Multi-tenant radio-based application pipeline processing server
An offloading manager is launched at a computing device. The offloading manager schedules execution of a first set of network functions of a first radio-based application at a first network function accelerator, and the execution of a first set of network functions of a second radio-based application at a second network function accelerator.
1. A system, comprising:
a server configured to run virtualized radio access network (RAN) network functions, wherein the server comprises one or more processors and a memory; and
a plurality of hardware network function acceleration devices, including a first network function accelerator and a second network function accelerator;
wherein the memory includes instructions that when executed on or across the one or more processors:
launch, at the server, an offloading manager, a first request handler and a second request handler;
cause the offloading manager to schedule execution, at the first network function accelerator in response to a message from the first request handler, of a first set of one or more network functions of a first radio-based application of a first client of a service of a provider network, wherein the first set of one or more network functions includes a network function of a physical layer of a radio-based application technology stack;
cause the offloading manager to schedule execution, at the second network function accelerator in response to a message from the second request handler, of a second set of one or more network functions of a second radio-based application; and
execute the first set of one or more network functions at the first network function accelerator and execute the second set of one or more network functions at the second network function accelerator.
2. The system as recited in claim 1 , wherein the first network function accelerator is incorporated within a first network function accelerator card connected to the one or more processors via a peripheral interconnect.
3. The system as recited in claim 1 , wherein the first set of one or more network functions includes a particular network function of a first category of network functions defined in a radio based application technology standard, and wherein the second set of one or more network functions does not include a network function of the first category.
4. The system as recited in claim 1 , wherein the memory includes further instructions that when executed on or across the one or more processors further:
cause the offloading manager to schedule execution, at the second network function accelerator, of a third set of one or more network functions of a third radio-based application; and
execute the third set of one or more network functions at the second network function accelerator.
5. The system as recited in claim 1 , wherein the offloading manager is implemented within a virtualization management component.
6. A computer-implemented method, comprising:
launching, at a computing device, an offloading manager, a first request handler and a second request handler;
causing the offloading manager to schedule execution, at a first network function accelerator of the computing device in response to a message from the first request handler, of a first set of one or more network functions of a first radio-based application, wherein the first set of one or more network functions includes a network function of a physical layer of a radio-based application technology stack;
causing the offloading manager to schedule execution, at a second network function accelerator in response to a message from the second request handler, of a second set of one or more network functions of a second radio-based application; and
executing the first set of one or more network functions at the first network function accelerator and execute the second set of one or more network functions at the second network function accelerator.
7. The computer-implemented method as recited in claim 6 , wherein the first radio-based application is executed on behalf of a first client of a provider network, and wherein the second radio-based application is executed on behalf of a second client of the provider network.
8. The computer-implemented method as recited in claim 6 , wherein the first set of one or more network functions includes a particular network function of a first category of network functions defined in a radio based application technology standard, and wherein the second set of one or more network functions does not include a network function of the first category.
9. The computer-implemented method as recited in claim 6 , further comprising:
causing the offloading manager to schedule execution, at the second network function accelerator, of a third set of one or more network functions of a third radio-based application; and
executing the third set of one or more network functions at the second network function accelerator.
10. The computer-implemented method as recited in claim 6 , wherein the computing device is installed within a server rack, and wherein the computing device occupies one rack unit ( 1 U) of the server rack.
11. The computer-implemented method as recited in claim 6 , further comprising:
causing the offloading manager to determine that the first network function accelerator has been configured for exclusive use on behalf of the first radio-based application, wherein the first set of one or more network functions is scheduled at the first network function accelerator based at least in part on the determination that the first network function accelerator has been configured for exclusive use.
12. The computer-implemented method as recited in claim 6 , further comprising:
causing the offloading manager to schedule execution, at the second network function accelerator, of a third set of one or more network functions of the first radio-based application; and
executing the third set of one or more network functions at the second network function accelerator.
13. The computer-implemented method as recited in claim 6 , further comprising:
causing one or more metrics obtained from the first network function accelerator to be presented via a programmatic interface.
14. The computer-implemented method as recited in claim 6 , further comprising:
causing the offloading manager to obtain a performance metric of the first network function accelerator, wherein the first set of one or more network functions is scheduled at the first network function accelerator based at least in part on the performance metric.
15. The computer-implemented method as recited in claim 6 , further comprising:
causing the offloading manager to determine a particular application area, from among a set of application areas defined by a standards organization, to which the first radio-based application belongs, wherein the first set of one or more network functions is scheduled at the first network function accelerator based at least in part on the particular application area.
16. One or more non-transitory computer-accessible storage media storing program instructions that when executed on or across one or more processors:
launch, at a computing device, an offloading manager;
cause the offloading manager to schedule execution, at a first network function accelerator of the computing device, of a first set of one or more network functions of a first radio-based application, wherein the first set of one or more network functions includes a network function of a particular layer of a radio-based application technology stack;
cause the offloading manager to schedule execution, at a second network function accelerator, of a second set of one or more network functions of a second radio-based application; and
execute the first set of one or more network functions at the first network function accelerator and execute the second set of one or more network functions at the second network function accelerator.
17. The one or more non-transitory computer-accessible storage media as recited in claim 16 , wherein the particular layer is a physical layer, and wherein a network function of the first set of network functions whose execution is scheduled at the first network function accelerator comprises one of: a coding function, a rate matching function, a scrambling function, a modulation layer mapping function, a precoding function, a resource mapping function, a digital beamforming function, a Fast Fourier Transform (FFT) function, a cyclic prefix insertion function, a cyclic prefix removal function, an inverse FFT function, a de-mapping function, a channel estimation function, a pre-filtering function, an equalization function, a demodulation function, a descrambling function, a rate de-matching function, or a decoding function.
18. The one or more non-transitory computer-accessible storage media as recited in claim 16 , storing further program instructions that when executed on or across one or more processors further:
cause the offloading manager to determine a maximum number of network function accelerators to be used for the first radio-based application based at least in part on a category of compute instance launched at the computing device at the request of the first client.
19. The one or more non-transitory computer-accessible storage media as recited in claim 16 , wherein the first radio-based application is a private 5G network application, and wherein the second radio-based application is a public 5G network application.
20. The one or more non-transitory computer-accessible storage media as recited in claim 16 , storing further program instructions that when executed on or across one or more processors further:
launch (a) a first compute instance of a virtualized computing service at the computing device in response to a first launch request and (b) a second compute instance of the virtualized computing service at the computing device in response to a second launch request, wherein the offloading manager schedules execution of the first set of one or more network functions in response to a request from the first compute instance, and wherein the offloading manager schedules execution of the second set of one or more network functions in response to a request from the second compute instance.