Operating Systems Discusion

Part 1: Memory Management

Discuss the limitations of memory management schemes in early systems. What were the most problematic aspects of these schemes? Why were they sufficient for the first three generations of computers?

Part 2: Respond to two peers

Peer 1 Reese

Earlier memory management schemes were Single-User Contiguous. This scheme would only run one process at a time as long as the process did not use more than the available amount of memory the system had. Additional processes would need to wait and would not be processed until previous process was finished. The next scheme was Fixed Partition. This scheme broke the memory into pre-configured sections. Each section could run a process but only if it would fit in one of the sections of memory. Once a process started in a section that entire section of memory could not be used by anything else until that process finished. The sections could be resized but the system would need to be restarted. The next type of scheme was Dynamic Partitions. This scheme works great when the system first starts, the memory is broken up into sections of needed size, but much like a hard drive when adding and deleting files, the memory becomes fragmented and becomes difficult in any large processes need to be run. This fragmentation issue was addresses with Relocatable Dynamic Partitions. This allows memory to be assigned based on the required space, but once a process finished all fragmented gaps are compacted, always leaving as much memory available as possible. In the early days of computing, not many processes needed to be finished in short amounts of time. As systems and processors became faster, the need to process items quickly became necessary.

Peer 2 Megan

The early memory management schemes had a few limitations. The first being, that a job had to be loaded and finished completely before that memory could be used again. In the single-user contiguous scheme that meant one job could only be done at a time. The second issue arises from wasted memory in the case of internal or external fragmentation. With fixed and dynamic partitions there is always the possibility of memory space being wasted. This occurs most notably when jobs are loaded in queue and the first available space is much larger than what is needed by the job but it gets loaded in anyway. With the partitions, multiple jobs can be run at one time but but there is going to be a queue for larger jobs. The early generations of computers were fine with utilizing the single-user contiguous schemes or even the fixed partitions because the technology was not as advanced as what we see today. What would take rooms of hardware to make up a computer in the 1940's takes up a thin 3 pound laptop today.