The Imaging Jukebox Blues

By Rich Payne

Do you still have a jukebox? No I am not taking about a Wurlitzer. This is about massive storage of DVDs. Recently I spotted a question on Information Zen (www.informationzen.org/forum ) from a user requesting information about jukeboxes. The most common answer to his question was, Why?

During the early days of imaging jukeboxes were essential and ubiquitous—mainly because of the high cost of magnetic storage. The jukeboxes used for imaging were first cousins to the original music jukebox. As those of us a certain age recall, jukeboxes stored vinyl records, would pick your record with a picker arm, place it on a turn table, where a needle would read the groves in the vinyl. Voila, music—for a nickel a play (for those of you under 40, you’re welcome for the history lesson). The modern data jukebox stores DVDs. When the software requests data, the correct DVD is selected and placed into a reader where a laser translates it into data or documents so that information can be viewed by the user. Most DVD Jukeboxes have multiple readers.

Initially the data jukebox was used as the primary storage location for the early document imaging systems. Because of the mechanical nature of the jukebox there could be a delay of 15 to 30 seconds between images. Then we began to use RAID (Redundant Array of Inexpensive Disk drives) to store the latest information so users could view it faster. Some sophisticated operations used something called pre-fetch software to pull expected data from the jukebox in advance and place it on the RAID. Eventually as the cost of RAID dropped the need for jukeboxes was diminished and when it cost more to purchase a terabyte of jukebox storage than it did for a terabyte of RAID we stopped purchasing jukeboxes.

Just because we have stopped purchasing jukeboxes doesn’t mean that many of us still have them around and in use. Many of us are at a point where we have to make important decisions about our jukeboxes. Here are some things to take under consideration.

Why replace your jukebox? As your jukebox gets older the mechanical functions become problematic, sometimes the actual DVDs will begin to disintegrate (this happened to my company and we discovered that a batch of DVDs produced one month in 2004 was poorly manufactured, disintegration did not start until 2006 when we tried to display 2004 documents. ). Another important consideration is the software that operates your jukebox. Our software purchased in 2001 is no longer supported and new software is prohibitively expensive and not certain to work.

All that being said many of us are facing a decision on retiring our jukeboxes. I am sure that there is someone out there that has ‘special’ circumstances but for the rest of us here are some jukebox recommendations.

Don’t purchase another jukebox. You should take a look at the cost of RAID versus Jukebox,can’t think of any circumstances where a jukebox would be more cost effective. Don’t be swayed by outdated standards or requirements which require jukeboxes. Security considerations can be easily overcome, just talk with your auditor or information security officer. Legality is not an issue as far as media but Records Retention is important. The security considerations include proper disaster recovery plans and documented destruction of outdated documents. Most auditors and electronic security officers understand that most procedures concerning jukeboxes are outdated.

Stop feeding the beast. Take action to put your ongoing data into a RAID. In some cases a RAID exists and companies are still blindly following obsolete procedures to put backup DVDs into the old jukebox.

Make plans for your jukebox.

If you jukebox contains information that has a short retention develop a plan to remove it from service as soon as the retention is met. Meet with your auditor and information security personnel along with data owners and develop a consensus on its retirement.

Replace your jukebox if it contains information that has a retention schedule that is long term. You will have to decide what defines long term. Permanent certainly, ten years yes, five years possibly. A rule of thumb is to calculate how long it will take to convert your data to RAID and if the retention has more than five years after you have finished converting then convert it.

Develop a plan to convert your data. This generally means copying your data to the RAID. You need to make sure that you have RAID space and converting data takes time. Be sure to include documentation for data that cannot be converted or has become corrupt. Data conversion can be very time consuming and is usually done after hours. I remember a rolling carrousel that we had in the South Carolina Retirement system which contained an alphabetical cross reference on two by three inch cardstock. Years after we had an electronic alphabetical cross reference we still printed and filed these little pieces of paper. It took me six months just to stop feeding the beast and it saved twenty man hours a week.

Rich Payne has been in the industry for over 19 years and is noted for his insightful, practical and humorous presentations. He also has a collection of 12” glass optical disks which came from an antique data jukebox.