System, Method, And Service For Providing A Generic RAID Engine And Optimizer

Abstract

A generic RAID engine system accepts an access request, accepts a metadata input comprising a layout description and, optionally, a plurality of resource optimization objectives, accepts a dynamic input comprising a dynamic state of an I / O stack comprising the generic RAID engine and a fault configuration of a plurality of storage devices in the I / O stack, and accepts RAID code input comprising information about the RAID code used by the I / O stack. The metadata input, the dynamic input, and the RAID code input are utilized to transform the access request into individual device reads and individual device writes such that RAID code relationships for the storage devices are maintained at all times. An optional optimizer module selects strategies that meet the resource optimization objectives.

Description

FIELD OF THE INVENTION[0001]The present invention generally relates to data storage. More specifically, the present invention relates to providing fault tolerance for a storage system comprising RAID coded storage volumes.BACKGROUND OF THE INVENTION[0002]Conventional applications requiring reliable data storage typically use commonly available RAID (redundant array of independent disks) levels to protect against data loss due to media or disk failures. Although conventional RAID technology has proven to be useful, it would be desirable to present additional improvements. With a marked rise in the quantity of stored data and no commensurate improvement in disk reliability, a variety of RAID codes are becoming essential to containing data management costs. However, rolling out new RAID codes is challenging as well as cost prohibitive since the new RAID codes require significant development, testing, and tuning efforts.[0003]Until recently, protecting customer data from loss due to med...

AI Technical Summary

Benefits of technology

The present invention provides a system, service, computer program product, and method for a generic RAID engine and optimizer. The system can accept an access request, metadata input, and dynamic input to transform the access request into individual device reads and writes while maintaining RAID code relationships. The system can also select strategies to meet resource optimization objectives. The layout description comprises a description of each disk in the storage devices, while the resource optimization objectives include minimizing disk I / O and memory bus bandwidth. The dynamic input includes information about the data cache and fault configuration. The system can also execute read and write requests, and provide a method for the user to specify input and invoke the generic RAID code engine utility. The technical effects of the present invention include improved performance and efficiency of RAID storage systems.

Problems solved by technology

However, rolling out new RAID codes is challenging as well as cost prohibitive since the new RAID codes require significant development, testing, and tuning efforts.

Only a few RAID codes were supported in traditional RAID implementations (firmware) for a variety of reasons: firmware complexity, software maintainability, and field upgrade difficulty.

Increased complexity increases development and test costs.

The firmware becomes a collection of special cases making it hard to perform path length optimizations.

In addition, performance tuning becomes difficult, if not prohibitively expensive.

. . ” code blocks makes firmware readability harder, making the firmware more prone to bugs.

Upgrading firmware and drivers is not a task that a storage administrator relishes, given the propensity of upgrades to trigger other problems.

No single RAID code satisfies all aspects of data storage.

Using a single RAID code in this application is challenging.

These data sets typically span many disks, indicating a relatively high probability of simultaneous failures.

Furthermore, while disk capacities are increasing, the hard error rate (HER) due to media deterioration, channel errors, etc., has remains relatively constant.

Combined, these trends raise the frequency of data loss events as data sets grow unless RAID codes with a higher fault tolerance are used.

Consequently, providing various data reliability levels using cheaper but less reliable disks requires a greater variety of RAID codes.

RAIDframe provides no simplification or automation of error handling.

Furthermore, RAIDframe has no ability to automatically tune performance.

Further, tuning of this performance approach is only minimally automatic, if at all.

Non-XOR codes such as LDPC and Reed-Solomon have been developed, but these have not become popular since they offer no special advantage over XOR based codes.

An example of such an error path is, on a read, encountering a read error due to a failed sector or disk.

The need for such a solution has heretofore remained unsatisfied.

Images