Assessment: Recording at the Edge
When edge-based storage first became commonplace in the security industry, many saw it as an approach solely designed to tackle issues regarding bandwidth management. However, it actually offers significant benefits in surveillance system design, and delivers a better level of security too. Benchmark looks at the options to assess the potential of edge recording and see if it adds benefits for surveillance archiving needs.
Edge recording is increasingly commonplace in today’s surveillance marketplace. It enhances security, safeguarding against the loss of footage during infrastructure outages. It allows a more flexible approach to system design. It enables the creation of cost-effective solutions. It adds value to surveillance systems of all sizes and complexities. Given these benefits, it’s hard to understand why edge recording still remains a second choice behind a centralised approach.
In many applications, the case for recording at the edge is very strong. Centralised recording has been, and some would argue still is, a cornerstone with regard to surveillance system design. This is not because it offers better functionality, is more secure or aids overall performance. It isn’t because customers demand it, or because it is ‘best practice’.
It is because, historically, surveillance technology was restrictive, and centralised recording was a necessity if systems were to be cost-effective.
When networked solutions first appeared in the surveillance sector, centralised recording remained the dominant method. Many manufacturers simply replicated analogue topology, but with network connectivity. Manufacturers, integrators, installers specifiers, consultants and even end users were used to centralised recording, and as a result few questioned it efficiency.
Network-based solutions are well suited to distributed archiving. They offer the flexibility to use distribute storage without fragmenting elements of the overall system. Many VMS solutions can manage edge recording sources from a variety of locations, while to the operators it appears as if the system is one unified entity.
So, why is edge-based recording a good thing? A centralised approach can be detrimental to performance. Design becomes unnecessarily cumbersome, with many benefits of network infrastructure being ignored. Management elements must address vulnerabilities which the centralised model introduces. A centralised approach can introduce unnecessary duplication, deliver lower performance and cost more.
In many cases, video footage and associated data will only be needed if an incident occurs. However, a centralised model has to transport all the data back to a central point, regardless of whether it contains anything of interest or not. This results in the need for bandwidth management, which means when something does happen, the critical video will probably be ‘throttled’ to ensure all the other streams of irrelevant data can also be transported; after all, an incident could happen anywhere!
Edge-based recording eliminates issues with a single point of failure, it allows flexibility with regard to bandwidth and storage media, it provides potential for different approaches dependent upon risk, and it simplifies scalability and future expansions.
There will always be some applications where a centralised approach remains the best choice, but in the majority of cases edge recording offers significant benefits.
Edge recording is not limited to archiving at the camera or encoder. Video can be archived on numerous networked devices: NVRs, NAS or SAN storage devices, servers or appliances, memory cards in nodes, etc.. There is no need to be restricted in the choice of media.
In many cases, the lack of a need for a centralised control room frees up personnel to manage the video system in a more effective way. With a centralised system, the topology can create issues. Systems are generally accessed as a singly entity. Distributed solutions can be accessed as either a single system, or as individual subsystems, dependent upon needs. Management of the system is flexible; the integrator and the customer decide how best to manage it.
Edge recording allows a high degree of flexibility, with the ability to mix and match devices based upon the requirements of specific parts of the system.
Systems can operate without an NVR or servers, or they can take a hybrid approach, allowing regular archiving to a central location when network traffic is low.
In recent years storage costs have fallen. However, it is not only the low cost of storage, but also the variety of ways in which it is packaged, that makes edge recording an attractive proposition.
Importantly, edge recording isn’t about eliminating the server or the NVR! These devices can play an important role in edge recording. Devices can be deployed near a group of cameras and operate within that cluster. Overall control can then be taken via management software.
Edge recording enhances the capabilities and performance of systems, and delivers genuine cost savings.
There will always be cases where centralised recording makes sense. However, by ignoring the possibilities of recording at the edge, integrators and installers could be missing a trick.
Axis Edge Recording
Axis Communications has offered edge recording as a standard feature on its cameras for many years. The company has moved to the smaller form factor microSD cards, and supports both microSDHC and microSDXC formats. Obviously, NAS devices can also be supported via a network connection.
What makes the Axis approach to edge recording somewhat different for small sites looking for a cost-effective security solution is the Axis Companion software. The software is free, and allows a network of up to 16 cameras to be configured, via a laptop or other standard PC. The cameras record onto edge-based SD cards, and if an event occurs the footage can be viewed and downloaded via the PC. When in operation, there is no need for the PC to be connected, and the system can operate without any server or NVR.
Using edge recording with Axis cameras is not limited to Companion-based systems. The recordings can be accessed via the camera, or can be used via a VMS which supports edge-based archiving.
With Axis cameras, memory cards are loaded using the microSD slot (on the rear of the camera with body-type models and inside the housing for dome devices). Once the card is inserted, the recording parameters can be configured via the camera’s menus. Which menus are used depends upon the type of recording required. For event-based recording the Action Rules can be used. This allows a trigger to be set: it can be a motion event, an IVA trigger via an on-board camera application, an input or a device event. The trigger can also be scheduled. It is possible to add additional conditions, thus creating a double-knock type event if necessary.
Axis cameras offer flexibility with regard to edge recording, and work well when accessed via a compatible VMS. Quality isn’t an issue, and recording replays smoothly. You may see an occasional skipped frame when reviewing footage at the highest resolutions using Companion, but most users wouldn’t spot this. If you use a professional Axis VMS, the problem doesn’t arise, either for SD cards or other devices such as a supported NAS or appliance.
Axis has delivered a good range of options for edge recording, and integrators and installers have a choice as to whether the video is accessed via a third party VMS or through Axis’ own range of software.
Bosch Security Edge Recording
The Dinion range of cameras from Bosch are popular devices for mainstream applications and support edge recording via the use of iSCSI devices or microSD memory cards. The cameras support local use of both microSDHC and microSDXC cards.
The connection of iSCSI devices is achieved via the network, while local archiving makes use of a slot on the rear of the camera or inside the dome cover.
Dinion cameras can support two recording tracks; one can be allocated to the iSCSI device and one to local archiving if ANR (Automatic Network Replenishment) is required. ANR ensures that if recording streams are interrupted, the footage can be captured locally until the network connection is re-established. For critical applications, this delivers a fail-safe solution. Both recording options can be password protected.
It is possible to create numerous recording profiles to be used for differing times of day, days of the week or for various shifts or site status changes. Each individual profile allows a video profile to be set for the streams. These include independent configuration of resolution, frame rate and bit rate. The video profiles are set in the encoder menus.
Recording profiles allow metadata to be added or ignored. The type of recording (standard or event-based) is also configurable. Pre- and post-alarm periods for recording can be adjusted.
A recording status page delivers updates and contains information about errors, media, buffer sizes and ANR progress.
The Dinion range gives a good degree of flexibility with regard to edge recording, and allows the functionality to enhance the level of protection on offer. The ANR feature is certainly a benefit.
Accessing edge recordings is simple. The playback options are reminiscent of an NVR and make use of a timeline. Playback controls include play/pause, jump forwards/back in one second increments, fast forward and fast rewind. There are also options to capture a snapshot or to manually control archiving.
When working with edge recordings via a compatible VMS, accessing recordings from the camera is straightforward and footage is replayed as expected. Quality is as configured, and recording replays smoothly without any juddering or skipped frames.
SD Cards in a Nutshell
Many feel that SD cards – the most commonly used cards for on-device edge storage – offer poor reliability. As with many things, when you buy memory cards, you get what you pay for. There are cheap memory cards that will fail in a few weeks if used a lot, but you can also buy cards that are reliable. The latter obviously cost slightly more.
As an example, Benchmark uses a number of high capacity microSDXC cards. They have been used heavily, in a wide range of environments such as temperatures approaching 35°C with high humidity, down to well below freezing. They have been used for data, images and HD video.
MicroSD cards are smaller than the standard SD units and are used by most manufacturers. Currently MicroSDXC cards offer capacities of up to 1TB which, we considered against many mainstream NVRs or appliances, delivers a significant capacity.
When used with video surveillance, devices should support SDHC as a minimum, but preferably SDXC cards in either standard or Micro formats.
While capacity choice is straightforward, it is important to be aware of the card’s speed ratings. SD cards have a class rating which states minimum data transfer rates, and a UHS speed which identifies the minimum rates for writing video.
The speed class shows the card’s minimum transfer rate. A Class 10 card (10MB/s) is required to HD video, while a V30 card (30MB/s) is required for 4K UHD video. There are also V60 and V90 cards (60MB/s and 90MB/s respectively) available.
SD cards have sometimes been questioned with regard to suitability for security applications, but they do offer a solution and with increased capacities, in many applications they will be suitable either for short-term back-up or to create a temporary solution.
While NAS devices or NVRs are more popular where clusters of devices are supported, the potential on offer from SD cards is well worth exploring.