System and method for network and computation performance probing for edge computing
Described herein are systems and methods of for measuring and assuring performance in networked applications where edge computing is utilized, which may comprise determining or improving network and computational performance by the use of edge probing. Edge probing enables rapid measurement and assessment of networking and computational performance, where performance comprises speed latency, for a device using an edge compute node. Edge probing can be used to assess performance of and assign converged networking and computing infrastructure to applications and devices.
1. A method of probing edge computing performance, comprising:
generating a plurality of edge probes that facilitate generation of time information related to a plurality of edge compute nodes;
transmitting, from a transmitter, a first edge probe among the plurality of edge probes to a first edge compute node and transmitting a second edge probe among the plurality of edge probes to a second edge compute node, the first and second edge compute nodes being in the plurality of edge compute nodes;
receiving, at a receiver, from the first edge compute node a first response corresponding to the first edge probe and receiving from the first edge compute node a second response corresponding to the second edge probe, the first response comprises network time information of the first edge compute node and edge node compute time information of the first edge compute node, the second response comprises network time information of the second edge compute node and edge node compute time information of the second edge compute node, the edge node compute time information is related to time due to computations occurred within the first or second edge compute node; and
selecting a preferred edge compute node based at least in part on time information derived from the first and second responses, the preferred edge compute node being within the plurality of edge compute nodes.
2. The method of claim 1 wherein each of the plurality of edge probes specifies one or more edge computing tests for an edge compute node.
3. The method of claim 1 wherein the time information is a measure or estimate of time delay.
4. The method of claim 1 wherein the first or second response from the first or second edge compute node includes a timestamp.
5. The method of claim 1 wherein a computation is performed, with a computation time delay being part of a round-trip response delay.
6. The method of claim 1 wherein a networking delay is further estimated by networking tests.
7. The method of claim 1 wherein the computation is designed to provide uniformly comparable computing times that enable comparisons between performances of the first and the second edge compute nodes.
8. The method of claim 1 wherein the first and the second edge probes and their responses are used to estimate one or more of speed, latency, jitter, statistical distributions, averages, statistics, loss, availability, reliability, repeatability, throughput, scalability, privacy, security, efficiency, computing speed, computing time, determinacy, repeatability, scalability, throughput, CPU, graphics, memory, disk, or network performance of the first edge compute node and the second edge compute node.
9. A network device comprising:
an edge probe generator that generates a plurality of edge probes that facilitate generation of time information related to a plurality of edge compute nodes;
a transmitter coupled to the edge probe generator, the transmitter transmits a first edge probe among the plurality of edge probes to a first edge compute node and transmits a second edge probe among the plurality of edge probes to a second edge compute node, the first and second edge compute nodes being in the plurality of edge compute nodes;
a receiver that receives from the first edge compute node a first response corresponding to the first edge probe and receives from the second edge compute node a second response corresponding to the second edge probe, the first response comprises network time information and edge node compute time information of the first edge compute node, the second response comprises network time information and edge node compute time information of the second edge compute node, the edge node compute time information is related to time due to computations occurred within the first or second edge compute node; and
an edge compute node selector that selects a preferred edge compute node based at least in part on time information derived from the first and second responses, the preferred edge compute node being within the plurality of edge compute nodes.
10. The network device of claim 9 wherein each one of the plurality of edge probes specifies one or more edge computing tests for an edge compute node.
11. The network device of claim 9 wherein the edge node compute time information at the first edge compute node and the edge node compute time information at the second edge compute node correlate to one or more performance characteristics of the first and second edge compute nodes relative to a particular type of application.
12. The network device of claim 9 wherein the preferred edge compute node is selected by determining which edge compute node within the plurality of edge compute nodes has desired time information.
13. The network device of claim 12 wherein the desired time information is the shortest latency or delay.
14. The network device of claim 9 wherein the preferred edge compute node is selected by determining which edge compute node within the plurality of edge compute nodes provides a highest performance associated with a particular application.
15. The network device of claim 14 wherein the preferred edge compute node is selected based on at least one of an estimation of jitter, computed statistical distributions related to performance, calculated averages related to performance and measured determinism of computation times.