Research Project on Python

Hi,

I am doing a research project for my Bachelor of IT (honours) on Machine Learning for Cloud Security.

This research paper discusses the Fraudulent Resource Consumption (FRC) Attack and uses Support Vector Machines (SVM) to detect cloud-based FRC attacks. Fraudulent Resource Consumption (FRC) attacks are created by slowly using cloud services’ metered resources. The attacker’s goal is to abuse the utility pricing model by stealing cloud resources. This skilful resource overuse results in a significant cost burden for the client. These assaults employ low-intensity HTTP requests per hour, like legitimate users. Due to this, FRC attacks are difficult to detect. FRC is an Economic Denial of Service (EDoS) attack that targets cloud adopters’ financial resources by increasing their costs. Unlike DDoS assaults, which can temporarily block legitimate users from accessing services, EDoS attacks can significantly increase cloud users’ costs. Support-vector machines (SVMs, also known as support-vector networks) are supervised learning models that examine data for classification and regression analysis.

Now I want guidance for a script that Can capture this generated FRC traffic, run SVM on it for training and then run SVM on it for testing. Training can be one script, and testing can be another

script.

I shall be highly grateful if you could kindly guide me in this.

Thanks & regards,

Osama Faheem

Maybe ask a professor or someone at your university?
Or look for a more machine-learning focused website. I highly doubt you’ll find someone here who has significant expirience in this very specific field of applied ML.

as Jagaya said this is very specific, but worth a try checking the ethical hacking courses from the freeCodeCamp youtube channel

1 Like

Consult your advisor.

In the meantime, assuming you want to classify traffic as an FRC attack (or not) and have simulators for both FRC and non-FRC traffic available, the basic procedure is straightforward:

  • Generate lots of labeled samples of both FRC and non-FRC traffic
  • Separate the samples into training and test sets. K-fold validation is useful here.
  • Train an SVM on the training data. There are lots of SVM libraries available. Alternatively, write your own. It’s not hard.
  • Validate performance on the test sets.

If you don’t have traffic simulators, you have more work to do. Again, consult your advisor.

jrm

1 Like