When it comes to OT (Operational Technology) security in all its facets, security people from the OT world and IT security professionals quickly can end up in a situation of strong disagreement. Depending on the language they are talking, it might even appear that they seem being divided by a common language. While the discussion in English quickly will end up with a perceived dichotomy between security and safety, e.g. in German it would be “Sicherheit vs. Sicherheit”.
The reason for that is that OT thinking traditionally – and for good reason – is about safety of humans, machines, etc. Other major requirements include availability and reliability. If the assembly line stops, this can quickly become expensive. If reliability issues cause faulty products, it also can cost vast amounts of money.
On the other hand, the common IT security thinking is around security – protecting systems and information and enforcing the CIA – confidentiality, integrity, and availability. Notably, even the perception of the common requirement of availability is slightly different, with IT primarily being interested in not losing data while OT looking for always up. Yes, IT also frequently has requirements such as 99.9% availability. However, sometimes this is unfounded requirement. While it really costs money if your assembly line is out of service, the impact of HR not working for a business day is pretty low.
While IT is keen on patching systems to fix known security issues, OT in tendency is keen on enforcing reliability and, in consequence, availability and security. From that perspective, updates, patches, or even new hardware and software versions are a risk. That is the reason for OT frequently relying on rather old hardware and software. Furthermore, depending on the type of production, maintenance windows might be rare. In areas with continuous production, there is no way of quickly patching and “rebooting”.
Unfortunately, with smart manufacturing and the increased integration of OT environments with IT, the risk exposure is changing. Furthermore, OT environments for quite a long time have become attack targets. Information about such systems is widely available, for instance using the Shodan search engine. The problem: The longer software remains unpatched, the bigger the risk. Simply said: The former concept of focusing purely on safety (and reliability and availability) no longer works in connected OT. On the other hand, the IT thinking also does not work. Many of us have experienced problems and downtimes to due erroneous patches.
There is no simple answer, aside that OT and IT must work hand in hand. It’s, cynically said, not about “death by patch vs. death by attacker”, but about avoiding death at all. From my perspective, the CISO must be responsible for both OT and IT – split responsibilities, ignorance, and stubbornness do not help us in mitigating risks. Layered security, virtualizing existing OT and exposing it as standardized devices with standardized interfaces appears being a valid approach, potentially leading the way towards SDOT (Software-defined OT). Aside of that, providers of OT must rethink their approaches, enabling updates even with small maintenance windows or at runtime, while enforcing stable and reliable environments. Not easy to do, but a premise when moving towards smart manufacturing or Industry 4.0.
One thing to me is clear: Both parties can learn from each other – to the benefit of all.