US 12,255,951 B1
Capacity management for radio-based applications
Diwakar Gupta, Seattle, WA (US); Benjamin Wojtowicz, San Francisco, CA (US); Upendra Bhalchandra Shevade, Washington, DC (US); Ximeng Simon Yang, San Francisco, CA (US); and Ishwardutt Parulkar, San Francisco, CA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Jun. 30, 2021, as Appl. No. 17/364,779.
Int. Cl. H04L 67/1031 (2022.01); H04L 67/567 (2022.01); H04W 28/084 (2023.01); H04W 28/16 (2009.01)
CPC H04L 67/1031 (2013.01) [H04L 67/567 (2022.05); H04W 28/084 (2023.05); H04W 28/16 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
in response to one or more programmatic requests from a client of a cloud provider network, configuring a first extension resource group of the cloud provider network at a premise external to the cloud provider network, wherein the first extension resource group comprises a first number of servers including a first radio-based application pipeline processing server, wherein the first radio-based application pipeline processing server includes one or more processors and a network function accelerator card of a first type;
executing a first set of network functions of a radio-based application at the network function accelerator card;
determining to implement a second extension resource group of the cloud provider network at the premise to change a capacity of the radio-based application, including selecting (a) a second type of network function accelerator card to use to execute the radio-based application in the second extension resource group based at least in part on an anticipated type of future workload of the radio-based application at the premise and (b) a number of resources to use in the second extension resource group based at least in part on an anticipated level of the future workload, wherein the second type of network function accelerator card is different from the first type of network function accelerator card in terms of:
a vendor or manufacturer,
one or more network functions accelerated, or
a performance capacity;
configuring the second extension resource group of the cloud provider network at the premise, wherein the second extension resource group comprises a different number of servers than the first extension resource group and at least one server that includes the second type of network function accelerator card; and
executing a second set of network functions of the radio-based application at the second extension resource group without interrupting end-user interactions of the radio-based application.