Systems and methods for predictive scaling of edge entities

A non-transitory computer readable medium comprising instructions which, when executed by a processor, enables the processor to perform a method for scaling a virtual network function (VNF) by analyzing VNF metrics data, the method comprising:

• obtaining, by a computing device, the VNF metrics data for a service function chain (SFC) executing on a second computing device,
• wherein the SFC comprises a first VNF, a second VNF, and a third VNF;
• determining, using the VNF metrics data, maximum processing times for the first VNF, the second VNF, and the third VNF;
• obtaining a constraint for the SFC;
• calculating a constraint violation probability for the SFC,
• wherein the constraint violation probability is expressed through a negative exponential function at least reflecting the maximum processing times for the first VNF, the second VNF, and the third VNF;
• making a first determination that the constraint violation probability is greater than an upper threshold;
• based on the first determination:
• making a second determination that the second VNF has a highest maximum processing time relative to the first VNF and the third VNF;
• and
• based on the second determination, scaling-up the second VNF,
• wherein scaling up the second VNF comprises initiating allocation of additional hardware resources on the second computing device to the second VNF.

A computing device, comprising:

• a processor;
• and
• memory comprising instructions which, when executed by the processor, enables the processor to perform a method for scaling a virtual network function (VNF) by analyzing VNF metrics data, the method comprising:
• obtaining the VNF metrics data for a service function chain (SFC) executing on a second computing device,
• wherein the SFC comprises a first VNF, a second VNF, and a third VNF;
• determining, using the VNF metrics data, maximum processing times for the first VNF, the second VNF, and the third VNF;
• obtaining a constraint for the SFC;
• calculating a constraint violation probability for the SFC,
• wherein the constraint violation probability is expressed through a negative exponential function at least reflecting the maximum processing times for the first VNF, the second VNF, and the third VNF;
• making a first determination that the constraint violation probability is greater than an upper threshold;
• based on the first determination:
• making a second determination that the second VNF has a highest maximum processing time relative to the first VNF and the third VNF;
• and
• based on the second determination, scaling-up the second VNF,
• wherein scaling up the second VNF comprises sending a command to a container orchestrator that is managing the second VNF to scale up the second VNF.