Edge Computing Will Change Everything – for Companies, and for Hackers
Internet of Things networks coordinate from few to hundreds or thousands of devices, sending data between their sensors and connected data centers or clouds. Given the massive amount of raw-data that scaled systems must transmit and process, IoT traffic can increase latency, crowd bandwidth, and compromise security when it is sent un-processed to a cloud server or from a central database to a network of devices. In order to meet these challenges, edge computing creates a processing perimeter at the “edge” of the network where logic and analysis can be performed in real time prior to data exchange with core systems.
Research firm IDC defines edge computing as a “mesh network of micro data centers that process or store critical data locally and push all received data to a central data center or cloud storage repository, in a footprint of less than 100 square feet.” Edge processing reduces the amount of traffic sent to and from devices and can significantly reduce data exchange times in critical applications such as healthcare or manufacturing. Edge computing can also increase the security of a local network, but it is subject to its own security vulnerabilities, most related to issues of remote access and management.
IoT Vulnerabilities at the Edge
While edge computing solves a number of data traffic issues in IoT networking, it can also present new vulnerabilities, resulting in a larger overall attack surface (sum of a system’s access points a potential attacker can leverage). The state of existing platforms makes networks more vulnerable at both the edge and at the user endpoint. Many attacks can occur, for example, because end-users do not, or cannot, change default passwords, letting bad actors gain access to the devices, exposing edge devices to attack as well.
Unsecured internet resources can be found and accessed easily. (There is even a search engine, launched in 2013, for finding unsecured IoT devices around the world.) In a 2017 “botnet barrage,” bots were deployed to search for devices running default passwords on a university campus. 5,000 IoT devices were attacked by 5,000 discrete systems trying to break through weak or default passwords.
While the vulnerability in attacks like these lies at the endpoint, edge computing complicates things, introducing potential new attack surfaces. IoT devices that connect to the public internet compromise security protocols at the edge. This is partly due to the current state of edge computing, in which full stack solutions that include sensors, software, and secure elements are rare. Many methods used to secure IoT networks at the edge can be insufficient. LPWAN protocols can become vulnerable if encryption keys are compromised. VPNs are subject to man-in-the-middle attacks.
The distributed nature of edge computing also opens a new, unwelcome frontier of physical risks. While the servers and devices that power traditional networks are usually housed in dedicated, often highly secure facilities, the same tiny data centers that make edge processing such a huge leap forward can also be a security nightmare.
Instead of living in data centers, these micro-centers devices are often deployed in the field which, when we’re discussing IoT edge, can be a corporate office, a farm, and anything in between. An attacker physically tampering with an edge device can bring down a network, or even harm one of its operators. Securing these devices is also far from trivial – while hardening them against physical attack is necessary, it’s also a tradeoff between security, cost, and the ease of upgrading and servicing edge data centers. Device makers also need to be aware of the risks, and make sure that devices can be easily secured to generate remote and local alarms on any sign of tampering.
Operator Error: the IoT Learning Curve
By far, the largest threat to IoT edge networks comes from installers, operators and other human actors who aren’t familiar with the rapidly shifting landscape of IoT cybersecurity. IoT devices outside corporate and industrial sites are often looked upon as toys rather than serious security risks and can be compromised if attackers have physical access to them. Devices need to be secured, both physically and with strong passwords, and they need to be updated regularly, procedures that many ordinary users may not regularly follow.
Best Practices for IoT on the Edge
Despite the potential for increased vulnerabilities, edge processing can be a safe way to manage IoT data traffic if devices are secure and proper authentication systems in place. The IEEE (Institute of Electrical and Electronics Engineers) recommends using edge computing for greater security, since edge software can verify the identity of IoT devices and prevent malicious devices from accessing the core network.
Some best practices for ensuring security at the edge of the network include implementing end-to-end encryption; including a long-term plan for an edge computing deployment that accounts for future use cases and the compatibility of components and software; and securing devices with strong passwords and encryption keys; and, if applicable, with biometric authentication systems. Devices connected to the edge should not run unnecessary services and connections to the public internet should be minimized or eliminated where possible.
Ultimately, most organizations should turn to partners with the expertise to provide a monolithic, rather than a “bolted-together” solution for an edge buildout. Beyond the right devices and software analysis tools, edge computing solutions require physical security and controlled physical environments. Organizations that partner with experts can better overcome the complex security challenges and reap the benefits of streamlined edge processing.