microsoft / kernel-memory

RAG architecture: index and query any data using LLM and natural language, track sources, show citations, asynchronous memory patterns.
https://microsoft.github.io/kernel-memory
MIT License
1.34k stars 252 forks source link

Async Kernel Memory Service doesn't support Multitenancy #676

Closed backendmaestro1 closed 1 week ago

backendmaestro1 commented 1 week ago

Context / Scenario

Async Kernel Memory Service currently lacks multitenancy support. This means each tenant (individual user or organization) having separate storage account and vector database would require separate instances of kernel memory service instead of having one KM Service catering to all tenants. This lack of resource sharing (in this case the KM service) )leads to increased operational complexity and resource consumption.

What happened?

Async Kernel Memory Service fails to isolate data and resources between different tenants. This lack of multitenancy support prevents the efficient utilization of the service in environments with multiple users or organizations.

Importance

The inability to use Kernel Memory efficiently due to the lack of multitenancy support is a significant issue for users who require cost-effective and scalable solutions. This limitation hinders the adoption of the service in shared or cloud-based scenarios where resource sharing is crucial.

Platform, Language, Versions

Platform: Microsoft Azure Language: C# (the programming language used to interact with the service) Versions: Latest and older versions of KM

Additional Information

Impact: The lack of multitenancy support limits the scalability and cost-effectiveness of the Async Kernel Memory Service in multi-tenant environments.

Workaround: Potential workarounds could involve manually provisioning and managing separate service instances for each tenant, but this would be cumbersome and expensive.

Desired Solution: A fix for the bug that enables true multitenancy support in the Async Kernel Memory Service, allowing secure and isolated data access for multiple tenants within a shared environment.