Software design is the important stage where the code is really put to work to deliver or build an business function and application. This is a stage where the SRS (Software Requirements Specifications) is finalized and signed off for design and development.
The major difficult in software design is to incorporate the business requirements as well as do threat modeling to understand the attack surface of the applications. Many applications do not show up problems in regular normal use but show up funny ways of responding when provided with a different input or action not generally considered as part of the application design.
The major areas that a threat model derived needs to address are – The Microsoft STRIDE model provides the following areas to be addressed as part of the design.
1. Spoofing
2. Tampering
3. Repudiation
4. Information Disclosure
5. Denial of Service
6. Elevation of Privilege
7. Integrity of Data
Even though all aspects are not covered most of them are covered in the above threat model. The best way to go about is to break the applications into the different threat vectors and address each one of them.
Each process, data store and elements that are part of the application design are specifically vulnerable to one of the vectors. A matix is prepared addressing each of the vectors’ effect on the application. All elements including web services, people who use the systems and the end points where data is handed over to another application are places of vulnerability. Once we map all the data flows from the application - In the form of a map , we identify areas that are possibly vulnerable to say Information Disclosure, and may have a threat profile associated with it.
The mapping will be very crucial to the way the application is designed. It always helps to think in the same way as a would be attacker does. In this way the model as it evolves will replicate a real threat scenario and the way an attack may be carried out. This is also one of the main reasons to differentiate between the design team and the team that threat models. This removes any bias the developer has for his system and is modeled in such a way that every aspect that can be exploited is tested thoroughly.
Though it is easier to write about threat modeling, it is an evolving science and needs to be modified in line with when it is required, new attack parameters are found and newer vectors are discovered, as they do the threat model that is used for design of the software needs to keep in tow with the evolving threat vectors
The major difficult in software design is to incorporate the business requirements as well as do threat modeling to understand the attack surface of the applications. Many applications do not show up problems in regular normal use but show up funny ways of responding when provided with a different input or action not generally considered as part of the application design.
The major areas that a threat model derived needs to address are – The Microsoft STRIDE model provides the following areas to be addressed as part of the design.
1. Spoofing
2. Tampering
3. Repudiation
4. Information Disclosure
5. Denial of Service
6. Elevation of Privilege
7. Integrity of Data
Even though all aspects are not covered most of them are covered in the above threat model. The best way to go about is to break the applications into the different threat vectors and address each one of them.
Each process, data store and elements that are part of the application design are specifically vulnerable to one of the vectors. A matix is prepared addressing each of the vectors’ effect on the application. All elements including web services, people who use the systems and the end points where data is handed over to another application are places of vulnerability. Once we map all the data flows from the application - In the form of a map , we identify areas that are possibly vulnerable to say Information Disclosure, and may have a threat profile associated with it.
The mapping will be very crucial to the way the application is designed. It always helps to think in the same way as a would be attacker does. In this way the model as it evolves will replicate a real threat scenario and the way an attack may be carried out. This is also one of the main reasons to differentiate between the design team and the team that threat models. This removes any bias the developer has for his system and is modeled in such a way that every aspect that can be exploited is tested thoroughly.
Though it is easier to write about threat modeling, it is an evolving science and needs to be modified in line with when it is required, new attack parameters are found and newer vectors are discovered, as they do the threat model that is used for design of the software needs to keep in tow with the evolving threat vectors
Comments