Selecting a CD/DVD Archive Solution

Using recordable CD and DVD discs to archive data makes sense for a host of reasons. CD/DVD technology provides many advantages, including cost, reliability, media longevity, and availability. The technology has obtained sufficient economic scale through its consumer applications to enable near universal availability - nearly every computer can read CDs and DVDs. And unlike tape and proprietary optical formats, reading a disc 10, 20, or 30 years from now is unlikely to present any difficulties due to obsolescence. Maintaining backwards compatibility as well as cross-platform compatibility have been two of the most significant factors contributing to the success of CD and DVD. The next generation of drives - either HD-DVD or Blu-Ray - is likely to offer compatibility with the previous standard, as evidenced by the recent announcement of a major manufacturer that their next-generation drives will also read DVDs. In fact, backwards compatibility will be one of the driving forces behind the acceptance of the next new standard.

Now that you have determined that CD/DVD is the right technology for your application, the next task is to pick the right solution. Purchasing a shrink-wrapped standalone software package (or downloading a freeware program) is fine for casual use by a single user, but if you are creating tens, hundreds, or thousands of unique CD/DVD discs on a regular basis, you need something more robust, preferably with a single source for support. Also, when you have distributed the capability for making discs, you are also distributing responsibility for archival, and risk losing control of intellectual property assets. A centralized system, utilizing automation, is the choice many different types of organizations have found to be the most cost effective.

At the outset, you must determine your production requirements; that is, how many discs do you need to record daily (peak production), as well as weekly and/or monthly numbers (sustained production). The system you select must have the capacity to handle peak production requirements, as well as having the durability to handle sustained production levels. Closely related to the raw number of discs produced is the amount of data that each disc contains. For example, in many imaging applications for healthcare and finance, each disc typically contain less than 100 MB (compared with a capacity of 650 MB for CD, and 4.5 GB for DVD). For some applications, data sets and/or file sizes are significantly more than 4.5 GB (the capacity of a DVD). In the past, this meant that an alternative such as tape was needed, but one vendor, Young Minds, Inc., provides disc spanning technology that can record file sizes up to 64 GB on DVD. There are additional considerations, such as many discs are unique (one-offs) and how many are duplicates. Here, you may want to consider creating at least two identical discs, one for onsite and one for offsite storage. The number of unique discs needed is far more critical than the number of duplicates needed. Many systems are available for generating large numbers of copies, but your choices are much more limited when you need to create unique discs.

Platform issues are just as important as production issues. Where does the data reside that needs to put to disc? If the data resides on different types of platforms and operating systems (for example mixed UNIX and Windows environments), you will need to consider what happens to the data when it is brought from one platform to another. Windows may capitalize the first letter of a file name. Other potential sticking points when moving UNIX files over to Windows include file permissions and pipes, symbolic links, multiple "dot" file names, deep directories, and file names with white spaces.

Burning large numbers of unique CDs and/or DVDs means handling a large amount of data. Moving data from the computer's hard disk to the CD/DVD drive is an I/O-intensive operation. Setting aside network bandwidth issues for the moment, consider what happens on a hard disk drive (HDD) when you prepare files to be recorded on CD or DVD. Operating system tasks periodically require disk accesses, as do other tasks that are running concurrently. Now add in the disc recording process, which may require reading widely scattered files across the HDD, and also writing an image file on the same hard disk. The HDD heads seek to new positions on the disk constantly, and performance quickly degrades. There are ways to minimize this problem; adding dedicated hard drives is one. Shutting down other processes will also help, as will moving the raw data to be recorded onto its own HDD or RAID array.

If your production window indicates that you will need to have concurrent recording processes (i.e., you need to record more than one disc at a time), you will need a system with more than one CD/DVD drive. There are several manufacturers that provide CD/DVD autoloaders with multiple drives. Look for autoloaders that can handle both CD and DVD recording at the same time. Most autoloaders also allow the use of a printer with the autoloader, so the disc is labeled with unique identifying information as soon as it is recorded. For nearly all applications, this is an absolute requirement to avoid mishandling and losing track of which disc contains which data.

Autoloaders vary widely in construction, capacity, and capabilities, as do disc label printers. Some entry level autoloaders include inkjet-based printers as an option, while high end autoloaders typically include thermal printers. Thermal printers usually provide a much lower cost per imprint than inkjet printers, are usually more durable, and are also significantly faster. Printer speed is important, especially in autoloaders with two or more CD/DVD drives. The printer can become a limiting factor on production in these cases, as recording jobs have to stop until the printer is available.

Other important factors to consider are whether direct-attach or network-attach is most appropriate. In many cases, this does not matter, but sometimes direct-attach is an absolute requirement (for security, administration, or other reason). Other considerations include: How will system integration be accomplished? Is an API included, or necessary? In many environments, scripting is the preferred methodology for adding automated disc recording.

One leading provider of automated CD/DVD solutions, Young Minds, Inc. (YMi), has a variety of CD and DVD recording solutions that fit most applications. As mentioned above, YMi has developed disc spanning technology that can handle automated recording of large data sets and files. YMi's solutions are scalable and run on most Unix platforms, including AIX, Linux, IRIX, and Solaris. Both network-attach and direct-attach systems are available.