One reply is from one of the original engineers working on EC2 that said this swapover had nothing to do with AWS and it's origins, and another reply is from a former distinguished engineer stating that they swapped from compaq/digital tru64 alpha servers, not sun.
In 2012 I joined Amazon in a a small data analytics org within the supply chain. We were tasked with reducing split shipments: multiple item orders that could not ship out on a single box. The top reason for splitting a shipment was because we simply didn't have the right combination of items in a single fulfillment center. By the time I had gotten there, my manager had already successfully implemented the first transshipment network, moving items from one warehouse to another so they could ship out with the right combination of items and reduce a split shipment. But by then we were reaching diminishing returns on transshipment, and splits were still rising nationwide, and I was given the chance to provide analysis and new insights.
I probably ran a million analyses that year, but one of the most salient was that the major reason for splitting shipments was because we we sold so few of one of the items that it didn't make sense to keep more than one or two in stock nationwide. Those types of tail items were typically stocked in one of three fulfillment centers, the largest of the fulfillment centers in three regions of the US. And the number of orders getting split shipments between those three fulfillment centers was massive...more than enough to cover the cost of daily cargo plane transshipments between the three fulfillment centers.
I ended up leaving that team and moving on, eventually leaving Amazon completely. Right about the time I left, Amazon made a huge anouncement: Amazon Prime Air (eventually renamed Amazon Air to distinguish between the drone delivery idea). The press releases made it sound like they were launching it to deliver packages nationwide, but a quick call to my former colleague confirmed that upon launch, the only items they were moving was unpackaged inventory ... cargo plane transshipment, as my analyses had pushed for a few years earlier. Since then, they've expanded to actually moving customer shipments, but the service was initially entirely justified based on an analysis of split shipments I had done years earlier.
I say this because all humans like telling stories about themselves and the oversized contribution they had on something that made it big. It's cool, and it makes us feel important. But I was hardly the only person that worked on that project. There were likely dozens of others who were pushing for cargo planes for different reasons and maybe the transshipment story was enough to tip it over the edge. There were many more who bought into a crazy idea really early and came to own it, at least far more than someone who did a throwaway analysis years earlier. And they all probably have different versions of the same story. And they can all be true simultaneously, at least true enough to matter. Because none of us have a perfect recollection of the past...we all have a history written into our minds that is colored by the experience that we lived.
This also applies in families, in retelling of historical events.
In addition, the capability of the story teller to enthrall can cause their version of the story to be weighted more strongly than others.
This can lead to a situation where the story told by the better “showman” gets an outsized credence.
Beware one person’s recollection, and recall that extemporaneous notes are the next best thing to a recording because it is that hard to remember what happened.
If your goal is influence, you must be at least as good of a story teller as someone else talking about something you have a memory of.