Intelligently pre-positioning and migrating compute capacity in an overlay network, with compute handoff and data consistency
Edge server compute capacity demand in an overlay network is predicted and used to pre-position compute capacity in advance of application-specific demands. Preferably, machine learning is used to proactively predict anticipated compute capacity needs for an edge server region (e.g., a set of co-located edge servers). In advance, compute capacity (application instances) are made available in-region, and data associated with an application instance is migrated to be close to the instance. The approach facilitates compute-at-the-edge services, which require data (state) to be close to a pre-positioned latency-sensitive application instance. Overlay network mapping (globally) may be used for more long-term positioning, with short-duration scheduling then being done in-region as needed. Compute instances and associated state are migrated intelligently based on predicted (e.g., machine-learned) demand, and with full data consistency enforced.
1. A method of distributed edge computing, comprising:
configuring application instances on edge nodes of an overlay network in advance of demand for compute capacity on the edge nodes, wherein the demand for compute capacity on the edge nodes is determined by collaborative machine learning wherein at least first and second edge nodes have associated therewith local machine learning models of computing demand that are distinct from one another, and wherein at least the first edge node adjusts its local machine learning model via transfer learning based on the local machine learning model received from the second edge node;
associating a client to a first of the application instances;
responsive to movement of the client, associating the client to a second of the application instances via a compute hand-off; and
maintaining data consistency with respect to interactivity among the client and the first and second application instances;
wherein, for a given edge node, the compute capacity comprises an amount of processing resources sufficient to support one or more of the application instances configured on the given edge node;
wherein at least one of the application instances is latency-sensitive and the client is a mobile device.
2. The method as described in claim 1 , wherein the overlay network is a content delivery network (CDN).
3. The method as described in claim 1 , wherein the mobile device is roaming in a radio access network coupled to the overlay network.
4. The method as described in claim 1 , wherein an edge node of the overlay network provides a mobile edge computing (MEC) function.
5. The method as described in claim 1 , wherein the application instances are pre-positioned on the edge nodes where the demand for compute capacity is anticipated based on historical data.
6. A computer program product comprising a non-transitory computer readable medium holding computer program code executable by a hardware processor to facilitate distributed edge computing, the computer program code configured to:
configure application instances on edge nodes of an overlay network in advance of demand for compute capacity on the edge nodes, wherein the demand for compute capacity on the edge nodes is determined by collaborative machine learning wherein at least first and second edge nodes have associated therewith local machine learning models of computing demand that are distinct from one another, and wherein at least the first edge node adjusts its local machine learning model via transfer learning based on the local machine learning model received from the second edge node;
associate a client to a first of the application instances;
responsive to movement of the client, associate the client to a second of the application instances via a compute hand-off; and
maintain data consistency with respect to interactivity among the client and the first and second application instances;
wherein, for a given edge node, the compute capacity comprises an amount of processing resources sufficient to support one or more of the application instances configured on the given edge node:
wherein at least one of the application instances is latency-sensitive and the client is a mobile device.
7. The computer program product as described in claim 6 , wherein an edge node provides a mobile edge computing (MEC) function.
8. The computer program product as described in claim 6 , wherein the overlay network is a content delivery network (CDN).