Nowadays, many chips adopt designs with lower single core performance but more cores, piecing together a CPU with many cores and high overall computing power, and calling it a cloud native product [1].
What are the characteristics of the cloud native load they are targeting that drive such a design? Why can they benefit from more cores and higher overall performance (although only weaker memory bandwidth compared to the growth rate of computing power)
I searched online for discussions about cloud native, but more focused on program development and deployment, and did not find any discussion on the characteristics of workloads themselves.