GATE CS 2003 Q25 — I/O Scheduling

To understand the effect of a larger block size in a file system, we analyze its impact on two key metrics:

1. Disk Space Utilization:

   In a fixed block size file system, space is allocated in units of whole blocks. If a file's size is not an exact multiple of the block size, the remaining portion of the last block goes unused. This is known as internal fragmentation.

   On average, each file wastes about half of its last block ($rac{ ext{Block Size}}{2}$). Therefore, a larger block size increases internal fragmentation, leading to poorer disk space utilization.

2. Disk Throughput:

   Disk throughput is the rate at which data is successfully read from or written to the disk. Every disk I/O operation incurs significant overhead due to seek time and rotational latency.

   With a larger block size, more data is transferred per disk access (amortizing the seek and rotational overhead). Additionally, fewer blocks are needed to store a file, which reduces the number of metadata lookups and disk seeks. This results in better disk throughput.

Thus, using a larger block size leads to better disk throughput but poorer disk space utilization.