Even in the 1940s, the use of video surveillance showed a tendency to expand in complexity. But today, video surveillance has reached a level of sophistication unimaginable seven decades ago. Today, we can deploy cameras with greater ease than ever before and collect vast quantities of data. Advances in cameras have made collecting data easy, but they present businesses with the new challenge of managing a volume of data they’ve never seen before.
The two most compelling challenges facing video surveillance professionals are both related to data storage: scalability and searchability. As data is created, it needs to be stored somewhere; past technologies like network attached storage (NAS) not only challenged IT managers to guess at their future storage growth rates in order to provision hardware, NAS is a ticking time bomb for organizations that expect to store large amounts of video content, because it has a practical limit of 2 petabytes. Deploying new nodes in a NAS environment becomes exponentially more difficult as the environment grows, and eventually growth hits a wall.
The other challenge lies in finding the right data when it’s needed. Again, the quantity of data stored contributes to the problem: the more data there is, the harder it is to locate the specific piece of data needed right now. Imagine looking for a needle in a haystack, and the haystack keeps getting bigger all the time.
While NAS struggles to handle the demand of modern video surveillance, another technology is offering a solution to both challenges. Object storage can scale limitlessly thanks to its flat file structure, and growing capacity is easy, since the system’s software can automatically incorporate new storage devices into the object storage cluster, dramatically cutting the amount of management overhead the system needs. At the same time, object storage enables metadata to be stored with video surveillance data, enabling fast, easy searching of unstructured data.
Despite its initial uses in research, video surveillance today is found in nearly every vertical market and in a wide variety of applications. It’s used to enhance security provide data for the military, government, public utilities, and vertical markets that include healthcare, retail, hospitality, transportation, education, and more.
Traditional use cases usually centered around Video Management Systems (VMS), which ingest feeds from dispersed cameras that are then viewed in real-time or are retrieved manually through a VMS. But use cases have expanded rapidly, from incident prevention and risk management to the real-time analysis of dynamic environments. Recent innovations in technology enable video data to work with other applications to address real business or administrative problems. For example, pairing video content with AI and facial analysis can allow retail employees to understand their customers even before they approach them. A trial in Japan uses vehicle analysis to deliver messages on dynamic billboards tailored to the demographics of the people who drive various styles of cars. And police and security officers are increasingly wearing body cameras, supplementing dashboard cameras in their police cruisers and cameras in interrogation rooms to use video as an integral part of criminal investigations. Even the food industry has benefited from video surveillance, driving quality assurance by automating the detection of potential mishandling of food that could lead to spoilage or illness.
As the imagination of business leaders is expanding how video surveillance data is used, advancements in camera technology are creating new capabilities and driving new applications. Analog cameras are being replaced with sophisticated digital cameras, taking videos and images with finer resolution. Cameras have grown in density. Today, a camera does not necessarily have just one lens; it may have multiple lenses in a single package, all recording simultaneously. Camera resolution has dramatically increased, allowing small details to be viewed and recorded, driving greater efficiencies at identifying events or incidents.
At the same time, AI and its various companion technologies – artificial neural networks, machine learning and deep learning – have allowed video surveillance systems to become smart visual intelligence platforms capable of triggering actions based on events as well as on patterns and predictive analytics. AI technologies through machine learning and deep learning are being rapidly adopted and integrated into the camera, storage and VMS to help drive real-time analytics.
These new use cases create and are dependent on massive amounts of data that needs to be stored, analyzed and retained over a period of time. Depending on the sector – for example, law enforcement or healthcare – the retention time for data may last from months to years, or perhaps even indefinitely. Also, depending on the event and incident, the data may need to be analyzed in real-time, and the same information may also need to be retrieved efficiently and shared securely with multiple parties. Analyzing video and audio data also requires association with metadata and tagging to make the data searchable and inherently useful for analytics purposes.
How successful an organization is at managing this is largely dependent on how well it manages data storage. In order for enterprises to set up and operate state-of-the-art video surveillance systems they have to address and resolve data challenges that could stop their plans in their tracks now and in the future.
First off, they need a plan to grow. This can mean more than just in terms sheer capacity – it can also mean developing a multi-cloud, multi-tier storage strategy. Multi-cloud refers to the use of more than one cloud providers’ services, usually in addition to on-premises storage. A multi-tier strategy sorts data by frequency of use or other criteria, allowing it to be stored in the right place across the storage infrastructure. For example, video content that is frequently used may be best kept in Tier I storage, on-premises in a system that delivers that data to the customer as quickly as possible. Other data that is less frequently accessed may be stored in the cloud or in storage tiers that have slower retrieval rates.
In order to use a multi-cloud strategy, it’s very likely that object storage forms the basis of the storage infrastructure. Its ability to scale out is unmatched, allowing data storage to grow at the speed of data, not at the pace prognosticated by IT. Its ease of management allows capacity upgrades to be made when the data demands it, not when IT has resources available. And, most importantly, since it allows the use of custom metadata, object storage is far more searchable than other forms of storage, allowing data surveillance devices to record data about the data, and making it possible for managers to add additional data to the data. This makes the data far more easily searched, a huge boon in isolating video for analysis and comparison.
The Cloudian storage solution provides several benefits for storing all the video and audio data generated by modern video surveillance solutions. Cloudian offers a highly scalable, secure, and fault-tolerant platform, along with embedded metadata tags that allow the cameras and VMS system to label the data, enabling managers to quickly search for specific patterns via integrated tools such as Elastic Search.
Consolidation of disparate storage silos with a single scalable storage platform: the solution reduces technical complexity, supports future growth needs and offers flexible deployment and configuration options, including tiering to AWS, Google Cloud Platform, Azure, or other cloud locations.
Performance: the solution can handle the largest environments, with bandwidth for HD and multi-megapixel content.
Petabyte-scalable: the solution can grow to handle extended retention time requirements. Systems can start small and easily expand with non-disruptive linear scaling options to meet data growth needs.
Unmatched data durability: with a choice of erasure coding and/or replication across nodes, datacenters and locations, the solution has no single point of failure, offering high data integrity.
Rich metadata tagging: the solution enables user-defined metadata to be stored with the actual media data. Users can store information such as scene content, sound clip descriptions, or image subject. This allows content creators to capture important attributes about the file and enable rapid media search.
Drop-in Integration: The solution is validated with popular cameras and VMS systems, and easily integrates with existing ecosystems without any modification.
Simplified storage management across all locations, including the cloud: A single-pane system across locations eliminates management overhead and complexity through a unified view of data.
Cost-effective economics: The solution reduces storage infrastructure and operating costs, providing a TCO that's approximately 50% lower than traditional storage systems
Video surveillance is experiencing growth at massive scale across multiple industries. With the increase in camera density and the scale of deployments, immense volumes of video media assets are being created that need to be stored and protected. Cloudian’s video surveillance storage solution offers an ideal platform to accommodate this unprecedented growth based on demand and needs. The solution offers simple, seamless data management across on-premises and public cloud storage to further optimize investment costs. In addition, the integrated advanced intelligence engine allows for data to effortlessly be searched, retrieved, and analyzed without the need for additional computational resources and investments.