An Overview of the Different Technologies Available for Backup
CONTENTS CONTENTS... 2 INTRODUCTION... 4 TAPE DRIVES... 4 Cost... 4 Quantity of Data... 4 Length of Archival... 5 Time to Backup... 5 Time to Restore... 5 HARD DRIVES... 5 Cost... 5 Quantity of Data... 6 Length of Archival... 6 Time to Backup... 6 Time to Restore... 6 ONLINE BACKUP... 6 Cost... 6 Quantity of Data... 7 Length of Archival... 7 Time to Backup... 7 Time to Restore... 7 OPTICAL DRIVES... 7 Cost... 7 Quantity of Data... 8 Length of Archival... 8 Time to Backup... 8 Time to Restore... 8 COST... 8
SECURITY... 8 REDUNDANCY... 9 SOFTWARE... 9 CONCLUSION... 10 WORKSHEETS... 11 INDUSTRY EXAMPLES... 12
INTRODUCTION Today s organizations rely too much on their data to not have backups. But backup systems are not the kind of glamorous, high impact projects that get management s attention; too many organizations do not think about backups until they actually need to restore data. Putting together a backup solution for your organization does not have to be painful. In first whitepaper of this series, introduction to Backup, the overall concept of backup was discussed. The second whitepaper, Factors to Consider, discussed the important factors to think about before deciding which backup technology was right for you. In this whitepaper, we will take a look at the common, mainstream backup technologies on the market, and see how they stack up. TAPE DRIVES Say the word backup to most IT professionals, and most likely they will think of tape drives. Over the years, tape has endured as the backup medium of choice. It has a great combination of cost effectiveness, speed, density, durability, and reliability. While tape seems to be the default choice for organizations looking to do backups, it has some serious competition now, particularly from hard drives. Tape Drives: A tape drive, which is also known as a streamer, is a data storage device that reads and writes data stored on a magnetic tape. It is typically used for archival storage of data stored on hard drives. Tape media generally has a favorable unit cost and long archival stability. Cost The tape media themselves have a fairly low cost per GB. At some price points and for some tape technologies, hard drives might be a bit cheaper. One thing to keep in mind here, is that more often than not, your servers already have all of the equipment they need to backup to a hard drive. But with a tape system, you will need to purchase a tape drive. And if your data exceeds what one tape can hold, you are probably going to need to purchase an autoloader, which tend to be quite expensive. You may also need to purchase a SCSI adapter, so be sure to double check the interface requirements of the drive you are looking at and the capabilities of the server you will be using. Quantity of Data Tapes can hold a lot of data. On top of that, most drives provide built-in compression, which drive manufacturers usually advertise as being able to double your storage (you will see these drives and tapes listed as having a capacity of, say, 300 GB/600 GB). Like any other compression system, your mileage may vary. Depending on your data, you may see better than 2:1 compression, or maybe your
data will not compress at all. The capacity of tapes is very closely linked to their price. Some tape drives can handle a wide variety of tape capacities, and some tape technologies offer many different sizes; others do not. If your storage needs are expected to change significantly over time, having a tape drive that allows you flexibility in terms of the storage capacity can save you big bucks down the road. Length of Archival Properly stored tapes archive well. Improperly stored tapes are a game of dice at best. Before selecting a tape system, go to the manufacturer s Web site, and take a look at the storage requirements of the media. If you do not think that you can provide conditions which meet those needs, than you should pass on tape. If you will be able to properly store your tapes, they will be quite reliable over the long term. Keep in mind, when a tape fails, it often takes the drive with it, which can require costly and lengthy repairs. The only thing worse than telling the boss that the tape broke during the file restore, is telling him that the tape is jammed in the drive to boot! Time to Backup Tape drives write sequentially to the tape, which means that they can write extremely fast. Time to Restore The amount of time that it takes to restore from tape depends on if you are going to be restoring a full file system or individual files. Because of the sequential nature of tape, having to grab one or two files can take a long time, especially if they are on the other end of the tape! HARD DRIVE Hard Drives: A hard disk drive (HDD), commonly referred to as a hard drive, hard disk, or fixed disk drive, is a non-volatile storage device which stores digitally encoded data on rapidly rotating platters with magnetic surfaces Over the last few years, more and more organizations have been turning to hard drives for backup, and with good reason: hard drives are fast, reliable, well-understood by staff, and are plummeting in price per GB while density skyrockets. For most organizations, the backup decision boils down to tape or hard drive? Many organizations follow a blended approach, using hard drives for near-line, frequent backups, and tape for offline (or offsite), less frequent backups. Cost Hard drives keep getting cheaper. Additionally, hard drives usually do not require any special hardware to use. Tape still holds a price advantage at many price points, but make sure to check prices when you make your decision. At the rate that pricing changes, there is no way to say definitively which costs less. Another nice thing about hard drives, is that if you need more space, you do not need to buy any new
equipment, just buy larger drives. Drives also have a side cost benefit as well: if you no longer need a drive for backups, you can put it in the part bin for other servers or desktop PCs. Quantity of Data Hard drives can store just as much as tapes, and lately, they can store even more. Unlike tape drives, hard drives do not have any kind of built-in compression. But at the rate that hard drives are dropping in prices and increasing in densities, chances are that you will be able to get bigger hard drives as your needs expand much more easily than you could get larger tapes. Length of Archival Hard drives can store data for quite a long time. Like tapes, hard drive should be carefully stored. Hard drives are sensitive in ways that tapes are now, though. Hard drives are extremely sensitive to sudden shocks and such, while tapes (not the drives themselves, though) are more forgiving on that end. However, hard drives are less sensitive to environmental factors such as heat, cold, and humidity. Of course, that does not mean that you can keep a pile of hard drives in a barn, but it does mean that you are more likely to be able to take a hard drive straight from storage and start restoring files. Time to Backup Hard drives are extremely fast for backing up. Time to Restore Hard drives can restore entire file systems very fast, and can also restore individual files almost as quickly. ONLINE BACKUP In the last few years, we have seen more and more services that have been traditionally performed on-premise become offered online. Backup is one of them. Online backup has been offered in a variety of different ways for at least 10 years now. Some of these systems have been aimed at businesses, some at consumers. Online Backup: A remote, online, or managed backup service is a service that provides users with an online system for backing up and storing computer files. Managed backup providers are companies that provide this type of service. All such solutions share the same weak link, which is connection speed. Nonetheless, depending upon your needs, online backups may be an option for you, even if they are functioning as a secondary system. Cost Online backup providers vary wildly in the price packages they offer, so there are no generalized statements that can be made about their cost.
Quantity of Data Like cost, the data capacity is variable amongst providers. A high quality business class backup vendor should be able to offer unlimited storage capacity if you need it. Length of Archival Assuming that your online backup provider stays in business, your backups should be theoretically available to you forever. Check the contracts though, some providers may limit your archive time, or charge extra for long-term storage of data. Time to Backup Even many of today s broadband providers that are advertising super fast speeds are not giving their customers nearly as much upload speed. Even customers on more traditional leased lines such as T1 and T3 lines have upload speeds that are mere fractions of the speeds offered by local SATA and SCSI systems in servers. If you are using online backup to protect a small amount of data (or data that infrequently changes), online backups may be for you. It may also be possible to send the vendor a tape or hard drive with an initial seed backup, and then synchronize only what has changed on a regular basis. Time to Restore Your time to restore files is dependent upon your connection speed as well. However, many Internet connections have higher download speeds than upload speeds, so you could be able to restore files more quickly than you can back them up; be sure to verify your connection speeds with your ISP. For individual files, even large ones, online backup should be fairly fast. OPTICAL (CD-R, DVD±R) DRIVES Optical Drives: In computing, an optical disc drive (ODD) is a disk drive that uses laser light or electromagnetic waves near the light spectrum as part of the process of reading and writing data. Some drives can only read from discs, but commonly drives are both readers and recorders. Recorders are sometimes called burners or writers. Remember when the 700 Mb offered by a CD-R seemed like a massive amount of space? Optical drives just don t have the capacity most businesses require today. On the other hand, they are quick, easy, the media is usually laying around the office anyways, and every nearly computer has the hardware for it. If you need a way to backup a small amount of essential data (if only for redundancy or on an ad-hoc basis), using CD-R s or DVD±R s may be just right for you. Cost Optical media is dirt cheap. And unlike hard drives and tape, you don t even need to shop online and wait for it to be delivered to get the best price. Just go down to your local discount store (you can even
get them at most food stores now) and you can pick up a 50 pack of discs for pennies each. Since nearly every computer in the shop will have the ability to read and write these discs, you won t have to spend a dime to use them either. Quantity of Data This is the biggest weak spot for optical drives. CD s are limited to 700 MB, and DVD s are capped at about 8.5 GB for the dual-layer versions. For a few organizations though, that may be just enough to meet their needs. Length of Archival When stored in a case, optical media should last for some time. However, questions regarding their long term performance have come up. For safety s sake, don t count on optical discs for storage beyond a decade or so. Time to Backup Optical discs, while slower than hard drives or tape, are still quite fast. Time to Restore Again, optical discs are fast, and their performance on individual files is almost as quick as their speed for sequential reads. COST There are a couple of additional notes regarding cost that are not specific to any particular technology. The reality is most organizations make cost their number one decision point. The problem with this is that what is cheap today may not be cheap tomorrow, and what is expensive today will probably be cheap tomorrow. For example, hard drives tend to drop in price more quickly than tape drives, and the price of tapes themselves stays fairly steady. At some data densities, you may find out that tapes are less expensive than hard drives now, and a year from now, they will be more expensive. If cost is your number one concern, try to keep sight of the downward curves in pricing, and take that into account. SECURITY The security issues around backups are going to be the same for all systems with a physical media. Your security issues will revolve around two major items: who can access the physical medium that the backups are on, and who can access the data once the physical medium is connected to a computer. The first question is a matter of how and where you store your backups. Are they in the system administrator s desk? Are they in a safety deposit box in a bank across town? Have you given them to a third party commercial archival company? The second question is up to the software you have used to perform the backup. Some software packages can password protect your data in a variety of ways. If this
is an important consideration for you (particularly if there is a chance that someone may be able to access the offsite backups without proper authorization), make sure to ask your software vendor about this functionality. If you choose the online backup route, you have no control over your data. Talk to your vendor and ask the tough questions to decide if their security measures meet your needs. Some questions that you may wish to ask: What kind of background check must your employees pass to work here? How many data centers do you have? Where are they located? If your main data storage is wiped out, how will you restore the data? Is our data encrypted in transit and on disk at your data center? Who can access our data, and how do they access it? REDUNDANCY How redundant your solution is depends largely on how much money you are willing to spend. For example, if you choose to backup to hard drives, and you do not want to buy extra drives so that you can rotate some offsite, then you will be left without redundant backups. If you are willing to spend some serious money, you can do backups to a near line system at night, and backup from the near line system to an offsite backup on a weekly or even a daily basis. Any quality online backup vendor will have plenty of redundancy of their own, but you will want to ask them for details to make sure that it meets your needs and expectations. SOFTWARE Of course, having a backup device without the software to perform the backups does not make much sense. All modern operating systems come with backup software of some sort which may be sufficient to meet your needs. Often, backup hardware (especially external hard drives and tape drives) come with a backup software suite of their own. And of course, the third party software market is filled with all sorts of various packages. In general, the software that comes built-in to the operating system is often sufficient to meet the needs of many organizations. These packages are not always the most feature rich, fastest, efficient, etc. But they have the advantage of being widely known and free. Typically, the software that comes free with external hard drives (usually of the one click variety) are designed for consumers, and are not going to meet the needs of a company. The software packages that come with many tape drives are often light versions of commercial grade backup packages and can do a decent job. Overall, you will want to look for these things in a software package:
Support hardware: Make sure that the software supports your hardware. If necessary, try to get a trial or evaluation copy and perform some test backup/restore cycles. Ease of use: Not every package is easy to use. If this does not sound like a big deal, imagine the damage that can be caused by the backups not being configured correctly because the software made no sense. Or worse, if you have a hard time using the software under ideal conditions, will you be able to make sense of it in the middle of a crisis? Compression: Some packages perform their own compression. Some can work with your existing hardware compression (for the record, you do not want to use both simultaneously). Make sure that the compression meets your expectations and needs. Security: Some software packages can protect your data if needed. You will want to know if it is truly encrypting the data at rest, or if it simply adds a flag to require a password to re-open the file. If you need ultra security, you need encryption. Application awareness: some packages are aware of typical server applications, and are capable of backing them up while they are running; some packages cannot. If possible, find a solution which is aware of your applications, otherwise you will need to back those applications up separately. Three common examples of applications which backup software needs to be aware of are Hyper-V, Exchange, and SQL Server. CONCLUSION While the variety of different backup solutions means that picking one involves some effort, it also means that you can custom tailor your purchase to fit your exact needs. For most organizations, the choice really comes down to hard drive or tapes? and within the tape category, there are a large variety of choices. Armed with the proper information, and having a good understanding of your organization s needs will make this process much easier and accurate. Here is a chart summarizing this information, and a worksheet for you to use when evaluated your choices, to help you make your decision.
WORKSHEETS Table 1: What s Important to You The whitepaper, Factors to Consider, included a table to be completed which would help identify the key factors that were important to you. As a reminder, the table is presented again below Factor Priority Special Considerations Cost Quantity of Data Security Length of Archival Redundancy Time to Backup Time to Restore Table 2: Which Solution is Right For You This whitepaper summarizes the performance of each type of solution according to those factors. Combine the results of Table 1 and Table 2 to determine which backup technology is right for you. Tape Drives Hard Drives Online Backup Optical Drives Cost Mid-to-expensive Mid-to-expensive Varies Inexpensive Quantity of Data Medium-to-large Large Unlimited Very limited Security Up to you Up to you Depends on vendor Up to you Length of Archival Long term Long term Long term Medium term Redundancy Up to you Up to you Depends on vendor Up to you Time to Backup Fast Very fast Slow Fast Time to Restore Fast (all files system), or slow (individual files) Very fast Slow Fast
INDUSTRY EXAMPLES Many industries have specific backup requirements. Here are a few examples which may help identify the situation that your own organization faces. Company A has clients in the legal industry. While Company A is not a law firm, their clients on occasion need to provide documents and files that contain privileged information. This information is not allowed to be backed up. If printed these documents must be securely destroyed when no longer needed. To accomplish this, Company A has created a special email address for any such files to be emailed to. This email address file is excluded from the backup scheme. This company has about 250 GB in network storage, 1 main file server, an Active Directory domain controller, and a number of application servers running as virtual machines. Overall, this company is performing full backups on a nightly basis to hard drives and retaining the most recent four (rotated nightly), and on a weekly basis, copying the most recent backup to an external hard drive, and rotating that drive off site. They rotate their offsite backups so that there is always three weeks worth of backups stored offsite. Once a month, they retire the oldest offsite backup and put it in a permanent archive, and rotate in a fresh set of drives. Company B handles extremely sensitive personal information on millions of people. This information includes medical history and insurance claims. The data itself (hundreds of GB) is not as important as the applications that process it (a few hundred MB), because the data can be re-provided by the customers, and is renewed on a weekly basis. This company backs up both the data and the in house applications nightly to hard drive, and on a weekly basis, makes an offsite backup to external hard drive that contains only the company s applications. The first backup of the cycle is a full backup, and subsequent backups are incremental backups. This prevents the clients sensitive data from leaving the premises. Company C is a typical organization with less than 1 TB of data to backup. Company C is performing nightly backups to tapes using an autoloader, connected directly to their main file server. On a weekly basis, they swap the tapes in the autoloader with the oldest set of tapes stored offsite. Once a month, they retire a set of tapes to be put into permanent storage. At the beginning of the cycle, they perform a full backup, with incremental backups on the subsequent nights. Company D works with data covered by HIPAA. As a result, they have to be extremely careful with where their data is stored and who has access to it. To simplify matters, none of the HIPAA data itself is very important to them; it is incidental to their work. Not critical to it. All data that might be covered by HIPAA is stored in a separate file location from all other data. That location is excluded from the backups. This organization is otherwise a fairly normal organization; they backup to hard drive nightly, and once a week, copy the data to tapes which are stored offsite. The backups at the beginning of the cycle are full backups, with incremental backups throughout the remainder of the week. They only use one set of offsite tapes, but periodically, the tapes are replaced with a fresh set, and the old tapes are overwritten with random data and sent to a third party for secure destruction.