Techniques for artificial intelligence capabilities at a network switch

A method implemented using a data processing system, the data processing system to be used in association with a cloud service provider system, at least one network, and multiple tenants, the cloud service provider system being configurable to comprise artificial intelligence (AI) processing resources, the AI processing resources comprising AI inference resources that comprise graphics processing unit (GPU) resource circuitry, the data processing system comprising interface circuitry and management circuitry, the interface circuitry to be communicatively coupled, via the at least one network, to the multiple tenants, the method comprising:

• processing, by the management circuitry, tenant service-related data, the tenant service-related data being configurable to correspond, at least in part, to AI service request data and AI service results data associated with the multiple tenants, the AI service request data and the AI service results data to be received by and transmitted from, respectively, the management circuitry via the interface circuitry, respective portions of the tenant service-related data being associated with respective ones of the multiple tenants;
• and
• generating, by the management circuitry, based upon management data, configuration data, the configuration data to be used in association with configuring at least certain of AI inference resources to implement, at least in part, AI inference services associated, at least in part, with the multiple tenants, the AI inference services being accessible by the multiple tenants based, at least in part, upon the tenant service-related data;
• wherein:
• the management data is associated with AI service registration data, the AI service registration data being associated with the AI inference services and the multiple tenants;
• the AI inference resources are to be assigned, at least in part, for use in implementing, at least in part, the AI inference services based, at least in part, upon service performance requirements data of the AI service registration data;
• the service performance requirement data is to indicate performance requirements to be fulfilled in providing the AI inference services;
• the respective ones of the multiple tenants are to be associated with respective portions of tenant identification data;
• the respective portions of the tenant identification data are to be associated with portions of the service performance requirements data;
• the configuration data is to be routed to at least certain of the AI processing resources so as to permit at least certain of the multiple tenants to access, based upon at least certain of the tenant service-related data, at least certain AI services associated with the respective ones of the multiple tenants;
• and
• the data processing system satisfies one or more of:
• the GPU resource circuitry comprises multiple GPU processing cores;
• the AI inference services are to be implemented, at least in part, using one or more neural networks, neural processing, and/or tensor processing to be implemented, at least in part, using the GPU resource circuitry and/or the compute resources;
• and/or
• the respective portions of the tenant identification data are configurable to be associated with respective portions of tenant billing data to be generated in association with the providing of the AI inference services.

At least one non-transitory machine readable storage medium storing instructions to be executed by at least one machine that is to be associated with a data processing system, the data processing system to be used in association with a cloud service provider system, at least one network, and multiple tenants, the cloud service provider system being configurable to comprise artificial intelligence (AI) processing resources, the AI processing resources comprising AI inference resources that comprise graphics processing unit (GPU) resource circuitry, the data processing system comprising interface circuitry and management circuitry, the interface circuitry to be communicatively coupled, via the at least one network, to the multiple tenants, the instructions, when executed by the at least one machine, resulting in performance of operations comprising:

• processing, by the management circuitry, tenant service-related data, the tenant service-related data being configurable to correspond, at least in part, to AI service request data and AI service results data associated with the multiple tenants, the AI service request data and the AI service results data to be received by and transmitted from, respectively, the management circuitry via the interface circuitry, respective portions of the tenant service-related data being associated with respective ones of the multiple tenants;
• and
• generating, by the management circuitry, based upon management data, configuration data, the configuration data to be used in association with configuring at least certain of AI inference resources to implement, at least in part, AI inference services associated, at least in part, with the multiple tenants, the AI inference services being accessible by the multiple tenants based, at least in part, upon the tenant service-related data;
• wherein:
• the management data is associated with AI service registration data, the AI service registration data being associated with the AI inference services and the multiple tenants;
• the AI inference resources are to be assigned, at least in part, for use in implementing, at least in part, the AI inference services based, at least in part, upon service performance requirements data of the AI service registration data;
• the service performance requirement data is to indicate performance requirements to be fulfilled in providing the AI inference services;
• the respective ones of the multiple tenants are to be associated with respective portions of tenant identification data;
• the respective portions of the tenant identification data are to be associated with portions of the service performance requirements data;
• the configuration data is to be routed to at least certain of the AI processing resources so as to permit at least certain of the multiple tenants to access, based upon at least certain of the tenant service-related data, at least certain AI services associated with the respective ones of the multiple tenants;
• and
• the data processing system satisfies one or more of:
• the GPU resource circuitry comprises multiple GPU processing cores;
• the AI inference services are to be implemented, at least in part, using one or more neural networks, neural processing, and/or tensor processing to be implemented, at least in part, using the GPU resource circuitry and/or the compute resources;
• and/or
• the respective portions of the tenant identification data are configurable to be associated with respective portions of tenant billing data to be generated in association with the providing of the AI inference services.

A cloud service provider system for use in association with at least one network and multiple tenants, the cloud service provider system being configurable to comprise artificial intelligence (AI) processing resources, the AI processing resources comprising AI inference resources that comprise graphics processing unit (GPU) resource circuitry, the cloud service provider system comprising:

• a data processing system comprising:
• interface circuitry to be communicatively coupled, via the at least one network, to the multiple tenants;
• and
• management circuitry to:
• process tenant service-related data, the tenant service-related data being configurable to correspond, at least in part, to AI service request data and AI service results data associated with the multiple tenants, the AI service request data and the AI service results data to be received by and transmitted from, respectively, the management circuitry via the interface circuitry, respective portions of the tenant service-related data being associated with respective ones of the multiple tenants;
• and
• generate, based upon management data, configuration data, the configuration data to be used in association with configuring at least certain of AI inference resources to implement, at least in part, AI inference services associated, at least in part, with the multiple tenants, the AI inference services being accessible by the multiple tenants based, at least in part, upon the tenant service-related data;
• wherein:
• the management data is associated with AI service registration data, the AI service registration data being associated with the AI inference services and the multiple tenants;
• the AI inference resources are to be assigned, at least in part, for use in implementing, at least in part, the AI inference services based, at least in part, upon service performance requirements data of the AI service registration data;
• the service performance requirement data is to indicate performance requirements to be fulfilled in providing the AI inference services;
• the respective ones of the multiple tenants are to be associated with respective portions of tenant identification data;
• the respective portions of the tenant identification data are to be associated with portions of the service performance requirements data;
• the configuration data is to be routed to at least certain of the AI processing resources so as to permit at least certain of the multiple tenants to access, based upon at least certain of the tenant service-related data, at least certain AI services associated with the respective ones of the multiple tenants;
• and
• the cloud service provider system satisfies one or more of:
• the GPU resource circuitry comprises multiple GPU processing cores;
• the AI inference services are to be implemented, at least in part, using one or more neural networks, neural processing, and/or tensor processing to be implemented, at least in part, using the GPU resource circuitry and/or the compute resources;
• and/or
• the respective portions of the tenant identification data are configurable to be associated with respective portions of tenant billing data to be generated in association with the providing of the AI inference services.

At least one data center system for use in association with multiple tenants, the at least one data center system comprising:

• at least one network;
• and
• a cloud service provider system configurable to comprise artificial intelligence (AI) processing resources, the AI processing resources comprising AI inference resources that comprise graphics processing unit (GPU) resource circuitry, the cloud service provider system comprising:
• a data processing system comprising:
• interface circuitry to be communicatively coupled, via the at least one network, to the multiple tenants;
• and
• management circuitry to:
• process tenant service-related data, the tenant service-related data being configurable to correspond, at least in part, to AI service request data and AI service results data associated with the multiple tenants, the AI service request data and the AI service results data to be received by and transmitted from, respectively, the management circuitry via the interface circuitry, respective portions of the tenant service-related data being associated with respective ones of the multiple tenants;
• and
• generate, based upon management data, configuration data, the configuration data to be used in association with configuring at least certain of AI inference resources to implement, at least in part, AI inference services associated, at least in part, with the multiple tenants, the AI inference services being accessible by the multiple tenants based, at least in part, upon the tenant service-related data;
• wherein:
• the management data is associated with AI service registration data, the AI service registration data being associated with the AI inference services and the multiple tenants;
• the AI inference resources are to be assigned, at least in part, for use in implementing, at least in part, the AI inference services based, at least in part, upon service performance requirements data of the AI service registration data;
• the service performance requirement data is to indicate performance requirements to be fulfilled in providing the AI inference services;
• the respective ones of the multiple tenants are to be associated with respective portions of tenant identification data;
• the respective portions of the tenant identification data are to be associated with portions of the service performance requirements data;
• the configuration data is to be routed to at least certain of the AI processing resources so as to permit at least certain of the multiple tenants to access, based upon at least certain of the tenant service-related data, at least certain AI services associated with the respective ones of the multiple tenants;
• and
• the at least one data center system satisfies one or more of:
• the GPU resource circuitry comprises multiple GPU processing cores;
• the AI inference services are to be implemented, at least in part, using one or more neural networks, neural processing, and/or tensor processing to be implemented, at least in part, using the GPU resource circuitry and/or the compute resources;
• and/or
• the respective portions of the tenant identification data are configurable to be associated with respective portions of tenant billing data to be generated in association with the providing of the AI inference services.