Even in the planning stages of a deployment, IoT security is one of the chief stumbling blocks to successful adoption of the technology.\nAnd while the problem is vastly complicated, there are three key angles to think about when laying out how IoT sensors will be deployed in any given setup: How secure are the device themselves, how many are there and can they receive security patches.\nPhysical access\nPhysical access is an important but, generally, straightforward consideration for traditional IT security. Data centers can be carefully secured, and routers and switches are often located in places where they\u2019re either difficult to fiddle with discreetly or difficult to access in the first place.\nWhere IoT is concerned, however, best security practices aren\u2019t as fleshed out. Some types of IoT implementation could be relatively simple to secure \u2013 a bad actor could find it comparatively difficult to tinker with a piece of complex diagnostic equipment in a well-secured hospital, or a big piece of sophisticated robotic manufacturing equipment on an access-controlled factory floor. Compromises can happen, certainly, but a bad actor trying to get into a secure area is still a well-understood security threat.\nBy contrast, smart city equipment scattered across a metropolis \u2013 traffic cameras, smart parking meters, noise sensors and the like \u2013 is readily accessible by the general public, to say nothing of anybody able to look convincing in a hard hat and hazard vest. The same issue applies to soil sensors in rural areas and any other technology deployed to a sufficiently remote location.\nThe solutions to this problem vary. Cases and enclosures could deter some attackers, but they might not be practical in some instances. The same goes for video surveillance of the devices, which could become a target itself. The IoT Security Foundation recommends disabling all ports on a device that aren\u2019t strictly necessary for it perform its function, implementing tamper-proofing on circuit boards, and even embedding those circuits entirely in resin.\nDiscovery and networking\nSecuring the connections between IoT sensors and the backend is arguably the toughest part to solve, in part because an alarming number of organizations aren\u2019t even aware of all the devices on their network at any given time. Hence, device discovery remains a critically important part of network security for IoT.\nThe main reason for this lack of visibility is that the nature of IoT as an operational technology, rather than one that\u2019s solely administered by IT staff, means that line-of-business personnel will sometimes connect helpful devices to the network without telling the people in charge of keeping the network secure. For network operations people, used to having a clear sense of the entire network\u2019s topology, this can be an unaccustomed headache.\nBeyond IT personnel working closely with the operational side of the business to ensure all devices connected to the network are properly provisioned and monitored, network scanners can discover connected devices on a network automatically, whether that\u2019s via network traffic analysis, device profiles, whitelists or other techniques.\nSoftware patching\nMany IoT sensors don\u2019t have a lot of built-in computing capability, so some of those devices aren\u2019t able to run a security-software agent nor accept updates and patches remotely.\nThat is a huge worry, because there are software vulnerabilities being discovered every day that target the IoT. An inability to patch those holes when they\u2019re discovered is a serious problem.\nMoreover, certain devices simply won\u2019t be able to be properly secured and made patchable. The only solution might be to find a different product that accomplishes the functional task yet has better security.