OUT OF THE SHADOWS … INTO THE LIGHTFrom STORAGE Magazine
Vol 7, Issue 7 - October 2007 Increasingly, both large and smaller sites are using hybrid strategies to protect their data, such as local hard disk backup from initial storage, followed by off-site tape for geographical redundancy. If the first hard disk fails, there's a second on-site copy to fall back on. But should the whole site succumb to a disaster, there's still a tape copy somewhere else. This is called D2D2T - or Disk-to Disk-to-Tape - backup. Sometimes also referred to as 'virtual tape', D2D2T takes advantage of the speed of disk by providing a short-term buffering or staging area for backup data prior to the data being written to tape. Its importance cannot be overstated in today's business environment where critical business information continues to grow at a staggering pace, as indeed do the regulations for compliance and governance. According to Steven S. Scott, VP and general manager, Breece Hill, protecting this data through its lifecycle is rapidly becoming the single most daunting task facing businesses today. He argues that one way businesses of all sizes can reduce total cost and improve efficiency is by implementing or changing to a D2D2T data protection strategy. "D2D2T allows businesses to execute a strategy that exploits the best
attributes of disk and tape when creating data protection solutions," he says.
"It provides businesses with a viable strategy of data protection that addresses
two core aspects of data protection - business continuity and disaster
recovery." According to Mike Walters, consulting systems engineer, NetApp: "Increasingly, companies do not want to rely on tape as their first path to recovering data. A recent NetApp webcast survey indicated that over 85 per cent of customers are looking to disk to accelerate the backup process and make it more reliable. There are various approaches that can be taken to meet these requirements." The first is to deploy a virtual tape library (VTL), he says. "These are designed to integrate into existing backup environments, as they simply 'look like tape' to the backup server. Being disk, however, they have significant performance and reliability advantages over traditional tape mechanisms. In addition, the random-access nature of VTL disk means that recoveries can easily co-exist with backups, removing the risk of possible device contention." In a disk-to-disk-to-tape architecture, a VTL will be deployed to push tape farther away from being a primary recovery mechanism and therefore it is important for customers to examine how easily VTL solutions can integrate with tape, he argues. "Do they keep costs down by maximising the use of physical tape? Can they take the tape processing away from the backup server to improve performance? In particular, what is the management needed to ensure the VTL performs as expected under load? It is rare for companies to have idle administrators and, if a VTL is used in conjunction with tape, it is essential it is easy to manage and self-tune." The second mechanism he suggests is to deploy solutions that integrate more intelligently with primary storage. "These types of solution only move changed data from the primary disk to the secondary disk, so instead of an incremental backup taking 10 per cent of data on a daily basis, the customer only has to replicate perhaps 0.5-1 per cent of data instead - a huge saving in disk and/or tape costs. “This positive impact of requiring less physical storage for each backup means that customers have the ability to keep far more backups online and accessible. This, in turn, means faster recovery. If the backup is across a WAN (wide area network) link, this can also have significant benefits in cost and performance,” he argues. "Organisations looking to deploy a disk-to-disk-to-tape architecture should
always examine the 'recovery point objective'," states Walters, "or from how far
back they typically need to recover information for the data being protected, in
addition to the speed of the backup process." The goal, he concludes, should be
"to craft a solution that enables them to efficiently keep backups online [ie,
on disk storage] to meet the vast majority of their recovery requirements - this
may be to hold data online for a week or perhaps 3-12 months. Only then can they
architect the appropriate solution to match those needs." "It's actually very hard for us to comment on which disk strategy to use. Whether companies opt to use virtual tape libraries, virtual tape, disk to disk to tape - we don't have any preference. However, what we always will suggest is to then archive off-site on tape. Keep in mind that for the SMB customer, especially the smaller sized companies, tape is more than adequate to securely and reliably back up and archive data." When you get to certain industries with specific requirements (for instance, retail, banking, healthcare, etc), she concedes, there might be a need to retrieve data more quickly, at which time disk can be implemented for intermediary backup and fast recovery. "I believe the problem is that there are so many different technologies and
ways to do disk-based backup that, for the SMB customer, it's often hard to see
the wood for the trees. Each customer needs to take a detailed look at his or
her specific requirements before implementing any backup solution. There are
many vendors and resellers who can help them select the most appropriate
solution, without 'over-selling'. It is always good to consider future needs and
data growth, but taking a realistic view of the requirements can save customers
a lot of money." "As a result, every night many companies back up to tape and transport the tapes to the recovery site," says Ian Masters, UK sales and marketing director at Double-Take Software. "What they don't understand is just how vulnerable their data and their business remain to threats, even after such a huge outlay of administrative effort and cost. "Given the ever-increasing demand for continuous data access, it is hard for companies to complete nightly backups to tape within the time window provided. Disk to disk backup and recovery is more reliable and less complex, so takes less time. This greatly improves the recovery point [amount of data the company is prepared to lose] and recovery time [amount of time it takes to regain access to business-critical data and applications]." It's important to consider not just data, but the server and applications that run the data. And with new technology available, it is possible to protect an entire production server, including its operating systems, applications and the data itself. With this entirely stored on disk, restoring the business to full operation is far quicker, adds Masters. "Advances in technology have made disk to disk backup easier to manage and less of a drain on system resources, including bandwidth. Replicating data to a central location, using software solutions, can reduce the cost associated with tape-only solutions and provide a higher level of recoverability for business-critical systems and data. "Many companies choose to incorporate a further line of defence by making a
tape backup of the data and system information stored on the recovery system.
This ensures that the production servers do not slow or go offline during the
backup window and that they have two full copies of their data." For mission-critical applications that require quick restore/retrieval times, disk is the better choice, he acknowledges. "However, not all applications are mission critical and have speedy retrieval access requirements. IT administrators should evaluate integrating disk with their current tape-based backup infrastructure to provide quick restores for those applications that demand it. With the use of a VTL, for example, storage administrators can implement disk without requiring changes to their existing processes, as this form of disk-based backup presents itself as tape." Including disk in the backup strategy makes sense for many companies, but most can't afford to keep more than a few days of their backup data on fast recovery media - the cost of conventional RAID arrays is simply too high. "However, the introduction of data de-duplication technology fundament- ally changes the role of disk in backup," he points out. "It gives users a cost-effective way to retain their backup data on disk for weeks or months, making their file restores fast and easy, and increasing the number of available recovery points. The advantage of enabling replication is a much more reliable disaster recovery process." Another option in the field of D2D2T is to implement data management software to generate revenue faster and store more data at a lower cost. "For long-term data storage, data management software automatically moves data between different disk and tape resources to reduce costs and protect content," Mackey adds. "Data location is virtualised, so that any file can easily be accessed for re-use, even if it resides on tape. "Also, the ability of tape to extend beyond running costs and performance issues to environmental and efficiency benefits is a point worth remembering. By appropriately using tape for storage, many data centres can significantly reduce power consumption, making the data centre more environmentally friendly and freeing up the limited supply of power for use elsewhere." Implementing a tiered storage architecture leverages hybrid disk/tape solutions. This coupling can offer the best of both worlds: fast retrieval from disk for mission-critical applications and cost-effective storage on tape for all other applications and long-term archive. ST |
|
The products referenced in this site are
provided by parties other than BTC. BTC makes no representations regarding
either the products or any information about the products. Any questions,
complaints, or claims regarding the products must be directed to the appropriate
manufacturer or vendor. Click here for usage terms
and conditions.
©2006 Business and Technical Communications Ltd. All rights
reserved. |