Sam Grinter, Market Analyst at IMS Research, in a prepared statement. “Recording video surveillance data is very intensive on storage write cycles. SANs offer the benefit of using virtualized, pooled storage which has the advantage of increased performance over file-based storage systems, often characteristic in DAS or NAS systems.”
Energy consumption, cooling and rack space are often overlooked, and experienced integrators and IT personnel should be brought in to eschew system-level, overheating disasters, Law cautioned.