Distributed caching cluster client configuration

摘要

A cache or other cluster is configuration-aware such that initialization and changes to the underlying structure of the cluster can be dynamically updated for use by a client. A client may use a client driver as an intermediary that is responsible for managing the communication with the cluster. For example, a client driver may resolve an alias from a static configuration endpoint to a storage node. The client driver may request an initial configuration from the storage node and then update configuration from one or more storage nodes that store current configuration of the cluster.

主权项

1. A computer-implemented method for a distributed memory caching configuration, comprising:
receiving, by a client device from a first memory caching node in a cache cluster, a current configuration of the cache cluster, the cache cluster comprising a set of memory caching nodes, each of the memory caching nodes in the set of memory caching nodes comprising a reserved memory space that stores the current configuration, the set of memory caching nodes being part of a first network, the client device being part of a second network that is separate from the first network, and the reserved memory space being exempt from rules of a caching protocol; defining, by the client device, a key space for each memory caching node in the set of memory caching nodes in the cache cluster based at least in part on the configuration; requesting, by the client device utilizing the caching protocol, a first set of values corresponding to a first set of keys from the cache cluster using the key space to determine which memory caching node to contact for each key of the first set of keys; monitoring, by the client device, responses from the cache cluster; determining, from the monitoring, that a latency of a response from a memory caching node has increased beyond a threshold; transmitting, by the client device utilizing an extension of the caching protocol, a provisioning request for a new memory caching node to be added to the cache cluster, the transmitting of the provisioning request being based at least in part on the determining that the latency has increased beyond the threshold, and the extension of the caching protocol being configured to modify the reserved memory space; receiving, by the client device from a second memory caching node in the cache cluster, an updated configuration of the cache cluster identifying changes to the cache cluster, the second memory caching node being different from the first memory caching node in the cache cluster, and the updated configuration being received independent from a request by the client device; determining a revised key space for each memory caching node in the set of memory caching nodes in the cache cluster affected by the updated configuration; and requesting a second set of values corresponding to a second set of keys from the cache cluster using the revised key space to determine which memory caching node to contact for each key of the second set of keys.