Dynamic memory management for embedded real-time systems

Dynamic memory storage has been widely used during years in computer science. However, its use in real-time systems has not been considered as an important issue because the spatial and temporal worst case for allocation and deallocation operations were unbounded or bounded but with a very large bound. TLSF (Two Level Segregated Fit) is a new allocator has been designed specifically to meet real-time constraints. These constraints are addressed in two axis: Time and Space. While the temporal behaviour of TLSF is excellent, 0(l), the spatial behaviour is as the best of the known allocators. The spatial constraint is specially required in embedded systems with mitided resources. An efficient and guaranteed use of memory is needed for these systems. In this paper we compare the temporal and spatial performances of the TLSF allocator comparing it with the most relevant allocators.