Patching Problems: How Do CNI Firms Protect Legacy Systems?Posted: November 28, 2017 Filed under: Uncategorized | Tags: air-gapping, Bad Rabbit, CNI, critical national infrastructure, european commission, IDS, legacy systems, NIS directive, notpetya, patching, unpatchable systems, virtual patches, wannacry Leave a comment
It might not have escaped your notice that critical infrastructure (CNI) organisations are increasingly being probed by nation state hackers. Traditionally, IT managers in these environments might have relied on “security-by-obscurity” to get by. But with many now connected to the public internet and running Windows systems, that defence is no longer valid. One of the main challenges appears to be mitigating risk on legacy systems which can no longer be patched.
I’ve been speaking to experts in this area for an upcoming feature and thought I’d share their best practice advice.
CNI under attack
There’s certainly no doubt that CNI firms are coming under attack far more frequently than they used to. Witness the alleged North Korean WannaCry attacks, which hit 34% of NHS Trusts and nearly 600 GP practices, leading to cancellation of an estimated 19,000 operations and appointments. Or arguably more sophisticated attacks – think China and Russia – designed to carry out reconnaissance work on key systems, or even in some cases disrupt power supplies, causing widespread black outs.
Recognising the uptick in threats, the European Commission’s NIS Directive will look to enforce a minimum standard of security across providers of “essential services” in the UK and Europe. However, from what I’ve heard, there are varying levels of awareness about the new law, set to come into force on 9 May 2018.
“Yes, awareness and activity has been on the rise, but generally still behind where it should be I think,” KPMG cybersecurity practice partner, Martijn Verbree, told me. “Many organisations haven’t realised to what extend this applies to them and the impact.”
So how exactly do CNI firms keep legacy, unpatchable, systems secure — appeasing regulators in the process? Most experts I spoke to pointed to SCADA and ICS systems as exposing organisations to most risk. Interestingly, air-gapping isn’t necessarily going to work, according to Michail Maniatakos, assistant professor at the NYU Tandon School of Engineering.
“Given the rapid increase in the numbers of mobile computation devices, air-gapping has proven to be an illusion as individuals enter air-gapped locations using their laptops, smartphones, smartwatches etc. There are documented cases of USB drives breaching the air-gap, ie Stuxnet,” he told me by email.
“The most appealing option is layered security, and the assumption that every layer can be breached. The advantage of this method is that even if a layer is breached, the window of opportunity for the attacker will be limited to go through all the layers of security — similar to the security offered by multi-factor authentication. Needless to say, this approach needs also advanced intrusion detection capabilities as well, in order to quickly understand whether an adversary is in your control network.”
BeyondTrust VP of technology, Morey Haber, explained that layered security should include ACLs and port filtering, which would have protected against threats such as WannaCry and Bad Rabbit.
“If networks are properly zoned, malware or ransomware that leverages exploits can’t attack additional assets if the ports and IP ranges are blocked against lateral movement from adjacent resources or untrusted zones,” he told me. “This mitigates the threat temporarily while patches are being deployed but is good practice to block an infection in the first place.”
Another key technology to consider is continuous network monitoring, baselining normal behaviour and then alerting when suspicious activity is spotted.
“This includes looking for command and control using DNS or communications obscure IP addresses or unapproved resources via lateral movement,” continued Haber. “It is not enough to monitor TCP/IP traffic alone but also all the supporting services from NTP or DNS that can be used in a modern attack. Deviations in these patterns may allude to a growing or persistent threat.”
He also recommended removing all admin rights — citing a Microsoft estimate from last year that 84% of vulnerabilities can be mitigated by doing exactly this. Risk exposure can be further minimised by turning off all unnecessary services, ports, and features, he added.
Advanced security controls are also vital, such as network and web firewalls, IDS/IPS.
“While these may be integrated technologies, when was the last time they were upgrade to the latest firmware, pen test for best practice rules, or planned for replacement due to end of life?” said Haber.
“While every security professional would agree the perimeter has dissolved due to the cloud and Internet, local area network resources still need to be protected. Using the latest technology and verifying your devices are not obsolete is key to defend against attacks and stopping modern threats like ransomware.”
In addition, there’s always the option of paying for extended support — a deal the NHS had with Microsoft until 2015, for example. This might be expensive, but with some firms like FedEx and Maersk claiming NotPetya cost them hundreds of millions of dollars, it might not be such a bad investment.