Belfast Telegraph

Software bugs inevitable, but they needn't threaten security

By Rahim Jina

It's all just software. Really, it is. Software runs our applications and software makes our hardware work. Software has always had bugs, due to the fundamental fact that humans, unlike computers, are not finite state machines which follow predictable paths.

The human brain is not wired like this, which means that as long as humans write code, code will always contain bugs. This raises the question for the future of artificial intelligence, since it is humans that will write the code that creates the intelligence, but that is a discussion for another day.

Where software runs does make some difference, but not as much as you might think. It's funny when an industry can create new markets for itself, just by reusing existing concepts and repackaging them. That's what cloud is. Cloud is the new mainframe, back when the paradigm was for computing to be performed centrally and then accessed remotely by a dedicated terminal - sound familiar? The cloud is simply just another bunch of big computers sitting somewhere else, where that someone else is paying the electricity bill instead of you (directly, anyway).

In fact, cybersecurity has not changed much in the past 15 to 20 years.

It has been rebranded often but the fundamentals are still there unchanged. Security vulnerabilities which give rise to hacking incidents and breaches can still be largely grouped into three categories: bugs in software (badly written code), misconfigured software (where protections are there but someone did not set them up right), and resource availability (someone else hogs all the resources and the system goes down).

Vulnerabilities still come in three flavours

The first category is where real complexity creeps in. Lots of software bugs can be leveraged to trick the software so that instead of receiving expected data, such as a name or phone number, it actually receives new code which effectively rewrites the actual software. While generally the most complex attacks to carry out, the impact of these can be devastating to a system.

Secondly, (mis)configuration issues are huge and contribute to about one third of vulnerabilities that we see affecting businesses over a given year. This is like anything from leaving a window in your house unlocked to leaving your front and back doors wide open. Some of the simplest attacks leverage these types of vulnerabilities and although they are generally quick and easy to fix, this assumes that you knew about them in the first place.

Finally, a resource availability breach is akin to using a giant loudspeaker at a concert, to overpower and drown out the music with noise. It's not that sophisticated and is a less-than-elegant attack, although the impact can be severe (unless you can do without online banking and streaming TV). The other two categories, however, are more Hollywood-hacker and exciting.

There are some cloud-specific issues but they all fit into the above categories. Ever hear of AWS S3 buckets? Well, some organisations were not configuring them correctly and leaving them exposed to the internet. Another common issue is to leave administration services exposed, allowing for easy attacks such as password guessing, which can lead to full compromise of a system. These types of issues have recently become very prevalent in a number of high-profile breaches by hacker groups with steampunk-sounding names. These all generally come down to misconfiguration and even though they might present themselves differently, the issue is the same as exposing any internal system to the internet unintentionally. If only we knew it was there.

Visibility is everything

How can we even start to get a handle on things? Visibility is everything. Visibility of not only what we have, but where it is, too. Only then can we get into where our vulnerabilities lie. We need to be able to map our attackable footprint in a meaningful way as vulnerabilities arise in both systems which change frequently and also those that don't. No software runs in isolation and bugs can be introduced in your software or the multitude of software that is not yours but which you need in order to make your code work. We cannot test what we cannot see and surely cannot secure that which we do not know about. Visibility is key.

Rahim Jina, chief operating officer and co-founder of Edgescan, discusses cybersecurity and the cloud ahead of Dublin Information Sec 2018 - Ireland's cyber security conference, at the RDS Dublin on Monday October 15

25% off today only - for tickets and more information on Dublin Information Sec 2018 see

Belfast Telegraph

Weekly Business Digest Newsletter

This week's business news headlines, directly to your inbox every Tuesday.