CONFIGURING DIRECT ACCESS TO HARDWARE ACCELERATOR IN AN O-RAN SYSTEM
Some embodiments provide various methods for offloading operations in an O-RAN (Open Radio Access Network) onto control plane (CP) or edge applications that execute on host computers with hardware accelerators in software defined datacenters (SDDCs). At the CP or edge application operating on a machine executing on a host computer with a hardware accelerator, the method of some embodiments receives data, from an O-RAN E2 unit, to perform an operation. The method uses a driver of the machine to communicate directly with the hardware accelerator to direct the hardware accelerator to perform a set of computations associated with the operation. This driver allows the communication with the hardware accelerator to bypass an intervening set of drivers executing on the host computer between the machine's driver and the hardware accelerator. Through this driver, the application in some embodiments receives the computation results, which it then provides to one or more O-RAN components (e.g., to the E2 unit that provided the data, another E2 unit or another control plane or edge application).
1 . A method of deploying an application in an O-RAN (Open Radio Access Network), the method comprising:
selecting a set of installation files for installing the application on a host computer, the set of installation files comprising a description for configuring passthrough access for the application to a hardware accelerator of the host computer;
using the set of installation files to install the application on the host computer, said using comprising configuring, based on the description, a program executing on the host computer to pass calls from a particular hardware accelerator driver associated with the application to the hardware accelerator without going through an intervening set of one or more drivers for the hardware accelerator that executes on the host computer between the particular hardware accelerator driver and the hardware accelerator;
said configuring allowing the application to bypass the intervening set of drivers when directing the hardware accelerator to perform operations for the application and to receive the results of the operations from the hardware accelerator.
2 . The method of claim 1 , wherein the application performs operations offloaded by an O-RAN E2 unit.
3 . The method of claim 1 , wherein the application executes over the program on the host computer.
4 . The method of claim 3 , wherein the program is an operating system of the host computer.
5 . The method of claim 3 , wherein the program is a hypervisor executing on the host computer.
6 . The method of claim 3 , wherein the program is a virtual machine (VM) and the application operates on a Pod that executes on the VM.
7 . The method of claim 3 , wherein the program is a virtual machine (VM) and the application operates on a container that executes on the VM.
8 . The method of claim 1 further comprising:
prior to performing the selecting and configuring, identifying the host computer from a plurality of host computers as the computer on which the application should be installed,
said identifying comprising:
determining that the application requires a hardware accelerator;
identifying a set of host computers that each comprise a hardware accelerator;
selecting the host computer from the set of host computers.
9 . The method of claim 8 , wherein selecting the host computer comprises
determining that the application will need to communicate with a set of one or more other applications that execute on the selected host computer; and
selecting the host computer as the set of other applications already executes on the host computer, said installation of the application and the set of other applications on the selected host computer reducing communication delay between the application and the set of other applications.
10 . The method of claim 8 , wherein the application is installed on the host computer in the set of host computers in order to install the applications that benefit from hardware acceleration on the set of host computers that offer the hardware acceleration.
11 . The method of claim 1 , wherein the hardware accelerator is one of a graphical processing unit (GPU), field-programmable gate array (FPGA), application-specific integrated circuit (ASIC), and structured ASIC.
12 . The method of claim 1 , wherein the application is a control plane application of the O-RAN.
13 . The method of claim 1 , wherein the application is an edge application of the O-RAN.
14 . A non-transitory machine readable medium storing a program which when executed by at least one processing unit deploys an application in an O-RAN (Open Radio Access Network), the program comprising sets of instructions for:
selecting a set of installation files for installing the application on a host computer, the set of installation files comprising a description for configuring passthrough access for the application to a hardware accelerator of the host computer;
using the set of installation files to install the application on the host computer, said using comprising configuring, based on the description, a program executing on the host computer to pass calls from a particular hardware accelerator driver associated with the application to the hardware accelerator without going through an intervening set of one or more drivers for the hardware accelerator that executes on the host computer between the particular hardware accelerator driver and the hardware accelerator;
said configuring allowing the application to bypass the intervening set of drivers when directing the hardware accelerator to perform operations for the application and to receive the results of the operations from the hardware accelerator.
15 . The non-transitory machine readable medium of claim 14 , wherein the application performs operations offloaded by an O-RAN E2 unit.
16 . The non-transitory machine readable medium of claim 14 , wherein the application executes over the program on the host computer.
17 . The non-transitory machine readable medium of claim 16 , wherein the program is an operating system of the host computer.
18 . The non-transitory machine readable medium of claim 16 , wherein the program is a hypervisor executing on the host computer.
19 . The non-transitory machine readable medium of claim 16 , wherein the program is a virtual machine (VM) and the application operates on a Pod that executes on the VM.
20 . The non-transitory machine readable medium of claim 16 , wherein the program is a virtual machine (VM) and the application operates on a container that executes on the VM.