FUSE Object Storage Device for Cluster File System Tiering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional Lustre cluster file systems face challenges in balancing storage capacity and IO throughput requirements, leading to suboptimal performance or excessive costs due to mismatched storage devices and IO operations.

Innovation Solution

The integration of a FUSE object storage device that provides a transparent flash storage tier within the cluster file system, allowing dynamic balancing of storage capacity and IO throughput without requiring significant changes to clients, object storage servers, or applications, through the use of a burst buffer appliance for storage tiering control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional Lustre implementations use traditional disk storage devices for object storage servers, then storage capacity is achieved, but IO throughput performance is suboptimal

Engineering Contradiction:
ImproveIO throughputVSAvoidstorage device configuration
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the storage system into multiple tiers: a first storage tier using fast flash storage devices for high IO throughput, and a second storage tier using capacity-optimized disk storage devices. The FUSE object storage device enables this segmentation by providing a virtual file system layer that can present different storage characteristics to clients, allowing the system to achieve both high performance and adequate capacity without requiring all storage devices to be complex high-performance units.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The FUSE object storage device acts as an intermediary layer between the clients and the heterogeneous storage devices. It provides a unified namespace and handles the complexity of managing multiple storage tiers with different characteristics. This intermediary abstracts the underlying storage complexity from clients while enabling optimized IO paths to different storage tiers, thereby improving overall IO throughput without requiring clients to directly manage complex storage configurations.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If high-performance flash storage devices are deployed to improve IO throughput, then performance is enhanced, but system cost becomes excessive

Engineering Contradiction:
ImproveIO throughputVSAvoidstorage cost
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent applies local quality by assigning different storage device types to different functional requirements: flash storage devices are used locally in the first storage tier to handle performance-critical IO operations, while disk storage devices are used in the second tier for capacity-optimized storage. The FUSE object storage device enables this differentiation by allowing clients to access high-performance flash storage for active data while storing less frequently accessed data on cheaper disk devices, thereby achieving enhanced IO throughput for critical operations without the excessive cost of deploying flash storage across the entire system.

Inventive Principle:
Principle #3Local quality

3Adaptability or versatility

If storage tiers with different device types are implemented, then dynamic balancing of capacity and performance is achieved, but device complexity increases

Engineering Contradiction:
Improvestorage tiering flexibilityVSAvoidmulti-tier configuration
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The FUSE object storage device provides universality by implementing a unified interface that can access both flash storage devices and disk storage devices through a common protocol and namespace. This multi-functional capability allows the system to dynamically balance between storage capacity and IO throughput by routing operations to appropriate storage tiers without requiring different interfaces or complex client-side configuration for each device type. The universal FUSE interface simplifies the multi-tier configuration from the client perspective while enabling sophisticated storage management on the server side.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS9858290B1Cluster file system with a fuse file system interface
Publication Date: 2018.01.02 EMC IP HLDG CO LLC
  • US9858290B1 patent drawing
  • US9858290B1 patent drawing
  • US9858290B1 patent drawing

AI summary

Cluster file systems are provided having a FUSE interface. A cluster file system comprises a FUSE object storage device that communicates with at least one object storage server of the cluster file system; and provides an interface between the cluster file system and at least one additional file system having a FUSE implementation. The additional file system optionally includes a burst buffer appliance. The burst buffer appliance can be configured, for example, to control the movement of data between first and second storage tiers. The burst buffer appliance can optionally communicate with a plurality of clients over the network and process requests from the clients. The FUSE object storage device provides an interface to, for example, a parallel log-structured file system and/or a Lustre clustered file system.