SATA Release 4.0 has been discussed but may never happen. There is a growing realization that what may eventually bring about the demise of SATA is not failure to introduce an even faster data transmission rate but the latency that is inescapable with SATA.
Latency may be thought of as simply delay. In fact, it IS delay. Latency is the additional time it takes from the beginning of a process to the end, where that time is built into the process by architecture, mode of working or some other factor.
In a network, latency is the time taken by a packet of data to move from one point to another. (Internet latency is simply a particular form of network latency). The factors that cause and add to latency are:
Propagation: the time it would take the data to move if it moved at the speed of light. Tiny, of course, but it’s there and it’s added to y the other factors in this list.
Transmission: different media introduce different amounts of delays since some (like fiber optics) are very fast transmission media and some (like copper wire) are much slower.
Processing through gateways: each node examines the data and the header. It all adds time.
Hardware delays: it doesn’t matter how fast data moves between two points; if any component at sending or receiving end is slow, latency will be added.
There is no question that SATA latency is higher than SAS latency. SAS offers up to more than double the rotational speed and read/write is also more precise and, therefore, faster. Research suggests that, in some cases, the latency of a SATA drive may be more than double that of a SAS drive.
Of course, how much this matters – and whether it justifies the extra cost of SAS drives – will depend on the environment and the type of application.