Object Store Federation with Unified Namespace and Single-Hop Access

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object storage systems lack a unified namespace, scalability, and high availability, leading to inefficiencies in cross-federation traffic and limited scalability across data centers.

Innovation Solution

A federation architecture that creates a single uniform namespace across multiple object store instances, allowing clients to access buckets and objects with a single hop, and ensures high availability through a distributed consensus protocol backed by a global metadata service.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a unified namespace is implemented across multiple object store instances, then accessibility and ease of operation are improved, but device complexity increases

Engineering Contradiction:
ImproveaccessibilityVSAvoidsystem complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

A global metadata service is introduced as an intermediary component that manages the unified namespace across multiple object store instances. This service maintains mapping information between global namespace identifiers and local object store identifiers, allowing clients to access objects across federated stores through a single namespace without directly managing the complexity of multiple distributed systems. The intermediary absorbs the complexity of cross-federation routing and resolution.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If object storage systems are federated across multiple data centers, then scalability is improved, but device complexity increases

Engineering Contradiction:
ImprovescalabilityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The object storage system is segmented into independent object store instances that can be distributed across multiple data centers, each maintaining its own local metadata. The global metadata service segments the namespace management function from the actual data storage, allowing the system to scale by adding new object store instances without requiring centralized management of all metadata. This segmentation enables independent scaling of storage capacity and metadata management.

Inventive Principle:
Principle #1Segmentation

3Loss of energy

If cross-federation traffic is reduced through unified namespace, then loss of energy is improved, but device complexity increases

Engineering Contradiction:
Improvecross-federation trafficVSAvoidsystem complexity
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

The global metadata service performs preliminary resolution of global namespace identifiers to local object store identifiers before data access operations. By pre-establishing and caching these mappings, the system avoids unnecessary cross-federation traffic during data access, as clients can directly access local object stores once the global namespace is resolved. This preliminary action reduces redundant network communication across federation boundaries.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12596682B2System and method for object store federation
Publication Date: 2026.04.07 NUTANIX INC
  • US12596682B2 patent drawing
  • US12596682B2 patent drawing
  • US12596682B2 patent drawing

AI summary

In some aspects, an apparatus includes a processor and a memory. The memory can include programmed instructions that, when executed by the processor, cause the apparatus to receive a request to create a bucket. The memory can also include programmed instructions that, when executed by the processor, cause the apparatus to, in response to receiving the request to create the bucket, write, to a first object store, a first mapping that maps an identifier of global metadata to an identifier of a second object store where the bucket is to be created, write, to the second object store, a second mapping that maps the identifier of global metadata to local metadata, and replicate, to the second object store, the first mapping.