BY William Pao
In the age of the internet of things (IoT) and artificial intelligence (AI), devices, including IP cameras, are generating more data than ever. Deciding how to properly store this data is critical. Depending on a user’s needs, several approaches are available. These include storage on the edge, the cloud, or a combination of both. As a leading storage solutions provider, Seagate Technology has solutions for all of the aforementioned architectures.
ecent findings by IDC predict the cumulative amount of data worldwide will grow from 33 zettabytes (ZB) (33 trillion gigabytes) in 2018 to a whopping 175 ZB by 2025. A lot of this data is generated by IP cameras for both security and business-intelligence purposes. Consequently, storing this data effectively and maintaining data integrity becomes crucial. “With the advent of ultra-high definition cameras and artificial intelligence features, surveillance systems are amassing more data and storing it longer than ever before,” said Robert Yang, VP of Asia-Pacific Sales at Seagate Technology. “Gone are the days when systems integrators can simply ignore the storage aspect of video surveillance systems.”
Storage Strategy: Edge vs. Cloud
According to Yang, users should ask themselves whether they want to go with edge or cloud storage, or a combination of both. “The edge refers to a device or appliance that stores captured data closest to where the footage is captured, while also analyzing that data in real-time, before transferring the data to the backend. The analysis is typically conducted in deployment or detection mode where the learning is already done and coded in chip. However, analysis can also be done through basic learning capabilities because data sets are smaller at the edge. Heavy compute capability is expensive and not feasible in large numbers on the edge, hence light compute with either CPU or low-cost GPU,” Yang said. “Meanwhile, the cloud is the large centralized backend storage server or IP-SAN where pools of data are aggregated and stored. Analysis of unstructured datasets, including multiple sources of data not limited to surveillance, occurs in the cloud. Deep learning and big data analysis are also conducted in
the backend often via more powerful GPUs.” According to Yang, when determining data storage needs, internet speed is another factor to consider. “ The speed of transfer depends heavily upon the internet connection and distance between the camera sensor and server. As a result, uploading files or streaming them may take some time. “This, coupled with the potential of monthly subscription fees for public cloud accessibility, can be enough to make end users think twice about cloud-based options.”
The architecture that a user chooses ultimately depends on their own case scenario. Regardless the strategy they adopt — edge, cloud or both — Seagate has the solution for it. For those opting for edge storage, the user can benefit from surveillanceoptimized hard drives. These can
Edge or cloud, Seagate has your storage needs covered