| Popis: |
The evolution of mobile systems from hardware and software standpoints, and their increased use in our daily life has direct implications on all the sub-systems of a smartphone. The storage sub-system, in particular, has changed deeply over the years (in terms of density, power consumption, functionalities, form factor, throughput, and latency) and has become one of the core systems in a smartphone. To search for system inefficiencies and to explore innovative storage solutions, it is important to have a comprehensive view of how the storage device interacts with the OS, System-on-Chip CPU, and SW layers. But standard benchmark tools are not suitable for this purpose. Therefore, we propose a methodology for live tracing the storage activity on commercial phones used by different users. With such methodology, we quantify the activity on the storage sub-system and observe possible IO patterns during a few days of usage of three different commercial smartphones. Moreover, we investigate also system level effects of introducing a low-latency emerging memory as storage device. |
