Backup & Disaster Recovery
Overview
Many companies face the challenge of protecting rapidly growing data dispersed throughout the organization and even in multiple remote locations. This expansion, resulting from years of server proliferation and decentralization has dramatically increased IT spending. Improving and automating consolidation of data, backup, and disaster recovery eliminates many of the hidden management and administrative costs of delivering IT services to distributed users.
Steady improvements in hardware and software technologies and the reduction in communications bandwidth costs have made data mirroring and replication a very practical alternative to many of the backup and recovery processes that have been used in the past. Network Attached Storage (NAS) has also become widely accepted as the lowest Total Cost of Ownership (TCO) alternative for controlling storage costs. It has been a natural platform to use in enhancing the process of insuring data availability.
Developments in iSCSI over the recent years have brought this relatively young technology to the forefront. iSCSI is an elegantly simple standard protocol to allow the transfer of standard SCSI storage device commands over standard Ethernet infrastructure. This allows systems with the appropriate software to transfer data between them with SCSI block commands. By adding iSCSI software to NAS systems, they can now manage data in block form as well as file formats.
MDI’s iSCSI enabled Express Stor NAS allows the user to automatically provide for multiple levels of data protection by the use of iSCSI mirroring, file snapshots, real-time replication with fail-over, and traditional backup. The following steps illustrate a possible implementation of this economical and easy-to-use technology.
Existing Configuration:
The configuration in to the right is commonly found in many establishments, with perhaps a variation in the quantity of each type of server. Many times, particularly in remote offices, each server has it’s own tape drive and an administrative person is responsible for a periodic backup of the data. Tapes are then sent to a vault or a central location. Sometimes they are verified for recovery capability. Often, they are not, leaving the recovery process at risk.
Step 1, Add NAS:
The addition of a NAS into the configuration, allows one or more file servers to be retired, reducing costs and offsetting the cost of the new NAS system. Users now are mapped to the NAS Server for file storage, and a variety of backup utilities can be implemented to also backup client workstation disks. The NAS system becomes the focal point for file data and the base system for additional functionality.
Step 2, Implement iSCSI:
The MDI NAS system has iSCSI software pre-installed. By installing Microsoft’s iSCSI initiator (a free add-on) on the SQL and Exchange servers along with file mirroring software provided by MDI, the entire configuration 
is iSCSI ready. By provisioning an ISCSI LUN (disk image) for each of the SQL and Exchange servers on the MDI NAS, a new disk volume will appear to the Volume Manager on SQL and Exchange Servers. Via standard Windows commands, these volumes can be initialized and assigned a drive letter. Files on the SQL and Exchange servers that the user wishes to protect can be selected and with a simple process can be mirrored to the respective LUN. After an initial mirror operation, all writes to the SQL and Exchange data files will be mirrored to the LUN on the NAS, providing a real-time copy of the data. If either server is lost, the LUN can be re-provisioned to another server (same or different location) and access to the data is restored.
Step 3, Activate Snapshot on the NAS:
The MDI NAS has built-in snapshot capability. The administrator can elect to create snapshots on a periodic basis of all of the files on the NAS (including the LUN’s) to enable point-in-time recovery of data. This enables rollback to prior states for both traditional files and the SQL and Exchange LUN’s. At this point, a live copy of SQL and Exchange data is kept locally, as are shapshots of normal data files and LUN’s. All of these functions are enabled to occur automatically. If desired, end-users can be given the capability to restore prior versions of their files themselves with a couple of mouse clicks.
Step 4, Implement Replication:
The next step is to replicate the data contained on the MDI NAS to an off-site (potentially central office) location. The off-site NAS can be the target for replication from multiple locations. This NAS unit is set up in a similar manner to the Express Stor in the diagram for Step 3. It can provide the same functionality for servers in its location and be a target for the replication of data from other locations.
Replication is automatically performed to the off-site system over a Wide Area Network. The NAS units are in constant contact, and if the first NAS should fail, the off-site NAS would automatically assume its identity so that file service users would not be aware of the failure. After notification, the Administrator can re-provision the copies of the SQL and Exchange LUN’s on the off-site NAS back to the original SQL and Exchange servers and the mirroring of the data from these servers would resume to the off-site unit. Traditional tape backup can be accomplished at the off-site or central location from the NAS.
Summary:
This example was designed to describe the functionality of MDI’s iSCSI Enabled Express Stor NAS appliance. It provides for a high level of automation, cost savings, data redundancy and the capability to minimize service disruptions caused by various outages. The risk of data loss is virtually eliminated and there is minimal requirement for expensive, labor-intensive traditional back-up processes.