Our research is mainly concerned with distributed and parallel systems for big data processing which typically aggregate many servers in a cluster or the cloud by using fast network technology, e.g. InfiniBand. We focus on distributed in-memory systems for interactive graph applications supporting fast online queries on billions to trillions of data objects. These queries typically result in irregular access patterns on graphs dominated by small data objects (32-128 byte) requiring a very good vertical and horizontal scalability and low-latency remote data access within microseconds. This results in many challenges from the application level down to the operating system.