Application computation offloading for mobile edge computing
Systems, apparatuses, methods, and computer-readable media, are provided for offloading computationally intensive tasks from one computer device to another computer device taking into account, inter alia, energy consumption and latency budgets for both computation and communication. Embodiments may also exploit multiple radio access technologies (RATs) in order to find opportunities to offload computational tasks by taking into account, for example, network/RAT functionalities, processing, offloading coding/encoding mechanisms, and/or differentiating traffic between different RATs. Other embodiments may be described and/or claimed.
1. An edge computing device operable in an edge computing system, comprising:
processor circuitry configured to:
identify a set of available radio access technologies that provide the edge computing device with connectivity to a network of the edge computing system;
determine power consumption parameters for respective radio access technologies of the set of available radio access technologies;
identify application requirements of at least one application task of at least one application, the application requirements configured to indicate requirements to perform computational offloading of the at least one application to respective nodes available using respective radio access technologies of the set of available radio access technologies;
select a radio access technology from the set of available radio access technologies based on the power consumption parameters of the selected radio access technology; and
select a node of the respective nodes for computational offloading based on the application requirements, wherein the computational offloading causes transfer of the at least one application task to the selected node for execution using the selected radio access technology.
2. The edge computing device of claim 1 , further comprising processor circuitry configured to:
determine network characteristics for the network that indicate at least one of:
channel state information of the network at the respective nodes, backhaul state information for the respective nodes, an average data rate at the respective nodes, or an average round trip time (RTT) at the respective nodes;
determine resource parameters for the network that indicate at least one of:
a computational capacity at the respective nodes, currently available computational load at the respective nodes, a security level at the respective nodes, or a re-use degree of computational resources at the respective nodes; and
wherein the application requirements indicate at least one of:
a frequency at which the at least one application task is to be offloaded, a computational load for executing the at least one application task, an amount of data to be transferred for the computational offloading, or an amount of data to be obtained from a node after execution of the at least one application task.
3. The edge computing device of claim 2 , wherein the processor circuitry is to:
determine, for the respective nodes, a computation latency, communication latency, a computation energy consumption, and a communication energy consumption based on the network characteristics and the application requirements.
4. The edge computing device of claim 3 , wherein the processor circuitry is further configured to:
determine, for the respective nodes, a latency budget based on the computation latency and the communication latency; and
determine, for the respective nodes, an energy consumption budget based on the computation energy consumption and the communication energy consumption.
5. The edge computing device of claim 1 , wherein operations to select the node for the computational offloading, includes operations to:
select the node according to an offloading configuration, wherein the offloading configuration indicates that a selection of the node is to be based on:
a lowest latency budget among the respective nodes,
a lowest energy consumption budget among the respective nodes,
a lowest latency budget among a set of the respective nodes having an energy consumption budget that is less than an energy consumption threshold, or
a lowest energy consumption budget among a set of the respective nodes having s latency budget that is less than a latency threshold.
6. The edge computing device of claim 1 , wherein the respective nodes comprise a plurality of mobile edge hosts (MEHs) in the edge computing system, the edge computing device further comprising:
network interface circuitry to communicate with the respective nodes, wherein individual MEHs of the plurality of MEHs are located at or near a corresponding access node of a plurality of access nodes, and wherein the edge computing system is adapted to operate according to an European Telecommunications Standards Institute (ETSI) Multi-Access Edge Computing (MEC) specification.
7. The edge computing device of claim 6 , wherein the processor circuitry is further configured to:
obtain, over a first reference point, an application offloading request message from a user equipment (UE), wherein the application offloading request message is to request an identity of an individual node on which to offload application tasks; and
send, over the first reference point, an application offloading report message to the UE, wherein the application offloading report message is to indicate the selected node.
8. At least one non-transitory machine-readable medium comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations to:
identify a set of available radio access technologies that provide an edge computing device with connectivity to a network of an edge computing system;
determine power consumption parameters for respective radio access technologies of the set of available radio access technologies;
identify application requirements of at least one application task of at least one application, the application requirements configured to indicate requirements to perform computational offloading of the at least one application to respective nodes available using respective radio access technologies of the set of available radio access technologies;
select a radio access technology from the set of available radio access technologies based on the power consumption parameters of the selected radio access technology; and
select a node of the respective nodes for computational offloading based on the application requirements, wherein the computational offloading causes transfer of the at least one application task to the selected node for execution using the selected radio access technology.
9. The at least one non-transitory machine-readable medium of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to:
determine network characteristics for the network that indicate at least one of:
channel state information of the network at the respective nodes, backhaul state information for the respective nodes, an average data rate at the respective nodes, or an average round trip time (RTT) at the respective nodes;
determine resource parameters for the network that indicate at least one of:
a computational capacity at the respective nodes, currently available computational load at the respective nodes, a security level at the respective nodes, or a re-use degree of computational resources at the respective nodes; and
wherein the application requirements indicate at least one of:
a frequency at which the at least one application task is to be offloaded, a computational load for executing the at least one application task, an amount of data to be transferred for the computational offloading, or an amount of data to be obtained from a node after execution of the at least one application task.
10. The at least one non-transitory machine-readable medium of claim 9 , further comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations to:
determine, for the respective nodes, a computation latency, communication latency, a computation energy consumption, and a communication energy consumption based on the network characteristics and the application requirements.
11. The at least one non-transitory machine-readable medium of claim 10 , further comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations to:
determine, for the respective nodes, a latency budget based on the computation latency and the communication latency; and
determine, for the respective nodes, an energy consumption budget based on the computation energy consumption and the communication energy consumption.
12. The at least one non-transitory machine-readable medium of claim 8 , wherein the instructions to select the node for the computational offloading, includes instructions to:
select the node according to an offloading configuration, wherein the offloading configuration indicates that a selection of the node is to be based on:
a lowest latency budget among the respective nodes,
a lowest energy consumption budget among the respective nodes,
a lowest latency budget among a set of the respective nodes having an energy consumption budget that is less than an energy consumption threshold, or
a lowest energy consumption budget among a set of the respective nodes having s latency budget that is less than a latency threshold.
13. The at least one non-transitory machine-readable medium of claim 8 , wherein the respective nodes comprise a plurality of mobile edge hosts (MEHs) in the edge computing system, the at least one non-transitory machine-readable medium further comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations to:
communicate with the respective nodes, wherein individual MEHs of the plurality of MEHs are located at or near a corresponding access node of a plurality of access nodes, and wherein the edge computing system is adapted to operate according to an European Telecommunications Standards Institute (ETSI) Multi-Access Edge Computing (MEC) specification.
14. A method, executable by an edge computing device, comprising:
identifying a set of available radio access technologies that provide the edge computing device with connectivity to a network of an edge computing system;
determining power consumption parameters for respective radio access technologies of the set of available radio access technologies;
identifying application requirements of at least one application task of at least one application, the application requirements configured to indicate requirements to perform computational offloading of the at least one application to respective nodes available using respective radio access technologies of the set of available radio access technologies;
selecting a radio access technology from the set of available radio access technologies based on the power consumption parameters of the selected radio access technology; and
select a node of the respective nodes for computational offloading based on the application requirements, wherein the computational offloading causes transfer of the at least one application task to the selected node for execution using the selected radio access technology.
15. The method of claim 14 , further comprising:
determining network characteristics for the network that indicate at least one of:
channel state information of the network at the respective nodes, backhaul state information for the respective nodes, an average data rate at the respective nodes, or an average round trip time (RTT) at the respective nodes;
determining resource parameters for the network that indicate at least one of:
a computational capacity at the respective nodes, currently available computational load at the respective nodes, a security level at the respective nodes, or a re-use degree of computational resources at the respective nodes; and
wherein the application requirements indicate at least one of:
a frequency at which the at least one application task is to be offloaded, a computational load for executing the at least one application task, an amount of data to be transferred for the computational offloading, or an amount of data to be obtained from a node after execution of the at least one application task.
16. The method of claim 15 , further comprising:
determining, for the respective nodes, a computation latency, communication latency, a computation energy consumption, and a communication energy consumption based on the network characteristics and the application requirements.
17. The method of claim 16 , further comprising:
determining, for the respective nodes, a latency budget based on the computation latency and the communication latency; and
determining, for the respective nodes, an energy consumption budget based on the computation energy consumption and the communication energy consumption.
18. The method of claim 14 , wherein selecting the node for the computational offloading, further comprises:
selecting the node according to an offloading configuration, wherein the offloading configuration indicates that a selection of the node is to be based on:
a lowest latency budget among the respective nodes,
a lowest energy consumption budget among the respective nodes,
a lowest latency budget among a set of the respective nodes having an energy consumption budget that is less than an energy consumption threshold, or
a lowest energy consumption budget among a set of the respective nodes having s latency budget that is less than a latency threshold.
19. The method of claim 14 , wherein the respective nodes comprise a plurality of mobile edge hosts (MEHs) in the edge computing system, and further comprising:
communicating with the respective nodes, wherein individual MEHs of the plurality of MEHs are located at or near a corresponding access node of a plurality of access nodes, and wherein the edge computing system is adapted to operate according to an European Telecommunications Standards Institute (ETSI) Multi-Access Edge Computing (MEC) specification.
20. The method of claim 19 , further comprising:
obtaining, over a first reference point, an application offloading request message from a user equipment (UE), wherein the application offloading request message is to request an identity of an individual node on which to offload application tasks; and
sending, over the first reference point, an application offloading report message to the UE, wherein the application offloading report message is to indicate the selected node.