TO: Professor Jason Ellis

FROM: Nargis Anny

DATE: September 22, 2020

SUBJECT: 500-word summary

This is a 500 word summary of “A Smart Agent design for Cyber Security based on HoneyPot and Machine Learning”. The article highlights the rise of security risks that come with the rise of social media and the World Wide Web. We’re also introduced to the programs that keep the security programs running, as well as the setbacks it’s brings to computer systems worldwide.

In the article, GDATA states how every year there are over millions of Cyber attacks that have been discovered. These issues are often involves analysis tools that keep track information. However, the difficulty is keeping an eye on every problem that arises. With a better understanding of how Cyber attacks work, there’s a better chance of preventing future issues. HoneyPots is one of the most prominent cyber security programs to date. Developed in 1992, HoneyPots is utilized as a monitoring and detecting system that locates harmful malware. Now future attacks can be prevented before they even find a system to disrupt. Part Two talks about Anomilies, data which has to be protected from harmful versions of software. With Social Media sites such as Myspace or Facebook, these sites need to be observed in order for a social ‘Honeypot”, to detect harmful profiles, as well as any other threats out there. Authors suggest a linkage defense system, which can bypass the setbacks brought on by past tools that tried to work. The Linkage system has the Honeypot’s and the defense system coexist together by having their management and communication tools work together. This system is based on a SMNP model code used in network management. Now Future intruders will be blocked by firewalls, if they try to hack into the system. Machine Learning is where we learn that computers operate under the system program that it’s been assigned. The concept of “Machine Learning”, keeps the computers adjusted to data structure and how to operate properly. Machine Learning has training models that separate into two phases in order to function. The first phase is estimating the data through training, by demonstrating tasks like recognizing animals in images or speech translation. The second phase is production. Here we see new data pass through the system in order to get the computer to complete an objective. The K-Means algorithm helps maintain clustering from certain systems. Eddabbah indicates that the “K –Algorithim is a faster solution to the issue it still has major setbacks” (Eddabbah, 2020, Page 3). The Decision tree helps branch out all data structures in case of testing. Part 4 jumps back into HoneyPot, this explains the different security communication networks. The first part is HoneyPot deployment which can monitor either Internal or External attacks on the system. With this we can see attacks that are carried out or attempted on any network. With DMZ’s (Demilitarized zones), HoneyPot function as a way to provide public internet service away from the computer’s internal network. Next, we have networks like KFSensor, Netfacade, Specter and CurrPorts. KFSensor is a server that watches out for connections with the network. Netfacade allows numerous network hosts interactions through unused IP a dresses. Networks also have to direct security threats to the firewall and eventually the honeypot will separate it to see if it’s serious or not. To conclude, network security is a very serious problem due to constant evolving and threats are hard to manage.

References:

Kamel, N / Eddabbah, M / Lmoumen, Y/ Touahni, R “A Smart Agent Design for Cyber Security Based on Honeypot and Machine Learning”, Security & Communication Networks, (2020) ID 8865474 (9 Pages), 2020