Browse Prior Art Database

Storage Cluster Allocation Using Cluster Free Lists and Persistent Messaging

IP.com Disclosure Number: IPCOM000019570D
Publication Date: 2003-Sep-19
Document File: 6 page(s) / 64K

Publishing Venue

The IP.com Prior Art Database

Abstract

The present invention is an asynchronous dynamic volume mapping apparatus that decouples cluster management from cluster allocation in a networked storage system. The apparatus uses a multiprocessor architecture that divides volume mapping functions between processor elements to enable asynchronous operation and a cluster free list to store cluster pointers for use in allocating storage resources. The performance of the system is further improved by the incorporation of persistent messaging. Persistent messaging provides greater system reliability by providing mirrored copies of the metadata that describes the last system configuration in the event of a failure.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 27% of the total text.

STORAGE CLUSTER ALLOCATION USING CLUSTER FREE LISTS AND PERSISTENT MESSAGING

Robert Horn and Virgil Wilkins, Aristos Logic Corporation.

27051 Towne Centre Drive, Suite 290, Foothill Ranch, CA 92610

Abstract of the Invention

The present invention is an asynchronous dynamic volume mapping apparatus that decouples cluster management from cluster allocation in a networked storage system. The apparatus uses a multiprocessor architecture that divides volume mapping functions between processor elements to enable asynchronous operation and a cluster free list to store cluster pointers for use in allocating storage resources. The performance of the system is further improved by the incorporation of persistent messaging. Persistent messaging provides greater system reliability by providing mirrored copies of the metadata that describes the last system configuration in the event of a failure.

Description of the Invention

The present invention is a process for asynchronously and dynamically allocating storage volumes in a storage system using multiple storage controller processors in conjunction with cluster free lists. There may be a single cluster free list or multiple cluster free lists. For the purposes of this description, a cluster represents a defined size of physical storage space. The sizes are determined at the storage system design level to provide desired granularity and optimize data storage management. For example, a cluster size may be 1 megabyte, 16 terabytes, or 4 kilobytes. The use of cluster free lists enables persistent messaging, which further provides redundancy for system control data. In the event of a power failure, the system recovers and continues from the last known state saved in the persistent message.


Figure 1) asynchronous volume mapping apparatus with cluster free list.

Figure 1 is a diagram of a volume mapping apparatus 100. Volume mapping apparatus 100 includes a host 110 and a storage controller 120. Host 110 is any system that utilizes storage services from storage controller 120 via a data interconnect 190. Storage controller 120 further includes a cluster manager processor 130, a cluster free list 140 with a tail 150 and a head 160, a cluster allocator processor 170, and a volume map 180. The cluster management function is decoupled from the cluster allocation operation, and thus allows cluster management processor 130 to perform various cluster management operations independently of cluster allocator processor 170. Independent processing facilitates an asynchronous mode of operation.

Cluster manager processor 130 is a dedicated processor for executing storage system-level cluster allocation policies within storage controller 120. Cluster manager processor 130 is responsible for detecting new storage elements coupled to the networked storage system and for grouping new elements according to specific algorithms outlined in the system architecture. Cluster manager processor 130 detects, at a storage system level, when...