Advertisement
To print: Select File and then Print from your browser's menu
-------------------------------------------------------------- This story was printed from ZDNet Australia. --------------------------------------------------------------
Archiving: what's in store?

By Kire Terzievski, Technology & Business magazine
August 14, 2003
URL: http://www.zdnet.com.au/news/business/soa/Archiving-what-s-in-store-/0,139023166,120276948,00.htm




What's the difference between backup and archiving? And what are the pros and cons of the three most common media--tape, disk, and optical? We set a budget of AU$20,000 and asked three vendors to come up with a storage solution.

The amount of data that organisations are storing is increasing quite rapidly.

There are a few factors that are contributing to this growth. There are many industries that are legally bound to preserve large amounts of data for years, such as health care, financial services, and government. There's also widespread use of enterprise applications such as customer relationship management (CRM), enterprise resource planning (ERP), and supply chain management (SCM) that also increase storage needs. And of course, there's the explosion of e-mail, which companies also need to hang on to for a long time.

A recent survey carried out by InfoStor found that companies intend to purchase more tape and more disk storage. The survey also found that speed, capacity, and reliability were the main factors concerning those who make purchasing decisions.

There are many different strategies for archiving data, which brings on the debate about which archiving medium fits your needs the best.

Tape, Disk, or Optical
Sometimes picking the right solution to best meet your business needs may mean having a combination of these technologies, which would ensure your covering all your bases.

See the table below for a brief summary of the advantages and disadvantages of the various storage media.

Medium Advantages Disadvantages
Disk
  • Archiving to disk is faster than tape
  • Restores from tape are fast and convenient
  • Disk prices have fallen
  • Users aren’t sure whether they have archived their files sucessfully
  • Mirrored disks offer no protection against viruses, theft, or natural disasters
  • Tape
  • Offers higher storage capacities
  • Automated tape libraries still offer lowest cost per GB
  • Media can easily be stored off site
  • It has a long shelf life (30 years)
  • Very slow at restoring data
  • Optical
  • Generally quick at restoring randomly chosen files
  • Media can easily be stored off site
  • Duplicate copies can be made quickly and easily
  • In artificial aging tests, DVD-RAM has been determined to have an estimated shelf life of up to 300 years
  • Quite slow at archiving data
  • Backup or archive?
    Some people are not clear on the difference between backing up data and archiving. So before considering any of the technologies you should decide whether you need an archiving or a backup solution. Having a better picture of what the differences are will help you develop the best storage solution.

    When it comes to backing up data all you are doing is creating an exact mirrored copy of your company's data. This data is rarely restored and is simply a reserve in case something happens to the original data. Backing up data isn't costly and has typically been done using tape devices. The only problem here with tape is when you do have to restore certain data it may be difficult and time consuming.

    If you're going to be accessing your data over and over again you should be archiving your data. You can do this by logically placing data where it can be easily retrieved. That's the main difference but many companies still are confused about backing up and archiving. Many just think as long as they have a copy of the original data somewhere it will be fine.

    It can prove very costly to store data on network drives. You don't want to be using up valuable storage space on your network with infrequently accessed data and if by chance the network isn't accessible, you could find yourself in all sorts of trouble.

    On the other hand, if you were to use a recent tape backup to restore data you may find most of your staff will not have the patience or the technical know-how to find specific data from the tape.

    A simple concept called hierarchical storage management (HSM) was developed to provide the ability to define and automate intelligent, cost-effective decisions regarding where and when data is stored. The notion was based on the fact that not all data is equally useful. Some may be required immediately, but with less frequently accessed data, slower transfer rates are much more tolerable as well as less expensive.

    One of the main advantages of archiving is that it gives companies greater control over their data. For a large organisation, archive management software could enable them to operate a departmental storage system. Each department can then be given a specified amount of storage to archive data.

    Where an organisation has many sites, archiving software can be embedded within the company's intranet to make storage and retrieval simple and consistent across the organisation.

    Then there are optical discs. We found that using optical libraries and archiving software can simplify the process of storing data and retrieving it later. Using optical discs are the best choice for companies that need to store data which cannot be erased. DVD-R discs can hold 4.7GB of data and the library system that automate the loading of the discs mean that the storage capacity of the archive system can grow alongside the company's requirements.

    Also by using optical discs to archive data, you will be freeing up valuable disk space, which you would want to keep for live working documents. By storing information offline, you can ensure your data is protected even if something major happens to the server or network.

    What to look for when buying an archiving unit

    • Speed: The amount of data you have to archive and the amount of time you have available to finish archiving determines the level of performance you need.
    • Capacity: The maximum amount of data that can be written to a single tape cartridge/disc/drive. What is the total native and compressed (for tape) capacity?
    • Price: Look at the cost per gigabyte in terms of cost of cartridge/native capacity. Cost per gigabyte can vary quite dramatically. Ideally you would want to go for a tape that has a high capacity and is respectably fast and inexpensive.

    You should also consider the following questions:

    • Will you be randomly accessing files?
      If the speed of access to a particular file is of paramount importance then an optical jukebox would be the ideal choice because of their random access capabilities.

      If you are sequentially archiving data and if you are accessing given files infrequently, then tape technology may be your best bet. This is the prime reason why tape libraries are used by the majority to back up and archive data.

      Disk on the other hand is faster than both. You might recall we tested the Quantum DX30 a few months ago. This device can be used in conjunction with a large tape archive unit to give you the best of both worlds being able to quickly restore recent data as well as offload data that is more than a week old (for example) to tape or optical.

    • How much data is being stored?
      You can easily work out the total capacity of your library by multiplying the media capacity by the number of cartridges or disks the library can accommodate. If you need to archive large amounts of data you would most certainly choose to use tape. LTO 1 tapes can archive up to 200GB of compressed data. Hard disks on the other hand can archive data limited only by your budget, however the media cost per GB is significantly higher. Optical discs can't archive as much data, but have a lower cost per GB than tape or disk.

    • How much time do you have to transfer your data?
      Archiving a significant amount of data can take longer than you might think. Most tape drives have transfer rates of 15MB per second. This would mean you could archive approximately 216GB in four hours. Optical drives are much slower. They can transfer data at about 2MB per second, which equals 28.8GB in four hours. Hard disks usually transfer data much faster than tape. This is certainly the case with a Quantum DX-30 using fibre channel--45MBps+--but not so much with the Iomega NAS server using gigabit Ethernet we looked at in this review, which clocked in at 8MBps.

      Many tape and optical libraries have the option of utilising multiple drives, which can increase the throughput by recording several tapes or optical disks simultaneously. You should also make sure the interface you will be using is fast enough to take advantage of the library's full potential. Most likely you will be using some variety of SCSI, but fibre channel and gigabit Ethernet are also popular depending on the circumstances.

    • What are your future data storage requirements?
      To work out your future storage requirements it would be a good idea if you compared your data storage requirements over the last few years and then estimated your future needs from this data. If it seems like you have to archive twice the amount of data you archived last year, then you could find yourself running out of storage in the near future. The best way to address this problem is to look for libraries that can be expanded.

    • What's your budget?
      You need to compromise if the library you are considering to buy is the right one but your budget does not allow for it. Another way of looking at things is by looking at the cost of entry versus the cost of usage. Over time you may well see that you have spent more money on media than on the initial hardware.

      You should also check the warranty and service contracts that the vendors will give you. Take note that many of the drives are not made by the same manufacturer of the library and that the warranties may be different. There are many service levels that are available as well. Try to find a service level that suits your budget and a turnaround time for repairs that is acceptable.

    Roadmaps
    You will also need to look at scalability. Is it a proven technology that you are buying? Some tape formats for example have been around for years and continue to give us roadmaps but there are others that don't. In these cases you may find yourself with having to purchase more hardware in order to increase your storage capacity.

    The main driver in tape technology over the next few years will undoubtedly be the increase in the amount of data stored and archived. This will ultimately mean that archiving has to become more automated, more intelligent, larger, faster, and of course even more reliable than it is today.

    The Lynx autoloader we looked at in this review uses LTO 1 tapes. They have a native capacity of 100GB (200GB compressed) and can transfer data at 15MBps in native mode or 30MBps in compressed mode.

    There are many tape technologies that are out there. If you are after some more detail on the various types of media that's available we highly recommend you have a look at the April issue of T&B where we cover all the different tape formats that are on the market.

    DVD-RAM discs--as used in the JVC jukebox we looked at--were the first rewritable flavour of DVD to come to market, and has subsequently found competition in the rival DVD-RW and DVD+RW formats.

    With the same overall size as a standard 120mm diameter, 1.2mm thick CD, DVD discs provide up to 17GB of storage with transfer rates higher than CD-ROM and access times similar to CD-ROM. They come in four versions:

    • DVD-5 is a single-sided single-layered disc with a capacity of 4.7GB;
    • DVD-9 is a single-sided double-layered disc offering 8.5GB;
    • DVD-10 is a 9.4GB dual-sided single-layered disc; and
    • DVD-18 will increase capacity to a huge 17GB on a dual-sided dual-layered disc.

    The sustained data transfer rates for DVD-RAM is quite low in comparison to disk and tape. JVC claims it can achieve a burst rate of 10MBps and a sustained transfer rate of 2.77MBps (read speeds).

    The Iomega NAS uses IDE hard drives to store data, in this case 185.2GB IBM hard disks which spin at 7200RPM. SCSI hard disks on the other hand are faster than IDE drives but are also a lot more expensive. Current SCSI solutions use 36, 73, and even 146GB hard disk drives that operate at 15,000RPM.

    Tape: Lynx LTO 1/8 Autoloader A5040A (LTO 1)

    The Lynx Autoloader is really an HP product. In fact on the front panel it says it's a HP StorageWorks 1/8 Autoloader.

    The Lynx Autoloader is designed for medium-sized workgroups that have basic storage requirements. The HP Autoloader is designed to store 1.6TB of data and has eight slots that can each hold 200GB LTO 1 cartridges. The unit comes in a 2U rack mountable enclosure.

    The Lynx doesn't feature a Web-based management interface nor does it support multiple drives or barcoding. With this model, Lynx decided to keep it simple. Indeed this device was easy to set up, operate, and manage.

    Lynx LTO 1/8 Autoloader A5040A (LTO 1)All you have to do to get it up and running is attach a SCSI cable from the Autoloader to the server. Your archiving software should then automatically discover the Autoloader and drive and install the necessary drivers. Loading cartridges is a piece of cake--the unit accepts cartridges one at a time.

    The menus on the front panel were grouped into four categories: commands, information, configuration, and diagnostic. From the panel you can easily work out how to perform simple actions like importing and exporting cartridges as well as changing SCSI IDs. Lynx took some special care to design the power button, which prevent the Autoloader from being turned off accidentally. The only way you can shut it down is by inserting a pencil into the enclosure where the power switch is.

    The Lynx unit we tested uses Linear Tape-Open (LTO) tapes. We were advised the LTO 460 loader is due for release this month. Performance is expected to be the same as the Lynx LTO 460 drive that we tested a few months ago. Basically the LTO 460 loader will feature the same robotics as the product we are featuring in this review, but with the faster drive.

    The Lynx Autoloader was the fastest at archiving data, managing a sustained transfer rate of 19.17MBps. It was a different story when it came to restoring data where it only managed 3MBps. As expected, tape proved to be the least expensive technology in cost per GB. At only $1.18 per GB, you can see why it's still the most popular form of media currently being used.

    Product: Lynx LTO 1/8 Autoloader A5040A (LTO 1)
    Price: AU$16,125
    Vendor: Lynx Technologies
    Phone: 02 9663 4855
    Web: www.lynxtech.com.au

    Interoperability:
    Wide software support across all platforms.

    Futureproofing: ½
    Very fast storing, slow at restoring. Capacities are increasing rapidly.

    ROI: ½
    Least expensive medium.

    Service:
    1-year return-to-HP warranty, next day exchange.

    Rating:

    Disk: Iomega NAS 405M

    The Iomega Network Attached Storage (NAS) server is also a good way of increasing your network capacity. It was engineered basically to share information over networks.

    The NAS server features a reasonably fast CPU (Intel Pentium 4 2.0Ghz), 1GB of memory, and four fast hot-swappable IDE hard disk drives, each with its own dedicated IDE channel to give you a storage capacity of 720GB.

    The Iomega has a moderately deep 1RU form factor. The unit includes a simple but very robust rack mounting bracket as well as a pair of Cat 5 patch cables. The front panel styling looked very modern and includes status LEDs for the four-drive bays and systems power. At the rear of the unit you will find two 10/100/1000 Ethernet ports, serial port, two PS/2 ports, two USBN ports, a VGA port, and a SCSI port.

    Iomega NAS 405MThe NAS server also came with Iomega Automatic Backup Software. It's a very basic software application, which as the name suggests backs up your data automatically once you set up all the parameters.

    We had a gigabit Ethernet connection to the NAS server. Its performance was also affected by the network connection, which proved to be the bottleneck. We were able to get some modest results, however we should remind you that this is not indicative of the type of transfer rates you should be getting when transferring from disk to disk. As we had mentioned in the introduction we were able to stream over 45MBps from disk to disk using fibre channel. This disk-based solution didn't have that sort of performance. But with that sort of performance comes a high price tag.

    The NAS server managed a sustained write throughput of 8MBps, which is quite slow in comparison to tape and despite disk prices coming down in price all the time, it still remained more expensive than tape in terms of cost per GB. One area of surprise was reading data back off the disk--it was actually faster than tape at 4.27MBps. And of course you don't have to find the right tape, wind on to the right spot, and then start reading.

    Product: Iomega NAS 405M
    Price: US$6199 (local POA)
    Vendor: Iomega
    Phone: 02 8875 7851
    Web: www.iomega.com

    Interoperability:
    Wide software support across all platforms.

    Futureproofing:
    Limited storage capacity compared to tape and optical.

    ROI:
    Less expensive that optical, more than tape.

    Service:
    3-year limited warranty.

    Rating:

    Optical: JVC MC8100U

    The MC-8100U is a digital storage system from JVC Professional. It's primarily tailored for users who need vast amounts of nearline networked archiving storage with random access.

    The MC-8100U features industry-standard removable media and you have the choice of any combination of drives including DVD-RAM and CD-RW. It can feature up to four drives while the 8200U and the 8600U can hold up to six drives. The MC-8100U can house 100 discs, which means it can store up to 940GB of data on double-sided DVD-RAM media. The 8200U can store 1.88TB while the 8600U a whopping 5.64TB of data.

    JVC MC8100UAll the 8000 series jukeboxes use a high-speed disc changer and flip mechanism. The flip mechanism comes in handy if you're using double-sided media. It will actually flip a disc on the fly, which eliminates the additional time it would normally take to transport media to a specified location. You can also buy an optional disc printer. The printer is fully automated which means it will print labels in full colour directly onto the discs after recording.

    The JVC jukebox didn't ship with any software; you have to purchase that from a third party supplier. We downloaded a trial version of PointSoft Jukebox Manager off the Internet. This worked quite well with the JVC jukebox. It detected the jukebox first up and managed it without any problems.

    The JVC performed as we had expected. We knew it wasn't going to be too quick at moving data onto disc. It managed to write to at approximately 400KBps and read data from disc at 2.44MBps, which is pretty much what JVC said it should do. The JVC was also the most expensive per GB.

    Product: JVC MC8100U
    Price: AU$19,999
    Vendor: JVC
    Phone: 02 9370 8888
    Web: www.jvc-australia.com

    Interoperability:
    Loads of third-party software available for most platforms.

    Futureproofing: ½
    Provision for multiple drives and label printer.

    ROI: ½
    Most expensive archiving solution, but cost of optical jukeboxes has dropped considerably in the last year or so.

    Service:
    1-year parts and labour warranty, option to extend.

    Rating:

    In summary

    There is no Editor's Choice in this comparison--each of the products tested were excellent for what they are designed to do. At the end of the day, tape still remains the least expensive in terms of cost per GB, however disk is breathing down its neck.

    We recommend you use tape-based products for backing up data, rather than archiving. Tapes aren't something you should be constantly accessing. An optical solution would be the more logical choice here especially if you weren't too concerned about data transfer rates. Disks on the other hand should be employed when speed is of paramount importance.

    Subscribe now to Australian Technology & Business magazine.

    About RMIT Test Labs
    RMIT Test Labs RMIT IT Test Labs is an independent testing institution based in Melbourne, Victoria, performing IT product testing for clients such as IBM, Coles-Myer, and a wide variety of government bodies. In the Labs' testing for T&B, they are in direct contact with the clients supplying products and the magazine is responsible for the full cost of the testing. The findings are the Labs' own--only the specifications of the products to be tested are provided by the magazine. For more information on RMIT, please contact the Lab Manager, Steven Turvey.

    Copyright © 2009 CBS Interactive, a CBS Company. All Rights Reserved.
    ZDNET is a registered service mark of CBS Interactive. ZDNET Logo is a service mark of CBS Interactive.