ARM Holdings designs a simple processor...other companies licenses this design and modify it in any way they want. As they modify and build upon their previous modifications, they in essence create their own "architecture". You can see it more clearly now with the addition of the A15 and it's power consumption issues. Each ARM licensee are scrambling to develop an "architecture" that will blend the A7s and A15s into the same SoC but somehow mimimize the use of the A15s to reduce power consumption. Samsung is trying to implement their Exynos using the big.Little concept while nVidia is doing their own implementation (Tegra 4). The problem is that the logic of when to use the A15s and when to use the A7s reside on the software. The software developers need to add this logic in each of their programs. This means EVERY software (including OSes such as Android) will have to have special code written specifically for the Exynos "architecture" and Tegra "architecture". Therein lies the problem. Software developers do not want to do this. The amount of maintainance needed to maintain code for each and every "architecture" and their iterations would be too cumbersome. Samsung gave up on their Exynos for their next Galaxy products. nVidia is proclaiming to the world how great their Tegra 4 is, even when it is not even close to finish. No doubt they are trying to convince phone makers to sign up and thereby convince to software developers to support their "architecture". In the end, there can only be 1 or 2 architectures that software developers can support. Intel will be one.
Big/little has several usage models, but none of them require end user software changes (ie, Software developer support). The simplest usage model doesn't even need any kernel/OS changes, while the more complex model does need kernel/OS changes they are small and are easy to add/maintain.