Method of adjusting service function chains to improve network performance

Abstract

Some embodiments of the invention provide a method for monitoring and adjusting a service chain that includes several services to perform on data messages passing through a network. For a service chain implemented by a set of service paths each of which includes several service nodes that implement the services of the service chain, the method receives, from a set of service proxies, operational data relating to data transmission characteristics of a set of operational service nodes. The method analyzes the data transmission characteristics. In response to the analysis of the data transmission characteristics, the method alters the set of service paths implementing the service chain.

1 - 30 . (canceled) 
     
     
         31 . A method for monitoring and adjusting a service chain comprising a plurality of services to perform on data messages passing through a network, the method comprising:
 for a service chain implemented by a set of service paths each of which comprises a plurality of service nodes that implement the plurality of services of the service chain, each service node connecting to the network through a service proxy that forwards data messages between the network and the service node:
 receiving, from a plurality of service proxies, operational data relating to data transmission characteristics of a set of operational service nodes in the plurality of service nodes; 
 analyzing the data transmission characteristics; 
 in response to the analysis of the data transmission characteristics, altering the set of service paths implementing the service chain. 
   
     
     
         32 . The method of  claim 31 , wherein altering the set service paths comprises adding a new service path by instantiating a new service node to implement a particular service of the service chain for the new service path. 
     
     
         33 . The method of  claim 32 , wherein instantiating the new service node is in response either (i) to operational data indicating that a throughput of one or more service nodes implementing the particular service is beyond a threshold throughput, or (ii) to operational data indicating that a latency of one or more service nodes implementing the particular service is beyond a threshold latency. 
     
     
         34 . The method of  claim 31 , wherein altering the service paths comprises eliminating a service path by deactivating an existing service node that is implementing a particular service for the service path. 
     
     
         35 . The method of  claim 34 , wherein deactivating the existing service node is in response either (i) to operational data indicating that a throughput of one or more service nodes implementing the particular service is beyond a threshold throughput, or (ii) to operational data indicating that a latency of one or more service nodes implementing the particular service is beyond a threshold latency. 
     
     
         36 . The method of  claim 31 , wherein altering the service paths comprises restarting an existing service node that is implementing a particular service. 
     
     
         37 . The method of  claim 36 , wherein the restarting is in response either (i) to operational data indicating errors in the existing service node, or (ii) to operational data indicating that the existing service node has fallen below a performance threshold. 
     
     
         38 . The method of  claim 31 , wherein a particular service node implements parts of multiple service paths of the service chain. 
     
     
         39 . The method of  claim 38 , wherein altering the service paths comprises reducing a number of service paths implemented in part by the particular service node. 
     
     
         40 . The method of  claim 31 , wherein altering the service paths comprises increasing a number of service paths implemented in part by a particular service node. 
     
     
         41 . The method of  claim 31 , wherein a particular service node and a particular service proxy for the particular service node execute on one host computer. 
     
     
         42 . The method of  claim 31 , wherein the operational data includes operational statistics derived from the data transmission characteristics received from the set of service proxies. 
     
     
         43 . The method of  claim 42 , wherein analyzing the data transmission characteristics comprises aggregating operational statistics data with previously collected operational statistics data. 
     
     
         44 . The method of  claim 43 , wherein the aggregating comprises calculating a weighted sum of the operational statistics data. 
     
     
         45 . The method of  claim 42 , wherein the operational statistics comprises data transmission characteristics of the service nodes received from the service proxies. 
     
     
         46 . The method of  claim 45 , wherein the data transmission characteristics of a particular service node comprise at least one of a throughput of the particular service node, and a latency time for the particular service node. 
     
     
         47 . The method of  claim 31 , wherein the service proxies collects data relating to packets received by their associated service nodes. 
     
     
         48 . A non-transitory machine-readable medium storing a program which when executed by at least one processing unit monitors and adjusts a service chain comprising a plurality of services to perform on data messages passing through a network, the program comprising sets of instructions for:
 for a service chain implemented by a set of service paths each of which comprises a plurality of service nodes that implement the plurality of services of the service chain, each service node connecting to the network through a service proxy that forwards data messages between the network and the service node:
 receiving, from a plurality of service proxies, operational data relating to data transmission characteristics of a set of operational service nodes in the plurality of service nodes; 
 analyzing the data transmission characteristics; 
 in response to the analysis of the data transmission characteristics, altering the set of service paths implementing the service chain. 
   
     
     
         49 . The non-transitory machine-readable medium of  claim 48 , wherein the set of instructions for altering the set service paths comprises a set of instructions for adding a new service path by instantiating a new service node to implement a particular service of the service chain for the new service path. 
     
     
         50 . The non-transitory machine-readable medium of  claim 49 , wherein instantiating the new service node is in response either (i) to operational data indicating that a throughput of one or more service nodes implementing the particular service is beyond a threshold throughput, or (ii) to operational data indicating that a latency of one or more service nodes implementing the particular service is beyond a threshold latency.