They worked great for enterprise clients with in-house global security operations centers (GSOCs), but for the end-user with one to a handful of sites, with no onsite security monitoring, these appliances were the wrong technology for the use-case. These end-users needed smaller appliances with a standard set of analytics, such as analytics to determine if correct PPE was being used on a job site, or for multifamily clients, analytics to determine if someone in uniform was approaching to drop off packages.
Edge AI Changes the Game
The first edge AI appliance was introduced by Vaido (then IronYun) around 2018. At the time, the company offered the larger analytics appliances for enterprise use, but came out with an edge appliance that could be used for smaller deployments or co-located on the edge with remote sensors to provide analytics packages to edge cameras. While innovative, the IronYun product took some time for adoption; after all, sometimes trends start slowly.
However, over the past few years, especially as the cost of bandwidth and cloud connectivity has dropped, the cost of edge AI chipsets like Nvidia and Hailo has come down, allowing the cost of adding edge analytics to an edge appliance to be much more reasonable.
Today, the idea of the edge AI appliance has gained ground. Vaidio is still there, but they have been joined by manufacturers like Camect and Hanwha Vision – all of which have developed an edge AI appliance to bring analytics to the edge without a specific camera requirement. The three here are listed only in the order provided above and by no specific recommendation.
How Edge AI Appliances Have Impacted Security Monitoring
The edge AI appliance has become one of the monitoring companies’ best technological resources, especially for monitoring clients with multiple locations – typically 24 cameras or fewer – with a mixture of camera solutions across the portfolio.
The edge AI appliance provides set analytics for any camera: AI-enabled, analytic-enabled, or without analytics. The three manufacturers listed above also provide vertical-specific analytics such as multifamily, retail, healthcare, car lot, and more.
The edge AI appliance has worked in two ways:
1. Pre-Set Analytic Packages – The client can standardize on a bolt analytic package that provides them a set type of analytics that does not depend at all on the camera vendor or the type of analytics the cameras may provide by default. This has enabled some customers to gain the advantages of camera analytics while still using older analog cameras, though the camera stream must be converted. Still, if analytics are desired, and the cost of replacing analog cameras is outside the budget, an edge AI appliance may be a cost-effective solution.
2. Standardized Edge Appliances – Monitoring companies are requiring their clients to implement a standardized edge AI appliance so that all cameras being monitored by the company are compliant with the monitoring company’s standard. This helps reduce operator training and helps the operators build muscle memory on the types of alerts provided by the analytics. The edge AI appliance also reduces the costs involved with manually reviewing video feeds with no activity.
A number of monitoring companies have built direct integrations into their edge AI appliance of choice, allowing the alerts to come in natively into the monitoring company’s software platform(s).
Technology Gains Footing
I frequently say, “The best technology in the world in the wrong use-case is the wrong technology.” Identify the use-case, then find the tech that solves the problem.
The edge AI appliance may be the perfect use-case for end-users to enable their monitoring provider to be proactive, respond faster, and reduce false alarms. While the technology is really just getting started, it is already a growing tech trend for monitoring companies. I would look for at least one or two more edge AI appliances to come to the market in the short term, offering on-premise and cloud-based analytic processing solutions.
