By now, this should be clear to anyone who manages any IT infrastructure for an enterprise. Yet, surprisingly, it has taken many IT leaders longer than you’d expect to focus on insider threats instead of the enemy at the gates.
This old idea of building a moat around a castle to keep intruders at bay is no longer applicable today. Indeed, focusing on it is missing the big picture, if you ask any cybersecurity expert worth his salt.
Even the Singapore government, which unfortunately started cutting off public servants’ Internet access in 2016 in an attempt to prevent serious cyberattacks, has come around to the reality that the threat has to be managed in a different way.
After all, some of the most serious data breaches of late have occurred because of human error. When the blood bank accidentally leaked the personal data of donors earlier this year, the database was inadvertently placed in the open by a third-party vendor running tests.
Another serious breach, which was revealed this year as well, involved a doctor who allegedly took the information of thousands of HIV patients here and simply downloaded the data onto a USB drive.
That’s a classic inside job, much like someone in a castle opening a backdoor to the enemy. No matter how strong your fortifications are, the perimeter defences are useless against such an attack.
Security experts are now calling for a policy to trust no one within the network. In other words, every user is viewed with the same suspicion even though they might have been granted access.
“Increasingly, customers are acknowledging that more attacks are coming from inside the network,” said Sudhakar Ramakrishna, chief executive officer of cybersecurity firm Pulse Secure, in an interview last week with Techgoondu.
“Who do you trust? Don’t trust anybody,” he added. “And don’t have artificial separation of what’s outside and what’s inside. Establish trust equally with everybody.”
In such a setup, when a user fires up his laptop to connect to a laptop, he has to be authenticated, possibly with multi-factor authentication. Then the way he uses his connection is constantly monitored so that any unexpected or suspicious behaviour can be flagged.
If someone in marketing tries to access files in the finance department’s file server, the system should be able to raise a red flag to alert an administrator. He can choose to investigate further from there.
Another problem with the old idea of maintaining a perimeter is that the perimeter is no longer clearly marked any more. With data moving in and out of multiple cloud platforms and being replicated and backed up, it is becoming impossible to draw a line and say “it’s safe over here”.
The key, say cybersecurity experts, is to make sure that everyone’s identity is verified and checked all the time, so they are who they say they are. This ensures that only those who should have access are able to access the data.
Of course, this still doesn’t guard against people stealing data if they are disgruntled with the company. Or, they may have clicked on a malware-loaded e-mail and opened themselves up to the bad guys.
This is where artificial intelligence (AI) comes in, to continuously scan for unusual behaviour after “learning” what is usual in the network. A person copying sensitive files and dropping them onto Dropbox afterwards should be looked into, for example.
The complexity of running an enterprise’s infrastructure today means that it is often not humanly possible to look through all the alerts that a system triggers. Machines have to help analyse that data.
Many companies have servers connected on their networks that they don’t even know of, said Sanjay Aurora, managing director for Asia-Pacific for Darktrace, a cybersecurity company.
As a result, these systems may not be patched up and may be open to hackers as backdoors, he explained.
As threats become more complex, the idea of cybersecurity has to change as well. Being a czar and trying to control everything may backfire, because staff will find ways around the restrictions.
Organisations have to look to role-based access control, rather than cutting off access in a broad manner, said Ramakrishna, who also called on cybersecurity teams to improve productivity, maintain visibility of assets and support compliance requirements.
However, they should forget the “indefensible walled garden”, he added. Instead, get users on their side through education and awareness of the latest threats, he advised.
“Security is about access, not just control,” he added. “You can only enhance things if you make things easily accessible. How to do that without compromising security or trying to lock everything down?”