Tenant

As explained in Designing Deep Learning Systems: A guide for software engineers

A customer organization that uses a shared application while maintaining a separate account and segregated data.

A tenant is a company or organization using the same application or service as other customers while retaining a separate account and segregated data. section 6.3.2

In a multitenant chatbot, each tenant can train a model on its own dataset. Requests are routed to that tenant’s model, improving fit for its business situation. section 6.3.2

A shared model-serving design provides:

• A model cache that keeps model graphs and dependent data in memory. section 6.3.2
• A model file server that supplies the models used by the service. section 6.3.2
• One prediction service capable of serving multiple tenant models of the same type. section 6.3.2

Tenant-specific intent classification

A school and a retail store share a chatbot application, but each has a separate account, dataset, and intent classification model. The application identifies the tenant and uses that tenant’s model for prediction. section 6.3.2

Figure 6.6 A prediction service with model caching for multitenant applications

Shows a prediction service using model caching to serve multiple tenant models. section 6.3.2

How Designing Deep Learning Systems: A guide for software engineers covers it

How deep it goes

Defines tenant identity and explains model routing, caching, and file serving without detailing implementation interfaces.

Examples

Uses schools and retail stores sharing a chatbot, separate datasets, tenant-specific models, and a cached prediction service.

What it assumes

Assumes familiarity with chatbot requests and intent classification while defining tenant accounts, data segregation, and model ownership.

What you can do after

Operate/configure a shared prediction service that selects tenant-specific models and caches their graphs and dependent data.

How it connects

Connects tenants with multitenancy, model serving, intent classification, model caching, and segregated data.

liveBookmanning.com

Read Designing Deep Learning Systems: A guide for software engineers on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Software Security for DevelopersA tenant is an independently scoped organization or customer population sharing an identity system while retaining separate users, configuration, and access boundaries.
• 

  liveBookmanning.com
  Platform Engineering on KubernetesA tenant is a team, customer, or other independent consumer that shares a platform while requiring separated access and resources.
• 

  liveBookmanning.com
  Just Use Postgres!A tenant is an independent customer or merchant whose data and operations are isolated within a shared multitenant system.

Compare how 4 books explain tenant →

Related concepts

• multitenancyuses · Multitenancy is an architecture in which multiple independent tenants share an application or underlying infrastructure while their data, resources, or access remain logically separated.
• policy bundlehas part · A policy bundle is a validated, packaged collection of authorization policies prepared for versioned deployment and runtime evaluation.

This explainer was generated by Manning's knowledge graph from the text of Designing Deep Learning Systems: A guide for software engineers. The section links point to the parts of the book it draws on.