Our use case for SandBlast Agent is that our team is set up in multiple geographies, such as, India, Sri Lanka, UK, North America, and Australia (where we have a bit of business). We have courses for an educational client which need go to market, schools, instructors for hire, and students. Given that there was COVID-19 and a lockdown, there was an increase in the digital demand for learning courses. So, we wanted to secure our courses from cyber attacks. Thus, we wanted an end-to-end security system in place that would prevent/save us from cyber attacks and protect our sensitive data.
Systems can be accessed on multiple devices, whether they be laptops, Macs, Windows, or mobile devices. Those devices could be connected to a home or public network on a platform, like a Chrome browser, Mozilla Firefox, or Safari. We have been able to track this through reports by seeing how vulnerable those agents are to attacks. Then, we determine how they can become more secure, so we can stay on the cloud and mobile devices. These are the areas where we are trying to use their reports and tighten our security, putting more systems in place to prevent attacks.
Cognizant had a malware attack recently, as the threat of cyber attacks has increased, and a lot of customer data was compromised. However, because this Check Point SandBlast technology was there in place, we were able to thwart the cyber attacks that were attempted. Most of the time, these attacks are college kids trying to do some phishing attacks or look into sensitive data. With SandBlast, it is possible to identify those attacks at the very source, preventing those attacks and keeping us secure.
Going forward, we are planning to extend it to authors and professors who are helping us author our content. For example, if there is an author who will be taking help from various professors in university or instructors in schools, then they will need to get their inputs. What happens is they expose their course to those authors on their networks, devices, laptops, mobiles, or tablets. They access the course through an application. Now, those authors and professors don't have an app login because they might be a third-party vendor. So, we are trying to have the SandBlast Mobile version on this site as well, based on the impressive performance of SandBlast, so our data remains secure and more users are able to utilize our systems and access our data. This will make it more valuable for our end users.
On the coverage part, there are malware, phishing, operating system exploitations, denial-of-service attacks, and man-in-the-middle attacks (MITM), so we have classified the attacks that can happen on a learning, educational system, like ours into five to six categories. With SandBlast Agent deployed in the cloud, we have good coverage to cover these attacks, as it is very extensive. The best part is (through our reports) we were able to identify the type of attacks. So far, our security has been 100 percent. We have not felt that a data breach has happened, so we are pretty happy with SandBlast Agent.