The trend of making objects smart by adding computing and communication capabilities is gaining momentum. A prediction made just two years ago that there would be 20 billion IoT devices on the planet by 2020 seems now to be a gross underestimate as everything from consumer devices to shipping containers is becoming an IoT device. Each of these devices is an endpoint that can act as a gateway to other systems and services connected through the network. Edge computing increases each node’s computing abilities, adding more processing power and storage to the point edge devices can perform tasks that were previously the responsibility of the servers they connect to. Edge computing is making increased use of IoT devices because distributing computing tasks at the source of the data reduces the load on centralized computing resources and speeds analytics. But the magnitude of IoT population calls for automated management. Protecting four layers of the IoT network architecture is critical because edge computing each requires its own attention as the stakes increase with the number of nodes.
Endpoint devices (IoT)
451 Research reports that attacks on IoT devices were ranked as top-level concerns by enterprises. Specifically, they pointed to poor authentication of IoT endpoints and how users access their IoT devices and applications. According to 451 Research, “As industrial equipment is increasingly connected to the Internet for data collection and analysis, enterprises open themselves to the sophisticated world of security intrusions.” CIOs need to deploy automated tools at least as sophisticated as those used by would-be attackers to protect the endpoints of their networks.
Physical network layer
IoT devices are typically connected to their networks wirelessly so it may seem irrelevant to consider protecting the physical layer of an IoT network. But even without wires the physical components still represent potential access points that can be breached. IoT devices come in all variety of configurations, and the fact that they typically rely on wireless networking doesn’t mean the device lacks other kinds of ports. Legacy manufacturing devices may have RS-232 ports for diagnostics while newer devices are more likely to have USB ports to allow local maintenance. US Cyber Security reports that 49% of companies are concerned about malicious insiders - individuals who could potentially gain physical access to any number of IoT devices and gain access to connected systems. IT needs strong controls over device configurations either through physical design characteristics or advanced security that can monitor and restrict access to physical ports.
In addition to the devices and the cabling connecting various parts of the network, IT needs to protect the platform. While this is well understood as part of traditional computing protection, it becomes more difficult and at the same time more important when IoT is joined to the infrastructure. Platform security specifically designed with a deep understanding of IoT devices should include authentication and authorization services as well as encryption so that rogue or compromised devices are discovered and their access restricted, protecting other segments of the enterprise.
Applications drive IoT deployments and applications can have exploitable vulnerabilities. In this way, IoT devices are no different from any other computing devices in that would-be intruders look for existing flaws they can use to gain access to the data, the device itself, and to the connected network and its assets. IT needs to consider IoT implementations with the same level of concern applied to any other computers attached to the network and assume that intruders will employ whatever methods available to gain access to them. This represents an additional risk above more traditional computers if only because the population of IoT devices is likely to be larger by orders of magnitude and infection of one node could quickly spread to all similar devices increase the risk to the overall network.
The challenge of protecting the quickly expanding use of edge computing in enterprise networks is substantial and requires dedicated attention as IoT deployments present intruders with additional targets and ways to access enterprise systems.