Tailored relief for congestion on application servers for real time communications

A method for managing load on a server for real-time communications on a packet-switched network, comprising:

• receiving, on a particular intermediate node of a real-time communications packet-switched network, usage data that indicates an amount of computational resources on the particular intermediate node consumed by a set of one or more application layer processes for a first type of data packets processed at the particular intermediate node, wherein the first type of data packets is an application layer protocol message type;
• determining whether the amount of computational resources consumed indicated by the usage data exceeds a processing threshold amount associated with proximity to performance degradation at the particular intermediate node, the processing threshold amount being below a processing budget for the particular intermediate node and the computational resources being processing resources;
• determining a particular source of data packets of a second type;
• and
• if it is determined that the amount of computational resources consumed exceeds the processing threshold amount, then sending to a first different node on the real-time communications packet-switched network a message that includes overload state data that indicates:
• the particular source of the data packets processed at the particular intermediate node;
• a reduction request to reduce an amount of the data packets that are sent to the particular intermediate node from the particular source for processing at the particular intermediate node;
• determining the second type that is both different from the first type and lower priority than the first type at the particular intermediate node;
• and
• increasing the processing threshold amount associated with performance degradation at the particular intermediate node for the first type.

An apparatus for real-time communications on a packet-switched network, comprising:

• a network interface that is coupled to a real-time communications packet-switched network for communicating therewith a data packet;
• one or more processors;
• a computer-readable medium;
• and
• one or more sequences of instructions stored in the computer-readable medium, which, when executed by the one or more processors, causes the one or more processors to perform:
• receiving usage data that indicates an amount of computational resources on the apparatus consumed by a set of one or more application layer processes for a first type of a plurality of types of data packets processed at the apparatus for real-time communications on a packet switched network, wherein the first type of the plurality of types of data packets is an application layer protocol message type;
• determining a particular source of data packets of a second type of the plurality of types of data packets processed at the apparatus from among a plurality of sources of the data packets of the second type;
• determining whether the amount of computational resources consumed indicated by the usage data exceeds a central processing unit threshold amount associated with proximity to performance degradation at the apparatus;
• and
• if it is determined that the amount of computational resources consumed exceeds the threshold amount, then sending to a first different node on the real-time communications packet-switched network a message that includes overload state data that indicates the second type, and a reduction request to reduce an amount of data packets of the second type that are sent to the apparatus, wherein the processing threshold amount is an amount below a central processing unit budget for a particular intermediate node and the computational resources being central processing unit resources;
• and
• determining the second type that is both different from the first type and lower priority than the first type at the particular intermediate node;
• and
• increasing the central processing unit threshold amount associated with performance degradation at the particular intermediate node for the first type.

A non-transitory computer-readable medium carrying one or more sequences of instructions for managing load on a server for real-time communications on a packet-switched network, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform:

• receiving, on a particular intermediate node of a real-time communications packet-switched network, usage data that indicates an amount of computational resources on the particular intermediate node consumed by a set of one or more application layer processes for a first type of a plurality of types of data packets processed at the particular intermediate node, wherein the first type of the plurality of types of data packets is an application layer protocol message type;
• determining whether the amount of computational resources consumed indicated by the usage data exceeds a central processing unit threshold amount associated with proximity to performance degradation at the particular intermediate node, the processing threshold amount being an amount below a central processing unit budget for the particular intermediate node and the computational resources being central processing unit resources;
• determining a particular source of data packets of a second type of the plurality of types of data packets processed at the particular intermediate node from among a plurality of sources of the data packets of the second type;
• if it is determined that the amount of computational resources consumed exceeds the threshold amount, then sending to a first different node on the real-time communications packet-switched network a message that includes overload state data that indicates the second type, and a reduction request to reduce an amount of data packets of the second type that are sent to the particular intermediate node;
• determining the second type that is both different from the first type and lower priority than the first type at the particular intermediate node;
• and
• increasing the central processing unit threshold amount associated with performance degradation at the particular intermediate node for the first type.