Well, IOzone benchmarks a file system by breaking up a file of a given size into records. These records are written (or read) in a different way, according to the given test, until the file size is reached.
For example, your command (iozone -R -r 1M -s 100m) asks IOzone to execute all its tests (e.g. read, re-read, write, re-write, etc) over a file of 100MB. Read/Write operations are split into records of 1MB. It means that 100 operations are done over records of 1MB to achieve the tests.
Have a look at the results. The first number is the size of the file. The second is the record length. Then, numbers correspond to the throughput recorded for the different tests. Some tests are done several time (e.g. read, write, etc). First time is done sequentially, second time is done by random location accessing.
The following document explains in details IOzone. Have a look at the description of the tests to understand their meanings.