By Seiji Shintaku, RTP Technology
A global manufacturing company with a large number of remote branch offices with limited network connectivity wanted to consolidate technology back to its main data centers in Michigan, Texas, the Netherlands, Brazil, and Singapore using disk-based backup solutions. One of the primary goals was to reduce its capital expenditure on legacy Windows server support costs, as well as to refresh its infrastructure.
We explored a variety of options, including continuing with the existing technology (which proved to be cost prohibitive) and upgrading the infrastructure to the most recent Windows version (which would have required patch management, hardware support and high maintenance in the remote branches, where technical resources were limited).
The solution was to upgrade to a NAS configuration with technologies from NetApp, for the following reasons:
- resilient technology at low cost in the remote branches
- no need for patch management and maintenance of Windows servers
- offsite backups
- local point-in-time snapshots for file recovery up to several weeks
- virtual filer for DR file service recovery
- removal of tape drives in remote branches and centralizing them in hub site locations
- global presence for support issues
A SnapMirror-to-tape solution was chosen, and deployed globally. Many of the remote branches had limited network connectivity. Utilizing a small LTO-2 tape library with both Fibre Channel and IP connectivity, we controlled the library remotely. Using NetApp's SnapMirror-to-tape solution, the majority of the data in the remote sites was backed up to tape. Once the tape backup was complete, we shipped it back to the hub site and restored it on the hub site filer.
We then established a SnapMirror relationship from the remote sites to the hub sites, and only 4KB block-level changes were sent over the WAN. This was implemented by rolling out NetApp filers to remote branches in North America, South America, Europe, Asia PAC, and limited areas of Africa.
The client had confidence in NetApp's RAID-DP technology and snapshots for the remote offices to act as first-tier data protection. NetApp's RAID-DP calculates parity in two ways: across the drives in a RAID group, and diagonally. This ensures the array can sustain a double-disk failure per RAID group and have no impact to the business. Using RAID-DP technology gives the filer as much resilience as a RAID-10 set, with less disks.
In addition, the client had 35 full sets of backups online locally by means of snapshots. Other NAS vendors use copy-on-first-write snapshot technology, which requires first moving a data block out of the volume to a copy-out area and then rewriting the block, thus incurring I/O penalties. NetApp's allocate-on-write snapshot technology uses pointers to 4KB blocks, and locks down those blocks while allocating new blocks as the file system changes. NetApp's snapshot technology uses zero disk space, is instantaneous, and does not incur any performance penalty.
In addition to using SnapMirror to retain offsite backups, we deployed NetApp's clustering technology in the hub site locations.
Key benefits included a reduction in IT resources and elimination of patch, OS, and hardware management in the remote branches. The solution reduced management costs by avoiding patch management, consolidated tape backup into main hub site locations and out of the remote offices, broke the upgrade cycle of OS and hardware maintenance, and gave the remote branches high resilience with online full backups.
SEIJI SHINTAKU is a principal consultant with RTP Technology Corp., a reseller, systems integration, and professional services organization, www.rtptech.com.