High-Performance AWS Networking                                      

Rather than reconstruct the networking service from the ground up, Amazon has created an AWS-like response: an additional set of offerings to address the needs of applications that require high-performance networking. This option leaves the vast majority of AWS users happily using the standard AWS offering while providing another option to the smaller portionof users who need better networking performance.The key phrase here, high-performance AWS networking, is built on three specialized instance types: High I/O, Cluster Compute, and Cluster GPU.                            

 

This arrangement makes a lot of sense — after all, if you need a lot of network capacity, you’re probably doing a lot of computing as well. It also makessense to not impose the higher cost of high-performance networking on users with more modest computing requirements. This approach is consistent withAmazon’s: Provide an inexpensive offering for those who don’t need morethan that, and extend the offering for those who want more and are willing to pay for it. This strategy is in direct contrast with almost all other cloud computing providers, who force all users to pay for expensive equipment, even if a user wants to run only a small or non-mission-critical application.