Technologies for providing service isolation, scalability, and proactive tenant migration in multi-tenant ecosystems
Assigneesalesforce.com, inc.
InventorsMahanthi Gangadhar, Vijaya Saradhi Pula
Life of this patent
Estimating the term…
Abstract
Systems, methods, and computer-readable media for performance monitoring and tenant migration in multi-tenant and/or cloud computing ecosystems are described. Embodiments may include technologies for scaling multi-tenant systems based on input load characteristics, proactive monitoring and tenant migration in multi-tenant environments, and service isolation in a multi-tenant environments. Other embodiments may be described and/or claimed.
Claims20
One or more non-transitory computer readable media (NTCRM) comprising instructions, wherein execution of the instructions by one or more processors of a master node of a plurality of compute nodes is operable to cause the master node to:
- identify a set of performance metrics of an ecosystem instance of a set of ecosystem instances in which one or more tenants in a multi-tenant system (MTS) reside, each ecosystem instance of the set of ecosystem instances comprising a set of service tiers to provide services for a discrete subset of tenants in the MTS, each set of service tiers being isolated from other sets of service tiers of other ecosystem instances, the set of ecosystem instances being operated by a plurality of secondary nodes among the plurality of compute nodes;
- determine an input load of the identified ecosystem instance based on the set of performance metrics;
- detect a trigger for generating one or more other ecosystem instances based on the input load;
- generate, in response to detection of the trigger, one or more other ecosystem instances based on the performance metrics, each of the one or more other ecosystem instances to include individual service tier sets that are isolated from one another and from the service tiers of the ecosystem instance;
- and
- migrate at least one tenant of the discrete subset of tenants from the ecosystem instance to another ecosystem instance of the one or more other ecosystem instances, and a service tier set of the other ecosystem instance is to provide services to the at least one tenant.
A multi-tenant system (MTS) to provide cloud computing services, the multi-tenant system comprising:
- a plurality of server computing systems to be employed as a plurality of compute nodes, each of the plurality of servers including at least one processor and memory, the plurality of compute nodes comprising a manager node and a plurality of secondary nodes;
- the plurality of secondary nodes are configurable to:
- operate a set of ecosystem instances in which one or more tenants of the multi-tenant system reside, each ecosystem instance of the set of ecosystem instances comprises a set of service tiers to provide corresponding services to the one or more tenants, and each set of service tiers are isolated from other sets of service tiers of the set of ecosystem instances;
- and
- the manager node is configurable to manage the set of ecosystem instances, and, to manage the set of ecosystems, the manager node is to:
- identify a set of performance metrics of an ecosystem instance of the set of ecosystem instances,
- determine an input load of the ecosystem instance based on the set of performance metrics,
- detect a trigger for generating one or more other ecosystem instances based on the input load,
- initiate generation of one or more other ecosystem instances of the set of ecosystem instances in response to detection of the trigger, and
- migrate at least one tenant from the ecosystem instance to the one or more other ecosystem instances.
A method to be performed by a manager node of a plurality of compute nodes, the method comprising:
- monitoring, by the manager node, a set of performance metrics of an ecosystem instance, the ecosystem instance comprising a set of service tiers to provide services for tenants of a multi-tenant system;
- determining, by the manager node, an input load of the ecosystem instance based on the monitored set of performance metrics;
- adjusting, by the manager node, a size of the ecosystem instance when the input load reaches an auto-scaling threshold;
- generating, by the manager node, one or more other ecosystem instances when a spin-off threshold is reached, each of the one or more other ecosystem instances to include individual service tier sets that are isolated from one another and from the service tiers of the ecosystem instance;
- migrate one or more tenants residing in the ecosystem instance to the one or more other ecosystem instances;
- and
- configure the one or more other ecosystem instances after the migrating such that traffic is routed to the one or more other ecosystem instances for consumption by corresponding ones of the one or more tenants.
Dependencies are inferred from the claim text.