Welcome to Kvariant.com

K-variant is a software system architecture to improve the security of time-bounded mission-critical systems. K-variant is a multi-variant execution environment in which variants are generated by automated source to source program transformations. The survivability of K-variant systems is measured by a metric called the survivability of an unsuccessful attack (Pu). The goal of K-variant systems is to increase Pu. The higher Pu, the more secure system.

Pu can be estimated analytically for a few cases. For other cases, the simulator is utilized. However, estimating Pu with the simulator is slow and impractical. Therefore, we estimate Pu by using machine learning techniques. You can access our Bayesian network tool from here .

On this page, you can find our latest neural network tool. The tool increases the speed of Pu estimations significantly.

In our latest experimental study, we developed an attack emulator to investigate the effectiveness of program transformations in K-variant systems. You can access the research information here .

Try neural network-based tool

[Collapse/Expand]
The probability of an unsuccessful attack (Pu) estimator
The size of memory (N):
+
-
The size of vulnerable memory (n):
+
-
The number of attacks (M):  
+
-
The number of variants (K):
+
-
Attack type (A):
v
   
Estimate with tool
Estimate with simulator

 

Benchmark

[Collapse/Expand]
Speed benchmark test

Range of N:

Range of n:

 

Range of M:

 

Range of K:

 

Attack types:

The number of random test cases:
+
-
Start Test