ROBUST MEMORY MANAGEMENT USING REAL TIME CONCEPTS

Memory fragmentation is the development of a large number of separate free areas.Memory management in embedded systems demand effective implementation schemes to avoid fragmentation problem.Existing dynamic memory allocation methods fail to suit real time system requirements.Execution times need to be deterministic and this motivates the need for allocation and deallocation to be done in constant time with the help of API's.In µ C/OS-II, memory allocation is semi-dynamic and a buddy allocator dynamic memory allocation algorithm is commonly used.Programmer must statically allocate a memory and partition the region using µ C/OS-II Kernel API.Tasks can only request pre-partitioned fixed-size memory space from µ C/OS-II.Memory allocation times are influenced by the ratio of memory allocation to the stack size of the task.In this research work memory management in LPC 1768 environment using RTOS µ C/OS-II is proposed.Effective sharing of memory blocks among tasks co exists with partition.The captured results shows that the memory allocation and deallocation suits real time.The implication of the work is that, the necessity to reserve a static set of locations ahead of time is eliminated so that memory can be allocated at compile or design time.