Disk-to-Disk Backup Is Coming?
The Arrival Of Affordable Disk-To-Disk (D2D) Storage - Part 1
by Diamond Lauffin
wwpi.com
Editor's Note: Tape vs. Disk for primary backup is a controversial area. Responses to this commentary are encouraged (mark_ferelli@wwpi.com). The second part of this article appeared in the December issue of Computer Technology Review. Click here for Part 2.
Historically, tape is the most widely used medium for backup and recovery of data stored on servers, disk arrays, and workstations. IT managers have used tape for archival purposes and have for years attempted to implement Hierarchical Storage Management (HSM) systems using this somewhat uncooperative medium. Since tape is used for such a variety of storage functions, in organizations of all sizes, it has become a fixture in today's networks.
What is it about tape that has made it such a widespread storage medium? It isn't tape's performance, availability, accessibility, scalability, or reliability. Despite continued advances in data transfer speeds, disk buffers, robotics, storage capacity, and other features, when it comes to storing and accessing data, tape still does not compare to the speed and efficiencies of hard disks.
While IT managers feel comfortably complacent with tape, the two primary reasons it has been utilized are cost and dependability. For years, it has been the least expensive storage medium, offering dramatic cost savings compared to hard disk or optical disk storage. This difference in cost between tape and disk-based systems has until recently been a barrier to the adoption of hard disk storage to replace tape.
Tape and Hard Disk Storage Costs
Prices per megabyte for tape and hard disk storage solutions vary widely, depending on the application. Enterprise hard disk storage solutions-including hardware, software, installation, and training-traditionally can total as much or more than $.40 per megabyte. Even less elaborate solutions can be quite expensive, with a total cost of approximately $.25 per megabyte. Conversely, a standard DLT-based tape storage solution-including a standalone drive, one cartridge, and Legato backup software-can range in price from about $.04 per megabyte to about $.20 per megabyte. A Compaq StorageWorks TL891DLX DLT Tape Library with a capacity of 400 gigabytes configured with a host bus adapter, seven DLT cartridges, and Computer Associates' ARCserve software sells for a bundled cost around $21,000, or $.0525 per megabyte.
Eliminating the Hard Disk/Tape Cost Differential
However, these numbers do not take into account the development of a new generation of disk-to-disk (D2D) backup storage solutions that are exceptionally powerful but far less expensive than the hard disk storage of the past, a significant new trend in storage. These new D2D solutions have dramatically altered the backup pricing model and now provide disk-based backup alternatives of similar capacities that are two to ten times less expensive than tape.
The cost advantage of these next-generation disk solutions over tape is based on initial purchase costs only. If you create a three-year total cost of ownership (TCO) model, the disk-based systems year after year continue to provide savings. In addition to the superior TCO, D2D backup solutions provide the highest degree of reliability and availability. Technology advantages include industry-defined RAID sets, full hot-pluggable redundancy, predictive and pre-emptive online monitoring, higher transfer rates, on-site/online random file accessibility, off-site/online random file accessibility for disaster recovery, and the simplicity of disk storage compared to conventional tape backup operations.
Very affordable D2D storage solutions are available that utilize RAID-5 implemented Advanced Technology Attachment (ATA) hardware to provide high-speed, high-availability, enterprise-class, full-featured backup, on-site, and off-site disaster recovery systems. These D2D solutions include features that traditionally have only been found on enterprise-class, very expensive RAID arrays. Among these features are active-active failover, mid-plain internal architecture, and no internal cabling. All components are provided with full hot-swap capability, fault tolerance, embedded phone home monitoring, and data replication that can be easily integrated into a Direct Attached Storage (DAS), Storage Area Network (SAN), or Network Attached Storage (NAS) configuration.
With new D2D technologies, and highly efficient manufacturing and distribution methods, the total cost of D2D storage can average 1.7 cents per megabyte, all-inclusive. The industry has been asking for D2D solutions like these for the last 10 years. By eliminating the cost advantage of tape and offering one of the most robust cross-platform backup utilities, D2D maintains superiority in performance, availability, accessibility, scalability, and ease of use. D2D has opened up the affordable options for primary online/on-site storage and secondary online/off-site storage for backup and disaster recovery applications.
The Reasons for Affordable D2D
The development and distribution of powerful, affordable D2D storage is a result of several factors. Suppliers that own their controller technology can produce systems for far less than a company that must rely on technology partnerships. Second, recently announced D2D product uses advanced ATA chip sets that were specially designed for optimum performance. These chip sets are individually surface-mounted on each drive carrier, providing individual drive channels for complete drive independence and the fastest transfer rates in the industry. These systems also combine the use of internal firmware and a software application specifically designed for hard disk-based backup. The D2D software allows 24x7 event-based, "windowless" backup with features such as direct file overwrite to maximize media use efficiency and true multi-version control for unlimited point-in-time file and system control.
Dramatic advances in data storage media for the consumer market have resulted in lower prices for the business market. ATA/IDE hard disk-based home audio and video recording systems are now shipping to end users from companies like Panasonic, Mitsubishi, Hewlett Packard, Microsoft, The Dish Network, and DirecTV, and have resulted in continued improvement and lower cost for ATA/IDE hard drives and the anticipation of a significantly greater market availability for disk sales.
It is important to note that hard disk manufacturers will eventually ship more disk drives for the personal, home, automobile audio, and video market than they do for the computer market. This is a market where tape has disappeared forever. This virtually unending growth for hard disk sales will enable companies to continue to provide hard disk-based backup and disaster recovery systems for substantially lower costs than tape.
While lowering the cost of hardware and software was an essential part in developing an affordable D2D solution, a cost-correct and efficient method of providing the solution to the user is equally if not more important to make D2D a reality. Over the last few years the cost of hard disk storage has been wrongly inflated by the major storage vendors, for the most part to cover inflated overheads and inefficient ways of doing business today. The inability or refusal to readjust to new and available business models have made it necessary for the giants of the storage industry to charge between 20 and 40 cents/megabyte for hard disk products and services. While in some cases these prices may be justified, it is clear today that these types of costs are out of line with today's storage needs.
With the continued improvement in reliability, compatibility, and ease of use, the "black magic" component of storage has been eliminated as evidenced by D2D systems today. These systems include auto-recognition in their firmware, eliminate the need for server-based communication or configuration software, and have evolved to true plug-and-play storage appliances. Delivering solutions through premier channel partners provides one of the industry's best models for both experience and geographic support. This correct-cost delivery and support model significantly reduces total cost of ownership while providing the answers for a better solution today.
In addition to the standard corporate data storage needs, affordable D2D solutions are ideal for a variety of vertical markets, such as digital video security, medical imaging, print graphics, animation and special effects, and telecommunications.
The Performance of D2D and Tape
Hard disk offers many advantages over tape as a storage medium. Now, as the cost differential between hard disk and tape disappears, the value and use advantage of hard disk storage becomes even more apparent.
The first and most obvious of the differences between the media is the performance of the two types of storage. When evaluating a medium for backup and recovery functions, the speed with which the medium accesses files and reads and writes data is probably the most significant factor in determining the medium's value. In this area, the contrast between tape and D2D storage is dramatic.
D2D: Faster Backups and Online Immediate Recovery Eliminates Downtime
Backing up data involves downtime for a network, and downtime has its costs. Naturally, backing up data as quickly as possible minimizes disruption of the network and interferes as little as possible with users' productivity.
Tape and D2D differ significantly in the size of the backup window they require. This is primarily a function of the speed with which they can access files and read and write data. D2D systems can access files much more quickly than tape because of D2D's ability to randomly store and retrieve data. It can do this because its read/write heads fly directly to a new location as soon as it receives an address from the CPU.
File access on tape media occurs much more slowly than on hard disk. A system must find the location on the tape at which the desired file is stored. Accessing an address on linear media that must be wound or unwound to find the location for read/write takes substantially more time than accessing an address on a circular disk. Although there have been significant improvements in tape-based application software, someone who uses tape must confront the process and speed restrictions of tape mounts and the slow transfer and restore rates of linear-based technology.
It is important to note that tape throughput is not simply a matter of accessing a file. It also depends on a number of factors, including differences between the tape system's transfer rate and the host system's transfer rate. Adaptive cache buffering can help compensate for these differences.
Not only do D2D systems locate an address more quickly than tape, they read and write data more quickly as well. These factors help account for the differences in backup windows and downtime on these media.
The slower file access and read/write times of tape make it an unattractive medium for the many organizations in which demand for online storage and fast access is increasing. Tape is simply too slow to accommodate the users' need for data.
When an application depends on stored data, poor application performance and the resulting delays-especially if they involve users-can be catastrophic. Tape's slow file access can be a considerable and costly liability. D2D's immediate access times and disk-based read/write capability offers substantial benefits such as enhanced productivity, profits, and ultimately both continued user and customer satisfaction.
D2D accesses files and reads/writes data at light speed compared to tape, and the more advanced D2D systems offer online random access backup to provide much faster backup than tape solutions and real-time agents for event-based incremental backups. These latest D2D systems incorporate live monitoring agents that provide direct access to changed files for event-based incremental backups, eliminating the need to scan a disk or file registry.
D2D should also provide hot database snapshots for real-time, online protection. This online access provides unrestricted restore options, including doubleclick restores for ease of use. Backup files appear to the operating systems as the original files. In the event of a server crash or a virus attack, an administrator can immediately redirect the data path to the last known best version. This provides the ability to be up and running immediately allowing for an offline server restore or rebuild.
Other Features of D2D
While a storage medium's performance is a very important factor, other factors must also be evaluated to determine a medium's suitability for backup/restore functions. These other factors include availability, scalability, system management and ease of use. Like any other factors in storage, certain costs and risks are associated with choices IT managers make.
TCO-The Real Cost of Storage
To fully appreciate the impact the new and affordable D2D can have on an organization's storage model, it is necessary to examine the obvious and not-so-obvious costs involved in implementing various storage solutions. To understand the true cost of any type of storage-tape, hard disk, or optical, it's necessary to consider the Total Cost of Ownership (TCO).
TCO certainly starts with hardware and software, such as servers, HBAs, RAID, tape drives, optical disks, and management applications, but also includes, among other things, the cost of managing storage (including personnel costs) and the cost of managing risk-weighing the cost of fault tolerance and availability against the cost of downtime.
The cost of managing storage is an area in which the contrast between tape and D2D is particularly clear. Tape solutions can be notoriously labor-intensive. Maintaining tape can involve such tasks as identifying and retrieving tape volumes, monitoring tape barcode sequences, loading and unloading tape, inspecting tape for wear, rotating tape, manual disaster recovery operations, and inventory management.
Of course, there are automated tape solutions-including loaders, stackers, and libraries-that can relieve some of this burden, but managing tape solutions is still plenty of work, especially compared to D2D. The large storage capacities, durability of media, fault tolerance, and highly automated management solutions associated with hard disk storage keep manual management of D2D data storage to a minimum.
The Cost of Managing Risk
The cost of managing risk is another significant component of TCO that is affected by the type of storage solution selected. Any storage solution involves some degree of risk-risk that data will become corrupted, lost, or be otherwise unavailable. An organization's choice of tape, hard disk, or optical storage dictates the type of risk to which the organization's data is exposed.
Tape has its obvious risks, including degrading and wearing out and the possibility that the medium could tear, or that a tape drive can fail to the point of damaging the tape itself. Hard disks and optical storage can crash, or suffer other technical difficulties. Because any system can be vulnerable to failure, it is essential to maintain fault tolerance for all systems-tape, hard disk, or optical. The majority of D2D systems today operate in highly reliable RAID 5 configuration, with fully redundant components that are hot swappable, with complete online preemptive monitoring for all online/on-site and online/off-site backup and disaster recovery functions.
A crucial step in avoiding equipment failure is being alerted to it before it occurs. The best way to do this is to maintain an event-oriented system that can alert managers to potential component or system failures 24/7. Sophisticated remote management tools in today's D2D systems alert managers to any problems. Web-enabled GUIs provide system configuration, event and component monitoring from any standard web browser.
The length to which an organization goes to protect its data usually depends on how critical the data is to the organization, and on the organization's financial resources. Managers must weigh the cost of fault tolerance and high availability versus the cost of downtime. The cost of downtime in some areas of financial services can be substantial-millions of dollars per hour. Though the dollar cost may not be as high, downtime can be devastating for smaller organizations, as well.
Beyond the obvious step of employing both full and incremental data backup, what can an organization do to manage risk? There are a number of topologies IT managers can consider. D2D can play a crucial role in all of them.
A Storage Area Network (SAN) can provide a very high level of availability and reliability. SANs can incorporate hard disk, optical storage, and tape media (including automated tape libraries). A SAN, of course, can be an expensive and complex system to implement and maintain. The best D2D systems support a variety of storage topologies, including SANs.
Some organizations protect their data by employing a hot disaster recovery site. In this situation, they are likely to utilize wide-area connectivity and remote backup. RAID can be an excellent choice, providing data availability and recoverability to protect non-critical applications
D2D not only supports a primary online/on-site system but also makes it possible to co-locate a secondary D2D online disaster recovery system off-site. This solution provides immediate access to data in the event of the catastrophic loss of a data center.
Combining RAID and SANs
It is also possible to use RAID and SANs together to protect data. This combination can enhance the protection provided by RAID, offering redundancy and eliminating the risk of a single point of failure. One of the traditional advantages of SCSI-based or Fibre-based disk products has been hot-swap capability and redundancy. Today's advanced D2D should offer both capabilities.
In addition to protecting data through fault-tolerant architectures, managers need to be concerned with a storage solution's capacity and scalability.
D2D systems can offer ample capacity, some starting at 160GB native/320GB compressed. Data is secured on a disk-based RAID system in a compressed format, delivering fault-tolerant unrestricted configuration options at a low cost per megabyte. Scalable systems are available that span from 160GB to multi-petabytes of online capacity.
In virtually every organization, data storage requirements are increasing dramatically every year. In many cases, storage requirements increase by 100% or more annually. D2D offers significant advantages in terms of scalability, compared to tape. With D2D, a manager can increase storage capacity with no additional application-based tape library slots or drive license charges.
Storage Virtualization
Storage virtualization can make storage management much more efficient, speeding access to files and enabling managers to optimize their use of hardware.
Virtualization in D2D environments accommodates unlimited amounts of storage efficiently. D2D provides centrally managed storage pooling and virtual volume allocations. Virtual data volumes can be dynamically created, expanded, deleted, or moved from place to place, and storage pools and volumes can be built from different physical storage devices.
D2D systems that have proprietary virtualization applications should allow configuration without interruptions, allow hot spare drives to be added and deleted either as a pool or an array-dedicated spare and data management features such as snapshots, bit-level replication and virtual tape tools.
Ease of Use
In addition to the highly automated nature of D2D, which simplifies management, D2D is easy to use in other ways, too. For example, configuration and setup can be accomplished in less than 20 minutes.
To load any type of RAID onto a system, it is necessary to have software on a server or workstation that is able to communicate with and configure the RAID system. Self-recognition features built into firmware enable a server or workstation to communicate with a RAID system and configure it.
Since an organization's storage resources do not always grow in a systematic way, it is essential to have versatile hardware. The most advanced D2D solutions are system independent, supporting SAN, DAS, or NAS configurations, and Full Fabric, Fibre, SCSI, or ATA. They provide support for TCP/IP, Ethernet, Data and SCSI over IP, and for Serial, LAN, WAN, or web connections.
These systems are also designed to be completely host- and operating system-independent (providing that the host can support external SCSI or Fibre Channel disk drives). They support Windows NT/95/98/2000, Novell Netware, SCO, Linux, MacOS, Solaris, AIX, and Unix and clustering such as Microsoft's Window's 2000 Datacenter (Cluster) Server. Internet-enabled management processors are compatible with all major web browsers.
Now that IT managers can take advantage of affordable D2D as an alternative to tape, they can enjoy hard disk performance, availability, accessibility, scalability, and ease of use for more applications than ever before. It is not likely tape will disappear as a storage medium, but in light of the price equality with D2D, the number of storage niches tape occupies should, near term and will long term, decrease significantly.
Diamond Lauffin is the executive vice president at Nexsan Technologies (Woodland Hills, CA). |
