Long Term Storage in a Surveillance Environment

New legal requirements force organizations to increase their data retention of video surveillance for up to a year. This amount of video data requires very large storage capacities, which will drive up the costs for the end user by a large amount. Surveillance data that is old is very rarely accessed, thus there is no need for a short retrieval time for older data. This thesis investigated different storage technologies that could be used in conjunction with a VMS for long term storage of video surveillance. The use case for this thesis was cannabis cultivation facilities in Canada, which are required to store their surveillance video for up to a year. A technical analysis, as well as an economical one, was made for the different systems, and some were tested in practice. Depending on the amount of data and how long it will be stored, different systems can be chosen. It was concluded that the Veracity Coldstore was a good alternative for the use case presented in this thesis.Video surveillance commonly include many cameras recording in HD, and storing the recordings on a digital storage medium, where the total size of all the recordings can become very large. Today, it is common to store video surveillance for 30-90 days. However, recent changes in the data retention laws – meaning how long the data has to be stored – of some countries, like Canada, means that data now has to retained for up to 1 year. This data requires a massive amount of storage space, and this amount of storage can easily become very expensive. Furthermore, for certain industries, like the cannabis cultivation industry, failure to comply to these new retention laws can mean an end to the business as it risks losing its license to cultivate cannabis. Therefore, new solutions for storing large amounts of data had to be investigated. Normally when storing large amounts of data, systems are often configured in something called RAID, which stands for Redundant Array of Independent Disks. RAID is a way for a computer to manage a number of hard drives as one, spreading out information across all of the drives. This increases the speed of the overall system. It also provides redundancy, meaning that the data is stored on several different hard drives; if one drive fails, the data is not lost forever. We investigated one such RAID system, along with other alternative techniques, such as magnetic tape and cloud archival storage. When we had decided which systems were interesting to us, we performed a technical analysis, as well as an economical analysis. With regards to these analyses, the most prominent solution was found to be a device which can spin down disks that are not being used – in effect rendering them inactive, or idle – as it had a relatively low price as well as very good fault tolerance. Furthermore, it also had very low power consumption, which further reduced the cost of the system along with it being a more sustainable solution. However, this spinning up and down of disks takes time – around 20s – and was considered to be a problem. A solution was designed so that new recordings are saved on the local storage; that is, on the computer on which the recordings are created. This is because, according to Axis Communications, newer recordings are much more often retrieved than older ones, and so they need to be immediately available in the majority of cases. Older recordings, then, are moved to the device in question, which acts as archive storage for the recordings