It’s easy to associate data protection with firewalls and antivirus programs, but what about the physical security of data? Jennifer Hones, key account manager at Milestone Systems, talks about the myriad of ways that video surveillance technology is safeguarding massive amounts of information in data centers.
If you’ve ever wondered where the files or photos you upload to â€œthe cloudâ€ go, you may venture to guess that this process relies on data centers. Whether near your location or far, data centers store, process and distribute large amounts of data utilizing computing and networking equipment. Because data centers house critical applications and proprietary data, they are vital to the operations of most businesses on the planet, making the physical security of the buildings extremely important. Every major company across all industries relies on data centers to move and remote store its data.
The main components of data centers include routers, switches, firewalls, storage systems, servers, and application delivery controllers. These components must be secured around the clock. Using advanced video surveillance equipment, fencing, drones, and AI, data centers must have the best protocols in place to prevent intruders and unauthorized personnel from entering.Â
Starting from outside the building, data centers typically have a wall of fencing around the property. Microwaves monitor long ranges and send alerts if a person approaches the fence. Video management software (VMS), together with intrusion detection, shock detectors, and climbing detectors, are used to detect potential intruders as they attempt to scale the fence. The VMS is connected to all the cameras in and around the data center and is programmed with license plate recognition (LPR) to identify authorized and unauthorized vehicles coming close to the property.
Assuming a person is able to scale the fence, geofencing with radar is activated. A geofence uses location data via GPS signals that set up an invisible barrier around the outside of the data center. Now the person is automatically tracked at every step with LiDAR. 3D images of LiDAR data are sent to the VMS to process, track, and classify the person entering. This, in turn, activates the alarm system.Â
If the threat is not human but is environmental, seismic detectors come into play. Or perhaps the person attempting to enter is doing so during challenging weather conditions, such as during a snowstorm. Thermal cameras with long detection ranges and high-contrast images perform well even in low-visibility conditions, such as when there is light rain, fog, smoke, or total darkness.Â
Once a person approaches any of the entry points to the data center, more technologies come into play. Visitor and vendor management is accomplished by activating access control, sometimes touchless. Access control denies entry to anyone except approved employees or visitors. Access control technologies may include fingerprint scanners, card readers, coded entry, among others. This may be enforced by facial recognition software and video analytics that identify who should or shouldn’t be on the premises. Some data centers are also utilizing iris/eye scanners to further ensure the accuracy of facial recognition software. Facial redaction blurs out people’s faces in the VMS to protect privacy while visitors and employees are on-premises. Facial redaction can only be lifted by approved administrators of the VMS.Â
Once inside the data center, authorized personnel goes through several layers of security to reach the highly protected server racks. Motion-activated cameras with built-in analytics should be stationed throughout the facility, with the capability of following an individual throughout their entire journey in the building as they move from room to room. At each room’s entry and exit point, access control either denies or accepts entry into that room ensuring only verified employees or visitors enter.Â
Warnings & Alerts
Any room in a data center can be made inaccessible by shutting down access when a sensor is triggered. When a sensor is activated, an alarm will alert the security team that someone is on-premises that shouldn’t be.Â
Hundreds of cameras with built-in artificial intelligence (AI) strategically placed throughout the premises can detect anomalies. For example, if someone is standing in a hallway too long, AI will send an alert through the VMS to security personnel. In certain high-profile locations and/or at certain times, drones are programmed to patrol the hallways and send an alert if anyone should enter the premises that should not be there. Lastly, far-reaching scanners are placed on top of the data racks themselves which are typically housed in cages to see if anyone is crawling on top of one of the appliances.
Cages, where server racks are stored, should require extensive authorization to enter. If the scanner detects an unauthorized person requesting access to a cage, a sensor goes off and automatically alerts the security team with video footage of the intruder. Sensors can even detect if a rack door is left open or accessed unexpectedly, triggering either pre-recorded or live audio and dispatching security.Â
The VMS can also be programmed to send alerts if someone is walking the wrong way down a hallway, standing in one place for too long, or if there is a fire or smoke.
Ultimately, intelligent video surveillance along with an integrated security program and plan provides thorough protection to data centers. Video surveillance can provide smart, high resolution and even thermal imaging detection when needed. As data center security technologies continue to evolve, new physical security measures will surely be incorporated.