We present the architecture of an automatic early warning system (EWS) that aims at providing predictions and advice regarding security threats in information and communication technology without incorporation of cognitive abilities of humans and forms the basis for drawing a situation picture. Our EWS particularly targets the growing malware threat and shall achieve the required capabilities by combining malware collectors, malware analysis systems, malware behavior clustering, signature generation and distribution and malware/misuse detection system into an integrated process chain. The quality and timeliness of the results delivered by the EWS are influenced by the number and location of participating partners that share information on security incidents. In order to enable such a cooperation and an effective deployment of the EWS, interests and confidentiality requirements of the parties involved need to be carefully examined. We discuss technical details of the EWS components, eva...