Virtual clusters to provide fault containment
Techniques to process service requests are described herein. In various embodiments, an identifier data associated with a service consumer with which a service request is associated is used to determine a subset comprising fewer than all members of a set of service instances. A selected service instance is selected from among the service instances in the subset to perform the service request.
1. A method of processing service requests, comprising:
receiving, at a load balancer or other service request distribution node, a service request;
extracting an identifier data from the service request, wherein the identifier data includes at least an Internet Protocol (IP) address associated with a service consumer;
using the identifier data associated with the service consumer with which the service request is associated to determine a total number of service instances of a set of service instances to be made available to service requests associated with the service consumer and to determine which service instances are to be included in a subset of service instances, wherein the subset comprises fewer than all members of the set of service instances and corresponds to the total number of service instances to be made available to service requests associated with the service consumer; and
selecting from among the service instances in the subset a selected service instance to perform the service request.
2. The method of claim 1 , wherein using the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset of service instances includes computing a key based at least in part on the identifier data and using the key to look up one or more service instance index values associated with the key.
3. The method of claim 1 , wherein using the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset of service instances includes computing a key based at least in part on the identifier data and using the key to determine dynamically one or more service instance index values associated with the key.
4. The method of claim 3 , wherein using the key to determine dynamically one or more service instance index values associated with the key includes using at least a portion of the key to generate a random number.
5. The method of claim 4 , wherein using at least a portion of the key to generate a random number includes using said at least a portion of the key to seed a pseudorandom number generator.
6. The method of claim 1 , wherein the subset includes a prescribed number k of service instances out of a broader population of n service instances.
7. The method of claim 1 , wherein using the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset comprising fewer than all members of the set of service instances is performed using an algorithm that ensures a same permutation of services instances is included in the subset as would have been determined for a prior service request, if any, from the same service consumer, absent changes in the set of service instances.
8. The method of claim 1 , wherein using the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset of service instances further includes iteratively choosing a service instance from an available pool of service instances that have not already been chosen to be included in the subset.
9. The method of claim 8 , further comprising adding an index or other value associated with each service instance to be included in the subset to a list of service instances that have already been chosen to be included in the subset.
10. The method of claim 9 , further comprising checking the list, for each iteration after a first iteration of choosing the service instance to be included in the subset, to determine whether the next selected service instance index or other value has already been included in the subset.
11. A service request distribution system, comprising:
a communication interface configured to receive data associated with a service request; and
a processor coupled to the communication interface and configured to:
extract an identifier data from the service request, wherein the identifier data includes at least an Internet Protocol (IP) address associated with a service consumer;
use the identifier data associated with the service consumer with which the service request is associated to determine a total number of service instances of a set of service instances to be made available to service requests associated with the service consumer and to determine which service instances are to be included in a subset of service instances, wherein the subset comprises fewer than all members of the set of service instances and corresponds to the total number of service instances to be made available to service the service request associated with the service consumer; and
select from among the service instances in the subset a selected service instance to perform the service request.
12. The system of claim 11 , wherein the processor is configured to use the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset of service instances at least in part by computing a key based at least in part on the identifier data and using the key to look up one or more service instance index values associated with the key.
13. The system of claim 11 , wherein the processor is configured to use the identifier data associated with the service consumer with which the service request is associated to determine which service instances are to be included in the subset of service instances at least in part by computing a key based at least in part on the identifier data and using the key to determine dynamically one or more service instance index values associated with the key.
14. The system of claim 11 , wherein the subset includes a prescribed number k of service instances out of a broader population of n service instances.
15. The system of claim 11 , wherein the processor is configured to determine the subset comprising fewer than all members of the set of service instances at least in part by choosing, iteratively a service instance from an available pool of service instances that have not already been chosen to be included in the subset.
16. A computer program product to process service requests, the computer program product being embodied in a tangible, non-transitory computer readable storage medium and comprising computer instructions for:
receiving a service request;
extracting an identifier data from a service request, wherein the identifier data includes at least an Internet Protocol (IP) address associated with a service consumer;
using the identifier data associated with the service consumer with which the service request is associated to determine a total number of service instances of a set of service instances to be made available to service the service request associated with the service consumer and to determine which service instances are to be included in a subset of service instances, wherein the subset comprises fewer than all members of the set of service instances and corresponds to the total number of service instances to be made available to service the service request associated with the service consumer; and
selecting from among the service instances in the subset a selected service instance to perform the service request.