I think it's important to point out that best practices always come back to application -- there are no blanket best practices.
A few years ago, Intel studied this with IOmeter and I believe they showed typical I/O rates for FC loops can remain pretty solid, without much degradation in the area of 45 drives. Again, I don't recall the workload model they used, but I was surprised that the number was as high as it was. The other thing I recall was that degredation occurred smoothly for numbers greater than that. In other words, the performance "knee" was not very pronounced.
There are, however, many things that can influence this, including the storage configuration with a number of drives in a RAID array and the level of RAID being used. Indeed, large subsystems are complicated things.
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .bphAaR2qhqA^0@/searchstorage>discussion forums.
This was first published in November 2003