Scalable and robust network management for cloud-based NAT environments

A method executed by a network management system (NMS) comprising:

• receiving a request to establish a management session from a managed element with a first DCM of a pool of device communication managers (DCMs), each of the DCMs executed by one or more processors of the NMS as a software container, the managed element comprising one of a plurality of managed elements within a network;
• issuing, via an application programming interface (API) gateway, a remote procedure call (RPC) from the first DCM to a first device operations manger (DOM) of a pool of DOM executed by the one or more processors of the NMS, each of the DOMs executed by the processors of the NMS as a software container, and each of the DOMs configured to present an API for performing operations on each of the managed elements according to device identifiers of the managed elements;
• establishing, with the first DOM, a persistent application-layer communication session from the first DOM to the first DCMs;
• directing, with the first DOM, commands from a set of one or more NMS applications to the first DCMs over the persistent application-layer communication session according to a mapping between device identifiers associated with the managed elements and network addresses associated with the DCMs;
• and
• issuing the commands from the first DCM to the managed element via the management session.

A computer-readable storage medium comprising instructions that, when executed, cause a processor of a network management system (NMS) to:

• receive a request to establish a management session from a managed element with a first one of a pool of device communication managers (DCMs), each of the DCMs executed by one or more processors of the NMS as a respective software container;
• issue, via an application programming interface (API) gateway, a remote procedure call from the first DCM to a first one of a pool of device operations managers (DOMs) executed by the one or more processors of the NMS, each of the DOMs executed by the one or more processors of the NMS as a respective software container, and each of the DOMs configured to present an API for performing operations on the managed elements according to device identifiers of the managed elements;
• establish, with the first DOM, a persistent application-layer communication session from the first DOM to the first DCM;
• direct, with the first DOM, commands from a set of one or more NMS applications to the first DCM over the persistent application-layer communication session according to a mapping between device identifiers associated with the managed elements and network addresses associated with the DCMs;
• and
• issue the commands from the first DCM to the managed element via the management session.

A network system comprising:

• a plurality of managed elements within a network;
• a network management system comprising:
• one or more hardware-based processors;
• a set of one or more network management system (NMS) applications configured to execute on the processors;
• a pool of device communication managers (DCMs), each of the DCMs executed by the processors as a software container, and each of the DCMs configured to accept and manage a management session for one of the plurality of managed elements;
• a pool of device operations managers (DOMs), each of the DOMs executed by the processors as a software container, and each of the DOMs configured to present an application programming interface (API) for performing operations on the plurality of managed elements according to device identifiers of the plurality of managed elements;
• an API gateway configured to route remote procedure calls (RPCs) from the DCMs to the DOMs via APIs exposed by the DOMs and according to device identifiers of the plurality of managed elements as specified within the RPCs,
• wherein the DOMs are configured to establish a set of persistent application-layer communication sessions from the DOMs to the DCMs, and
• wherein the DOMs are configured to direct communications from the NMS applications to the DCMs over the persistent application-layer communication sessions according to a mapping between device identifiers associated with the plurality of managed elements and network addresses associated with the DCMs;
• and
• a network address translation (NAT) device positioned between the NMS and the plurality of managed elements.