This post was inspired by j4m32's mention of solid state drives in his post about hardware security. I'm wondering if anyone has any experiences with solid state. I've heard it's much faster, though I've not tried using a solid state drive myself, so I'm not sure how much of a difference there is.
And I've also heard that solid state drives have the same problem that small flash drives have with a lifespan limited by the number of reads/writes. For flash drives that's not such a big deal because the reads and writes aren't so often, but for solid state you're reading and writing constantly, especially when virtual memory comes into play. So I wonder what the limitations are in that area.
Anybody have any experiences or thoughts?
I don't have any experience with using an SSD, but I have read a little bit about them. I have seen the improved read speeds make a huge difference with operating system boot times and other services - such as disc caching and especially boot time indexing for search services on Windows.
*EDIT: There are also two types of SSD, synchronous and asynchronous; I have no idea what the absolute difference is between them.*
The main advantages of an SSD are that:
1. The equivalent to the traditional seek time of a magnetic drive is far lower (talking micro/pico seconds or a few clock cycles), rather than the average ~8.9 milliseconds (thousands of clock cycles waiting) a 7,200RPM disc. Even with a SCSI 10K disc this time isn't reduced significantly to compete with an SSD.
This also becomes a problem where data is not stored uniformly on the disc (hence why defragmentation is ramed down users throats) more so with a magnetic drive than an SSD, an SSD can cope with high fragmentation before there would be any noticable slowness.
Typically the addition of a larger disc cache on magnetic drives helps minimise that and improving the throughput for large data file copying, but does not improve it for a large number of smaller files - the SSD leaves magnetic drives standing in both situations.
2. I am guessing that the primary reasons an SSD is quicker is because data can be accessed directly and can be pulled parallel in Nbit chunks to the SSDs internal cache before going out via SATA to the motherboard disc controller; Where as the magnetic disc has to read off the surface in a serial manner, automatically limiting its maximum performance.
I've head that SSDs do detieroirate in their copy speeds with time, which I don't know why they do, but what I have heard from other users is that it can be managed with software utilities that modify something in the OS kernel called TRIM.
By modifying TRIM it allows better marking of unused data blocks on the SSD rather than how the kernel would normally for a magnetic drive which works with [cylinders, heads and sectors]
I am taking a guess that an SSD doesn't have any concept of this really, and has to be translated by the SSD or disc controller from "CHS" to absolute offset on the SSD. TRIM
Hope that kind of answers a few questions