See, let me give you a couple of examples.
Imagine that you have a highly concurrent SaaS platform. And that you use, let say, a postgre db. If your app is mostly, or relatively, high on writes, then ssd will make a huge difference. HD IO ops are the bottleneck there, and thats why platforms that leverage async, non blocking ops are more effective (blocking ops: db and hd calls, http requests, or even array iterators). Thats why you see Amazon offering ssd solutions that perfectly work for those scenarios.
But what if, on the other hand, your app is low on writes, but high on reads? What you will see is that the overall impact of a faster store diminishes . Most probably, on those cases, the key factor are the cache layer on top of it. Those caches wont be running on an hd, although they can make use of it on misses, invalidations or preemptive loads.
Another example: a java app thats fully loaded on memory, and has no blocking ops (ie: a calculator app). Effect of ssd? Only on startup time, and thats it.
Anather example: A site that crawls other sites and fetches data. Most relevant blocking ops? http reqs. Thats why browsers mitigate that making them async from an arch standpoint(single, non blocking thread).
What is more impactful on an online game? faster hd, or proximity to datacenter?
Now imagine games. Lets say a game luke Limbo, or Inside. They are probably mostly fully loaded on memory. Its the calculator app on steroids, if you want. what is the effect of a faster hd? mostly neglectible. Most games rely on reducing the amount of these expensive io, blocking ops to the disc.
The fact that eventually, in a short term future, some devices will have much faster, closer to ram memory speeds wont shift paradigm instantly. It will take years to see gigantic changes other than load speeds. Thats why Sony and MS are marketing that as the main advantage. You wont see “2x improvements”, whatever the hell that means, due to ssd speed now nor on the midterm.
