Eye on Hosted Video: How to Keep Recording if the Network Fails

Aug. 31, 2011
Maintain critical surveillance coverage with NAS and SD memory cards

Network failures are pretty uncommon these days, thanks to ISP innovations, network technical service and redundancy plans. Yet there can still be times when your network might be unavailable, like if a local switch or router crashes or someone accidentally unplugs a cable. And even in a hosted video application, video management systems (VMS) might be taken offline periodically for server upgrades and scheduled maintenance. This has led to a popular hosted video myth: If my network goes down, I lose my surveillance capabilities.
To avoid losing critical surveillance coverage during potential outage intervals, consider installing an onsite backup system to cache the video locally.
But wait a minute — isn’t one main selling point of a hosted video system that you don’t need onsite storage? Aren’t we getting rid of the DVR with hosted video? Yes, hosted video does replace the need for expensive, high-maintenance DVRs for onsite storage; however, if you cannot afford even a brief lapse in your hosted recording, there are two common off-the-shelf and inexpensive local storage options: Network Attached Storage (NAS) drives and SD memory cards.

NAS: Backup and Recovery on Demand
The NAS device installs on the same local area network (LAN) as the network cameras. The cameras can be programmed to stream video to the cloud and the NAS simultaneously 24/7, or just to the NAS if the connection to the cloud is lost. Once the connection is re-established, the NAS can upload its stored video to the cloud automatically, or be offloaded on either a scheduled or manual process, depending on how you have programmed your surveillance system.
Since the NAS is attached to the LAN, you can also view NAS-stored video from any desktop, laptop or mobile device that has access to the network. It is transparent to the user whether the video came from the NAS or the cloud.
A NAS device is easy to install because it can be registered to the hosted video service portal with a single click. It is also a cost-effective backup solution since multiple cameras can share the same storage device. The NAS can be programmed to record multiple streams at different frame rates and resolutions, giving end-users both a redundant safety net for storage and the freedom to configure the cameras in the system for various performance levels and image quality (including HDTV and megapixel) depending on a particular application and its security needs.
A high-quality NAS unit typically costs around $250 for two terabytes of storage, which is more than enough for up to 10 cameras to share.

SD/SDHC/SDXC Cards: An Alternate Video Backup Plan
Another backup option is the SD memory card. An SD (secure digital) memory card — whether SD, SDHC or SDXC format — is the same type of card found in digital photography cameras and Smartphones to store pictures, music and video. This solution is commonly referred to as “edge storage” because video storage happens inside the camera — at the edge of the system. The standards for memory card storage are set in the consumer electronics industry by the SD Association.
SD cards are a great solution for sites with one or only a few cameras. But the cost per gigabyte of storage is higher than with an NAS device.
There are two basic ways to use SD cards for video back up. One is to program the camera for continuous recording to the SD card regardless of whether the hosted video application is accessing it or not. In this scenario, the VMS must identify that it has a gap in its video reception and request the missing clip from the camera so that it can be merged into the recording database.
The other way is to instruct the camera to buffer video on the memory card whenever a communication session with the cloud is abruptly terminated instead of properly closed. Once communication is restored, the hosted video application in the cloud can retrieve the missing video, restore fail-over recording and delete the video on the SD card. If the hosted video system does not offload the SD card before it reaches full capacity, the camera will simply stop recording to the memory card.
While NAS recording currently has a strong communication link and automatic handshake with hosting systems today, be forewarned that activating an SD card failover option may require custom configuration by an integrator. As edge recording continues to improve, expect to see advanced filtering and analytic options in the not-too-distant future.

Choosing the Right Amount of Local Storage
Whether you opt for a NAS device or edge-recording, the amount of storage needed will entirely depend or the application and company’s security needs. If you plan to use the local storage strictly as a redundant safety net for network failures, then less is more; however, a NAS offers an attractive bonus: dual-streaming. This gives you the option of sending high-quality video to the local device for forensic evidence while saving on bandwidth by streaming lower-quality video to the cloud.
While 2TB NAS models start around $250, you could scale up to 12TB models for around $800 today. By the same token, SD memory cards come in a wide range of storage capacities and prices, which a quick Google search will prove. Basic SD cards range from 64MB to 4GB. SDHC (high capacity) cards offer more storage — from 2GB to 32GB. SDXC cards — the newest SD Association standard — start at 64GB and can go as high as 2TB, though the upper limit cards may take years to reach the market.
Obviously there are plenty of data size options to choose from — albeit NAS or SD — so your decision depends on a number of factors:
1. Which device/card your hosted surveillance system supports;
2. Which compression standard you use (H.264 requires 50 percent less storage than MPEG-4);
3. Your image resolution (higher the resolution, more storage); and
4. The frames per second you wish to record (more frames, more storage).
The following table approximates the storage capacity needed per day based on 24-hour continuous recording with H.264 compression in a scene with medium activity and complexity. Note that storage will vary based on complexity of scene.

Resolution Frame Rate Storage/day*
VGA 15 fps 3.5GB
VGA 30 fps (full) 6GB
HDTV 720p 10 fps 8.3GB
HDTV 720p 30 fps 18.1GB


Unless mandated by corporate, casino or government policies for video retention, recording continuously for 24 hours a day — especially at full frame rate HDTV video — would be considered highly unusual and frankly a bit overkill. In a typical hosted video scenario, the cameras would be programmed to record to the onsite device at different frame rates and resolutions depending on motion detection, alarm events, certain hours of the day and if network connection is lost. In other words, the above scenario is the maximum you would need per camera per day.
A more realistic hosted video onsite redundancy scenario, for example, would be based on motion detection recording and would look something like this:

Resolution Frame Rate Storage/day*
VGA 15 fps 0.7GB
VGA 30 fps (full) 1.2GB
HDTV 720p 10 fps 1.7GB
HDTV 720p 30 fps 3.6GB

*Both charts were calculated using Axis Design Tool, version 2, based on 30% compression and medium scene activity. For the second scenario, motion detection is assumed at 20%.

If you only plan to leverage onsite storage to protect against potential network failure, then obviously there are plenty of inexpensive storage options you can choose.

Deciding if a Local Surveillance Safety Net is Necessary
If your surveillance system runs on a local area network, the backbone is generally very reliable. With such high uptime, SD cards and NAS devices might be more redundancy than you need — much like wearing suspenders and a belt. But if you are operating in a very demanding application environment or have times of the day (e.g., cash counting shifts) when you cannot afford to lose video, then redundant storage is an absolute necessity. In fact, there are some hosted video integrators today who will not sell a system to a client without some type of backup storage. To them, there’s too much liability at risk.
If you do deploy an onsite storage option, it might make sense to choose a NAS and improve your overall surveillance system by storing crisp, higher resolution video locally. In addition to failover redundancy, you can save on network bandwidth consumption by storing higher frame rate, higher resolution video on the NAS while dual-streaming lower frame rates and lower resolutions to the cloud.
As with other storage devices on the market, SD memory cards and NAS devices continue to improve in speed and capacity while economies of scale are turning the technologies into commodity products. With prices continuing to drop, edge storage and NAS devices are becoming game changers in the world of hosted video.

Fredrik Nilsson is General Manager of the Americas for Axis Communications and author of the book Intelligent Network Video. He is a regular expert contributor on topics of networked video surveillance systems and cameras.