People today are increasingly reliant on smartphones, smart speakers and other gadgets. Most can’t imagine going more than a few hours without using a computer, and some of them spend most of their work days sitting in front of one. This shift towards a tech-centric culture means people are at a much higher risk of cyberattacks.
For now, a common way to safeguard against attacks is to make software patches and install them on users’ computers as necessary. Similarly, virus and malware scanners detect suspicious files and keep them quarantined in dedicated folders on a hard drive.
While using a prototype processor fitted with the chip, the people on the research team demonstrated how the tiny component successfully prevented every kind of control-flow hack, which is one of the most commonly used and dangerous attacks hackers carry out.
Research indicates malicious or criminal attacks leading to data breaches are the most costly, resulting in an expense of $157 per user. So, the longer an attack goes undetected, the more expensive the catastrophe becomes.
Researchers working for the US Army believe they found a method that allows detecting harmful network activity sooner than previously used techniques permitted. For example, distributed network intrusion detection tasks a small number of specialty analysts to monitor several networks simultaneously. Sensors on a protected system transmit data to analysis servers, which is a bandwidth-heavy process.
Most systems minimize the bandwidth used by only sending summaries of network traffic. But that means analysts only see snapshots and often spend too much time investigating false positives, or do not have enough details in context to notice genuine attacks.
The researchers hypothesized that malicious network activity manifests early. They developed a tool that stops network transmissions after a predefined number occurs. The next part of the investigation involves compressing traffic analysis to less than 10% of its original volume while sacrificing 1% or less of the cybersecurity alerts.
Large-scale data breaches at companies like Equifax and Uber make company leaders more aware of the potential consequences associated with poor cybersecurity. Even so, many business entities remain unprepared. A 2017 study polled thousands of international businesses and classified their cybersecurity readiness level as novice, intermediate or advanced.
In all cases, at least 70% of the companies fell into the novice category. Due to the popularity of the cloud for businesses and the fact that many are so unprepared concerning cybersecurity, some people are exploring specific ways to secure the cloud. The blockchain is one viable possibility for keeping valuable details, such as business intelligence information, safe from cybercriminals.
Most people know of blockchain technology associated with cryptocurrencies. Information gets verified and permanently added to a digital ledger. As such, it’s difficult to tamper with the content, especially since the blockchain gives visibility and transparency to all involved parties.
Many of today’s cybersecurity detection technologies can identify anomalies. When they detect activity that strays from the norm, the systems notify human technicians to take a closer look. A research team from MIT wondered if they might push cybersecurity forward by combining machine learning artificial intelligence (AI) with human intuition. Typically, platforms that use machine learning get smarter over time without input from people.
The technology works by poring over the data and grouping it into clusters through an unsupervised learning process. The goal is for the technology to figure out which strange events are likely cybersecurity attacks. However, the system doesn’t stop there. Next, it provides the clustered data to human analysts. Those people then apply their knowledge and experience when checking the algorithm’s findings.
The humans verify which events are genuine attacks, then give feedback used to make better models for the next set of data. Moreover, the existing models can get better from the updated data in a matter of hours. As such, there is a low to non-existent risk that scientists would rely on outdated algorithms for too long.
One of the challenging realities of cybersecurity is that risks can come from multiple sources. For example, a person might unknowingly download an attachment contaminated with malware. Or, an adversary could attack the entire network by focusing on a detected flaw. So, one practical cybersecurity approach entails looking for numerous kinds of threats and safeguarding against all of them as much as possible.
The four examples above all concern technologies in progress; this is a glimpse into the fruits of such research. A company called Mistnet recently launched a product called CyberMist. Advertised as the first multi-entity detection and response platform, this tool offers real-time prevention of threats and gives visibility associated with users, networks or hosts.
It combines edge computing and AI analytics to find threats in less than an hour. Continuously updated metrics on the product’s homepage indicate CyberMist had a 99% reduction in false positives over the last 30 days.
Even though many people feel unsettled when they think about how a cyberattack could affect them, the fact that cybersecurity professionals are exploring such promising ways of reducing or eliminating those incidents is hopeful.
Cybersecurity researchers know how crucial it is to keep threats at bay, and they aren’t afraid of considering all possible options. As technologies improve, so should the choices for people who want to bring more high-tech applications to the cybersecurity sector.
This content was originally published here.