Addressing the Data Protection Challenges
- Published on Monday, 19 November 2007 21:02
Network storage architectures offer significant potential benefits, but these architectures also bring a higher degree of complexity. Managing and protecting data on a network infrastructure requires a new level of intelligence and new data protection techniques that were not required when the storage resource was tied to a server. These may include auto detection and configuration, as well as sharing of storage resources between multiple servers, or the use of new network protocols to move data between shared storage resources across storage networks.
Not all data protection solutions offer the same level of support for network storage environments, so it is important to evaluate and select a plan that dovetails with one’s data protection needs and network storage architecture. Businesses looking to deploy these network storage architectures should carefully evaluate whether their existing data protection solutions are up to the challenge, and if their solution will allow them to exploit the potential benefits of network storage architectures.
When planning IT investments, data protection shouldn’t be an afterthought, but rather the information bedrock on which a business stands. The entire IT environment must reflect the needs of the business, permitting both data backups and restores when, where and as quickly as is necessary to keep a business running. All efforts should be made to ensure that critical data, whether it is user, accounting or inventory data, remain both available and consistent. Data must be quickly accessible by the application servers and user, it is also – and maybe even more – important to ensure that data remain intact and readily available even in case of a disaster.
Unlike the DAS model, network storage brings a challenging paradigm to data protection applications. This includes new ways to perform traditional data protection tasks, which not only decrease backup and recovery downtime, but help ensure that existing hardware and software be used optimally. Storage is an increasingly complex network resource requiring a data protection solution that can handle this level of complexity. It must also be able to auto-discover storage resources such as SANs and NAS and allow auto configuration of devices (tape, drives libraries). To ensure optimal utilization, the new paradigm must seamlessly share and exploit storage resources on the network. This includes sharing storage resources between multiple servers, tape drives, libraries, and disk.
New Device Protection Protocols Needed
New protocols must be found for protecting the devices that serve these new architectures.
NAS environments typically use two different approaches to backup and recovery. Some NAS devices use common operating systems (e.g. Windows-powered NAS devices) that allow installation of standard backup clients. Other, proprietary systems use open network data management protocol (NDMP). Both allow a user to attach tape devices to NAS systems that can then be managed by a central backup server. These can typically be shared as well, offering similar advantages to SAN implementations. Many NAS vendors allow tape devices to be connected through a SAN, which can further leverage the existing SAN infrastructure. As in most NAS implementations, the majority of data is stored on the NAS disk storage, so including these systems in the data protection scheme is very important.
A SAN typically connects servers and storage devices (such as disk and tape) using either fiber channel (FC) components or Ethernet components IP-based SCSI (iSCSI). These appear as regular direct attached devices to the server operating system and are accessed via standard block oriented storage protocols (SCSI).
Fiber channel based SANs are typically deployed in the data center and facilitate the sharing of storage resources between multiple application servers (e.g., data base servers, mail servers, etc.). iSCSI based SANs allow servers to access remote storage resources, such as for disaster recovery.
iSCSI is still an evolving technology, even though there are many production installations to be found today. iSCSI allows tape devices to be seen as classical SCSI tape devices by the clients or servers that attach to them through the network. This permits tape devices to be easily shared as common network technologies are used in favor of SAN infrastructures. For fast and reliable backup and restore processes, the network being used must support the data transfer rates demanded by business needs. Nevertheless, iSCSI can be a valid lower cost alternative to a SAN for tape device consolidation.
Going Beyond Traditional Attached Storage Options
SAN, NAS, and iSCSI allow different connectivity options that exceed classical direct attached storage options. While some environments still have single tape drives in servers, with administrators changing tapes manually on a regular basis, most users today realize this process must be automated to ensure appropriate data protection levels in a cost-effective way. It is important to consider the differences between storage networks and identify the technology that enables the user to best choose the data protection solution for the specific environment and business process.
A tape library in a SAN environment can easily be connected not only to the backup server but (and even exclusively) to the servers that actually own the data to be backed up. The data transfer can then be performed directly from the application server or file server to the tape library without the need to pass through the backup server. This reduces the traffic on the local area network (LAN) and potentially accelerates critical business processes using the LAN while increasing the backup performance. As tape drives can be shared in a SAN environment, the actual amount of tape drives required in an enterprise can be reduced, as the existing drives are used on demand.
Enter the new generation of client applications – the so-called “smart” clients – which add the intelligence to manage data and connectivity. While smart clients provide the benefits of a rich client model with thin client manageability, they also add more flexibility than traditional rich client applications. This “smart tier” allows users to utilize a device attached locally to a system to control the backup of data.
In short, the data protection solution must be flexible to accommodate different configurations and data protection schemes. For example, using distributed NAS architecture versus a centralized NAS architecture. Since today’s customers will probably use a combination of network storage solutions, data protection solutions should support different architectures.
The singular message to forward thinking companies is that data protection can never be an afterthought. It must be the heart of any mission critical environment, balancing the ever increasing risk of data growth with the TCO required to store and protect it. A sound data protection solution is both efficient and cost-effective, eliminating the “back-up window,” boosting IT efficiencies, reducing administrative and operational costs, and most importantly, ensuring the availability of vital information, the lifeblood of any organization.
Clearly, data protection presents an opportunity to exploit a competitive advantage. Companies must anticipate how a data protection solution fits into their growing environment. This solution must be flexible and adaptable enough to protect heterogeneous platforms, various storage typologies, and provide robust application support. Finally, the IT environment should reflect business needs, enable backup and, more importantly, data restores to ensure that the business not only survives, but succeeds.
Jet Martin serves as director of product management for San Diego–based BakBone Software (www.bakbone.com), an international data protection solution provider that develops and distributes data backup, restore, and disaster recovery software for network storage and open-systems environments worldwide. Send comments to Martin at firstname.lastname@example.org.