An adversary exploits a data structure shared between multiple applications or an application pool to affect application behavior. Data may be shared between multiple applications or between multiple threads of a single application. Data sharing is usually accomplished through mutual access to a single memory location. If an adversary can manipulate this shared data (usually by co-opting one of the applications or threads) the other applications or threads using the shared data will often continue to trust the validity of the compromised shared data and use it in their calculations. This can result in invalid trust assumptions, corruption of additional data through the normal operations of the other users of the shared data, or even cause a crash or compromise of the sharing applications.
The table below shows the views that this attack pattern belongs to and top level categories within that view.
The target applications (or target application threads) must share data between themselves.
The adversary must be able to manipulate some piece of the shared data either directly or indirectly and the other users of the data must accept the changed data as valid. Usually this requires that the adversary be able to compromise one of the sharing applications or threads in order to manipulate the shared data.
None: No specialized resources are required to execute this type of attack.