Browse Prior Art Database

A method for emulating a Storage Area Network (SAN) using Network Attached Storage (NAS)

IP.com Disclosure Number: IPCOM000022533D
Original Publication Date: 2004-Mar-19
Included in the Prior Art Database: 2004-Mar-19
Document File: 5 page(s) / 46K

Publishing Venue

IBM

Abstract

A method for emulating a Storage Area Network (SAN) using Network Attached Storage (NAS)

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 47% of the total text.

Page 1 of 5

A method for emulating a Storage Area Network (SAN) using Network Attached Storage (NAS)

Disclosed is a method for emulating a Storage Area Network using Network Attached Storage. A Storage-Area-Network (SAN) is a special-purpose, high-performance network that interconnects disks and servers, and facilitates disk sharing. When developing software for such a storage network, there is a need to cheaply and quickly test the software without setting up a real SAN, as SANs are expensive as well as complex and difficult to set up. It would also be of advantage to easily emulate a very large SAN, i.e., a SAN with a very large number of disks. In addition, it is desirable to easily access the SAN's data using regular file tools, for example for debugging purposes.

One known solution is IBM's Virtual Shared Disk (VSD). VSD is software that allows a server with direct-attached disks to share its disks with other servers, so it in fact also emulates a SAN. However, the disks in the VSD are physical disks. So if a very large emulated SAN is required, i.e., a SAN with a very large number of disks, there is still a need to physically install these disks and also disk adaptors and attach them to the servers.

There are two parts for the solution. First, each emulated SAN disk is represented by a single ordinary file in a directory on a plain-vanilla Network-Attached-Storage (NAS) appliance (e.g., an Network-File-System (NFS) server). These files are exported to all the servers that wish to participate in the emulated SAN. On each server, a device driver that associates a virtual disk device with a backing file is used. Thus, each imported file is associated with a virtual disk. Using this two-layer approach, all servers have shared access to the emulated disks and an emulated SAN is thus created. Since each disk is represented by an ordinary file, a further advantage is that since it is easy to create many files and these can be sparse, it is easy to emulate a large number of disks using relatively small disk space.

BASIC INVENTION

Implementing the invention is done by the following steps:

For each disk in the emulated SAN, the user creates a file on a network file system or

distributed file system (such as Network File System (NFS), Server Message Block (SMB) or Andrew File System (AFS)) hosted by a NAS appliance. These are ordinary files, but are used to emulate disks, and thus are named DBFs (disk-backing files). For simplicity, it is assumed that all the DBFs are created within the same directory, but this is not a requirement. The DBFs can be sparse. The length of the DBF is also the size of the emulated disk which it represents. The user can create many files to represent many disks.

Here is an example of how to create a sparse DBF on UNIX-style machines:

% dd if=/dev/zero of=dbf1 seek=262144k bs=1 count=0


1.

1

Page 2 of 5

The above command creates a sparse file "dbf1" of length 256MB, which is used to represent an emulated disk of size 256...