Sigcat Software Showcase 1993/Sigcat 93 Software Showcase dataDisc.ISO/research/library/docs/scchp10.asc

Text File | 1993-02-08 | 25.1 KB | 662 lines

                      CHAPTER TEN

  THE RELIABILITY AND INTEGRITY OF COMPACT DISC MEDIA




      CD-ROM AS A LONG-TERM PRESERVATION MEDIUM

                    Charles Dollar
     National Archives and Records Administration
                     Washington DC


The presentation opens with a brief discussion of the
interest of the National Archives and Records
Administration (NARA) in CD-ROM media. The mission of the
National Archives is to preserve and make available the
permanently valuable records of the Federal Government,
which largely have been in paper form. However, new
Federal agencies are increasingly using new information
technologies to carry out their programs and
responsibilities. The major challenge that new
information technologies cause for the National Archives,
and ultimately for all users, is that we are dealing with
technology independent records. A fundamental component
of technology as we understand it today is innovation
that fuels innovation.
     The following quotation from De laude scriptorium
(In Praise of Scribes), written by Johanes Trithemius
(1462-1516), contains an interesting comparison of the
longevity between two information technologies--words
written on parchment with words written on paper:

     All of you know the difference between a
     manuscript and a printed book. The word
     written on parchment will last a thousand
     years. The printed word is on paper. How long
     will it last? The most you can expect a book
     on paper to survive is two hundred years. Yet,
     there are many who think they can entrust
     their words to paper. Only time will tell.

     The point of citing this quotation is that it
highlights a fundamental problem of technology dependent
records, whether it is pen, ink and parchment, a printing
press, or optical media. No information technology will
last forever. Trithemius was content to let history judge
which of the information technologies was superior. The
critical issue for electronic information technologies is
that we cannot wait for the passage of time to confirm
the longevity of storage media and information
technologies.
     One section of the presentation tries to clarify
what is meant by the term "archival record." Essentially,
the word "archival" is very ambiguous when used in a
longevity context. Consequently, the American National
Standards Institute is reviewing changes in certain
standards that delete the word "archival" and substitute
"life expectancy" of the storage medium. In line with
this approach, it is appropriate for CD-ROM vendors and
users to discontinue the use of "archival" and
"archivability" when describing the longevity of CD-ROM.
     In recent years, several people have raised
questions about the long-term reliability of the CD-ROM
medium. Most of the reported problems appear to be
associated with the loss of the seal coat integrity. In
order to acquire a better understanding of the issue of
long-term reliability of CD-ROM, the United States
Geological Service (USGS) and the National Archives
organized a working group made up of representatives of
Federal agencies to meet periodically to review issues.
The Working Group hopes to foster more investigation into
methodologies for evaluating the longevity of CD-ROM.
     The key problem from the perspective of an archivist
for all optical media, including CD-ROM, is that too much
attention has been focused on the longevity of the
storage media when the central issue is facilitating
access over time to technology dependent records. In this
context, the presentation introduced maintaining the
readability and intelligibility of electronic records as
the primary vehicles for facilitating access over time.
     The presentation concluded with a brief summary of
three critical issues for CD-ROM technology. The first
critical issue involves media questions and the need for
an industry-wide accepted test methodology for predicting
life expectancy. The second critical issue is knowing
when to recopy CD-ROM. This requires understanding the
function of error detection and error correction code
(ECC) techniques used in CD-ROM and providing end-users
with specific information about the level of ECC
activity. The third and last critical issue for CD-ROM
technology is the development of alternative technology
migration strategies. Given that technology obsolescence
is inevitable, vendors must assure that users must have
open systems and interoperability across different
technology platforms minimize the cost and difficulty of
migrating the information on CD-ROM to newer generations
of technology.

Related Graphics to this paper:

%g DOL01.pcx;
%g DOL02.pcx;
%g DOL03.pcx;
%g DOL04.pcx;
%g DOL05.pcx;
%g DOL06.pcx;
%g DOL07.pcx;
%g DOL08.pcx;
%g DOL09.pcx;
%g DOL10.pcx;
%g DOL11.pcx;
%g DOL12.pcx;
%g DOL13.pcx;
%g DOL14.pcx;
%g DOL15.pcx;
%g DOL16.pcx;
%g DOL17.pcx;
%g DOL18.pcx;
%g DOL19.pcx;
%g DOL20.pcx;
%g DOL21.pcx;
%g DOL22.pcx;
%g DOL23.pcx;
%g DOL24.pcx;





        FACTORS AFFECTING CD-ROM RELIABILITY

                     Ron Kushnier
 Chairman, SIGCAT CD Reliability & Integrity of Media
                       Committee
        Naval Air Warfare Center, Warminster PA


The Naval Air Warfare Center, Aircraft Division in
Warminster PA has conducted several tests on optical
media, including CD-ROM for use in harsh environments.
These tests have been published and are available for
distrbution on a limited basis. An excerpt is presented
here.


QUALITY CONTROL DURING MANUFACTURE

Regardless of the manufacturing process, the strict
discipline of high quality control is mandatory in the
production of CD-ROM. The lack of proper quality control
procedures can be divided into two broad categories and
result in the following effects.

A. Electronic Contamination. Due to power-line
fluctuations, spikes, noise, etc., and errors caused by
electronic data storage and transfer from one media to
another.

B. Chemical and Mechanical Contamination. Due to chemical
impurities in the water, cleaning solutions, plastics,
etc., as well as excess mechanical vibration and other
machine process variations during manufacture.


EXTERNAL EFFECTS ON CD-ROM MEDIA DUE TO ENVIRONMENT

A. The environmental agents working against CD-ROM
reliability are the following:

     1. Time
     2. Temperature
     3. Humidity
     4. Corrosive solvents,label inks, and adhesives
     5. Ultra-violet, infrared and nuclear radiation
     6. Mechanical stress
     7. Use and handling
     8. Molecular breakdown of substrate and thin
     film layers due to aging

B. The mechanical and electrical effects as a result of
these agents are as follows:

     1. Loss of disk transparency and coloration
     2. Loss of reflectivity
     3. Increase in birefringence
     4. Delamination of thin film layers
     5. Delamination of label
     6. Migration of label ink into reflective layer
     7. Destruction and delamination of seal coating
     8. Warping of substrate
     9. Increase in defects due to scratches, pits
     & microcracks
     10. Oxidation and corrosion of reflective layer
     11. Complete physical destruction of the disk
     due to thermal stress and cycling
     12. Loss of servo tracking
     13. Loss of optical focus
     14. Increase of electrical noise level
     15. Loss in signal-to-noise ratio
     16. Increase in error rate
     17. Increase in data access time
     18. Optical head abrasion from disk medium
     19. Disk physically jamming in drive


VARIATION OF ENVIRONMENTAL EFFECTS

A. The environmental agents listed above have a varying
degree and effect on CD-ROM performance. The Time factor
plays a natural role in the gradual degradation of the
disk. Changes can be seen with the application of heat,
humidity, and corrosives in under 150 hours on some
media. The tests conducted at the Naval Air Development
center exposed CD-ROM disks to 80 centigrade (176F) at
90% relative humidity (R/H) for approximately 150 hours.
Similar tests done by Industry, on a number of sample
disks, have produced similar results at a less stringent,
60C (140F) and 90% (R/H) in the same time frame,

B. Compact disc product manufacturing techniques vary
greatly among different companies. The manufacturing
process for some CD-ROM's do not provide good
environmental protection. Other discs, whch incorporate
sufficient protection, can still function after many
thousands of hours of extreme temperature/humidity
exposure.


CATASTROPHIC CAUSES OF DISK DESTRUCTION

A. Observations, based on environmental test data,
indicate several possible causes of catastrophic CD-ROM
disk failure.

1. Penetration of contaminants into the metalized
reflective layer

a. This can result from manufacturing the metalized layer
out to the edges (both inner and outer) of the disk,
thereby not providing a hermetic seal to the environment.

b. The seal coat acrylate plastic or epoxy layer,
(approximately 10-20 um thick) is applied over the
metalized layer on the disk by a spin-coating process. It
is usually then, UV cured. Ineffective seal coat material
and it's application has been shown to be the "weak link"
in CD-ROM longevity. Breakdown of this material from heat
and humidity has been judged a major cause of disk
failure.

2. Paper labels for CD-ROM have been shown to peel in
high heat and humidity environments, taking the seal coat
and reflective layer with them.

3. Improper silkscreen and labeling inks can produce a
migration effect into the metalized layer, thereby
destroying the disk.

4. Extreme Thermal Stressing. Delayed reaction of optical
media to thermal stressing has been documented by more
than one source. A short time after one company cycled a
CD-ROM through the full MIL-E-5400 temperature extremes
(-54C to 125C), the disk, while idly sitting on a desk,
shattered severely. Disk warpage from extreme heating is
another cause of failure.

5. Radial scratches on a CD-ROM, especially over the
Table of Contents data, may render the disk unreadable.


APPENDIX: CARE AND HANDLING OF CD-ROM DISCS

The following guidelines represent the current thinking
for the care and handling of CD-ROM discs, by a number of
CD-ROM disc and drive manufacturers. The validity and
usefulness of most of these
guidelines have not been substantiated by Government
testing and
therefore are presented for information only.
     Wash your hands before contact with the disc. If
available, wear lint-free cloth gloves, finger cots, or
talc-free latex gloves. If you must wipe the disc, do so
with a soft, dry, lint-free cloth in a radial motion --
that is, from the inner to the outer hub -- not in a
circular motion around the disc like you might do for a
phonograph record. The most devastating scratches are
those which occur along a radius of the disc which can
obscure a long stream of pits.
     Certain cleaning agents and solvents can damage the
discs. Some of these include: gasoline, paint thinners,
benzine, acetone, carbon tetrachloride, chlorinated
cleaning solvents, ammonia, and household detergents
which contain ammonia. Do not clean with a water soaked
cloth. The use of isopropyl alcohol, the ingredient in
many commercial CD cleaning products, as well as certain
waxes and acrylic liquids, is still questionable.
     Do not clean the label side of the disc. Use of a
CD-ROM caddy is highly recommended during transport and
operation. Limit the amount of physical contact with the
disc. Always handle the disc by the outer edge and/or the
inner (hole) edge. Never touch the data surface.
     Discs like to "live" in the same conditions that
people do; that is:

      They don't like to be manhandled

      They don't like exposure to temperature extremes

      They don't like exposure to excess humidity

      They don't like exposure to high intensity UV
light


Related graphics to this paper:

%g KUS01.pcx;
%g KUS02.pcx;
%g KUS03.pcx;
%g KUS04.pcx;
%g KUS05.pcx;
%g KUS06.pcx;
%g KUS07.pcx;
%g KUS08.pcx;
%g KUS09.pcx;
%g KUS10.pcx;
%g KUS11.pcx;
%g KUS12.pcx;
%g KUS13.pcx;
%g KUS14.pcx;
%g KUS15.pcx;
%g KUS16.pcx;
%g KUS17.pcx;
%g KUS18.pcx;






            WHAT IS THE EXPECTED LONGEVITY
              OF RECORDABLE CD-ROM MEDIA?

                   Akio "Alex" Iida
                 Taiyo Yuden, Inc. USA


   ARCHIVAL VERSUS QUALITY CONTROL CD-ROM TESTING

                     Denis Oudard
                       Digipress


Digipress has developed a compact disk specifically for
long term archiving and preservation. The Century-Disc is
made of etched tempered glass and gold. The purpose of my
talk however is to share Digipress' experience in
conducting life expectancy tests. I will not go into the
details of the tests, which are available here, but
rather to present some of the lessons we've learned while
investigating life expectancy issues.

      Tests that predict life expectancy for a
     specific medium, do not always work for all
     media and materials.

     Tests that predicted fairly well the life-expectancy
of paper failed to predict the vinegar syndrome of
cellulose based microforms. The microform had been
predicted to last 500 years. We now know that microforms
kept in an office environment are plagued by a sudden
decay mechanism around age 45. We do not know what might
or will go wrong in new material, therefore we do not
know how to test for it.

      Test conditions which would seem to be
     harshest are not always the ones we think.

     The conditions for getting glass to show iridescence
are 42C (107F) and 100% humidity, with condensation. Any
other temperature, whether higher or lower will not
produce the same results. In other words, why should 75C
(167F) and 75% humidity be the conditions that accelerate
the aging of polycarbonate, organic dyes and lacquer? It
does not even provoke iridescence in glass.

      Aging tests (Arrhenius and Eyring tests)
     only look at part of the picture.

     Aging of medium is a combination of chemical,
mechanical, optical and other reactions and of resistance
to accidents. Aging tests using the Arrehnius and Eyring
models only try to measure one of the parameters:
chemical stability. For a medium which one hopes to keep
for a hundred of years, there exists the possibility of
it encountering disaster. Therefore a certain ruggedness
is necessary if one is serious about long term archiving.
One should consider resistance to shocks, water and
floods, temperature shocks, temperature extremes, ultra
violet light and other rays, acidity, etc... To test
these conditions, we have used various tests, choosing
standardized test from different industries, when
available.
     Finally, I would like to contrast two different
types of tests: aging tests and quality control tests.
The two are often confused because they make use of what
could seem to be a similar methodology.
     Quality control tests are designed to insure that
certain parameters during manufacturing have been met.
For example, if the laquer of a CD does not cover the
aluminum layer completely, eventually the aluminum will
oxide and loose of its reflectivity. Placing such a CD in
a warm and humid atmosphere for a few days will expose
the problem rapidly. This simple QC procedure has never
been a way to predict that CDs or CD-WO will last for 86
years. Even if we were to accept Arrhenius or Eyring
valid aging methodologies, which I have suggested were at
best very incomplete ones, we would need to conduct a
battery of tests. Measurements are taken at appropriate
intervals, carefully choosing the parameters one wants to
track and their indicators and using different batches
that are tested at different temperatures in order to
estimate a multiplication factor to calculate a "normal"
condition life expectancy.
     It is encouraging to see that there is an effort by
people involved in standardization to take these elements
into account. Until these standards are approved however,
the warning of caveat emptor rules the laser waves.

Related Graphics to this paper:

%g IID04.pcx;
%g IID05.pcx;
%g IID06.pcx;
%g IID07.pcx;
%g IID08.pcx;
%g IID09.pcx;
%g IID10.pcx;




    ARCHIVAL LIFE EXPECTANCY OF 3M CD-ROM MEDIA

                    Bob Schoonover
            3M Optical Recording Department


CD-ROM has enjoyed considerable growth and acceptance
over the last several years, as all of us at this
conference know very well. We all rely on the information
distributed on CD-ROM's to be 100% error-free from the
first time we read the disc until the data is obsolete.
Recent articles seem to imply that all CD-ROM's will
self-destruct over time, and have raised awareness and
concern over the use of CD-ROM as an archival media. A
comprehensive article in the Wall Street Journal last
fall questioning the lifetime of CD's is a prime example.

CD-ROM Manufacturers Durability Comparison

Unfortunately, we have found these concerns are not
entirely invalid. This data covers the performance of
CD's made in Japan, Europe and the United States after
stress-testing at 60C and 85% relative humidity. Changes
in BLOCK- ERROR-RATE were monitored as a function of time
in the chamber. Some discs were not playable after only
100-200 hours while others continued to perform very well
after several thousand hours. We concluded that ALL CD'S
ARE NOT MADE THE SAME.

The Problem

This helped us to define the problem...to be able to
estimate the lifetime of CD- ROM's. The topic of CD-ROM
lifetime has become more of an issue in the last several
years, as end users and customers ask us about lifetimes.
There are no useful real-time data for archival storage
for this relatively new technology, and we find there are
no definitive technical studies on CD lifetime. We
therefore set the objectives for our study to provide
these capabilities:
     -- estimate life expectancy of 3M CD-ROM's at any
     reasonable combination of temperature and humidity.
     -- define the percentage compliance of our disc
     population.
     -- determine statistical confidence limits on the
     estimates.
We commissioned Bill Murray in our Research Laboratory to
initiate an appropriate study to develop a statistically
sound, technically comprehensive determination of the
life expectancy of 3M CD-ROM's. I will provide a summary
of the methods used and some examples of the results of
this study. The complete test methodology will be
presented by 3M to an ANSI subcommittee involved with CD
life expectancy evaluations in Montreal on June 4-5,
1992.

The Eyring Model

The Eyring equation is an accepted textbook model for
estimating product life expectancy, and has two prime
advantages for this study:
     -- it provides for the simultaneous modeling of two
     variables; we selected temperature and humidity as
     the most critical environmental factors involved in
     long term storage of CD-ROM media.
     -- the Eyring equation has been derived directly
     from the laws of thermodynamics, and thus provides
     a sound scientific basis for the model.

The Solution

The specific test cells or chamber conditions are listed
in this table. All of these higher test conditions are
intended to accelerate chemical changes related to
archival storage that would occur much more slowly at a
more reasonable temperature and humidity ... but not to
determine how the product performs while at these
particular high conditions.

Accelerated Lifetime Testing Conditions

The conditions for each chamber are noted here. All discs
were tested for 4000 hours total residence time in the
chamber, and each cell contained 16 discs ... a total of
144 discs. The change in mean BLER was selected as the
response for this model. You will note that the matrix is
not symmetrical ... there are 4 different RH's tested at
the high temperature, but only one at the low temperature
... and two at the in-between temperatures. Also, the
time interval between BLER testing was shorter for discs
at the high temperature and longer for discs at the low
temperature, to allow for different acceleration factors.
This model and the analytical methods used easily
accommodate these differences.

CD-ROM Lifetime Study Results

This is an example of the resultant test data from just
one of the cells. This does not tell us anything about
lifetime, but is simply the raw data. You will note that
these 3M discs had a mean-BLER of about 2 at the start,
and increased slightly over the 4000 hour test. The
Industry Specification allows a maximum BLER of 220 while
3M's Internal Manufacturing Specification allows a
maximum BLER of only 20 counts per second.

Estimated Years To Reach BLER Of 50 and 220
This is an example of the type of results that were
obtained from this study. In this case, we are looking at
the lifetime estimates for a fairly warm storage
condition of 30C (86F) and 90% relative humidity. The
horizontal axis is a logarithmic scale of life-expectancy
in years, ranging from 100 to 4000 years. The vertical
scale is labeled "survivor function" and simply defines
the percentage of the disc population that is in
compliance with the estimate. The two solid curves
represent two different END-OF-LIFE assumptions. The
right curve assumes END-OF-LIFE when the mean BLER
exceeds the Industry Specification of 220. We believe
this is not a reasonable assumption, since the disc will
certainly exceed 220 in some locations when the mean BLER
equals 220, and the data may very well not be
recoverable. We prefer a much more conservative approach,
using an END-OF-LIFE of 50 as shown by the left curve.
This is less than one-fourth of the Industry
Specification and in our experience the disc will be
playable with the data recoverable. Finally, the two
dotted curves represent 95% confidence limits on the
lifetime estimates. Thus, this model does meet the
original objectives of estimating life expectancy at any
reasonable combination of temperature and humidity,
defining the percentage compliance of our disc population
and stating the statistical confidence limits on the
estimated lifetime. As an example of how this data can be
used, note a line drawn horizontally from 60% compliance
or 0.6 on the vertical axes to the first dotted line.
Reading down to the time scale, the estimated lifetime is
about 350 years. This technical conclusion may be stated
as:

     With 95% confidence, we estimate that 60% of 3M
     CD-ROMs  will last 350 years at 30C (86F) and 90%
     relative humidity without exceeding a mean BLER of
     50.

This sounds impressive, but in fact contains a
significant caveat. An estimate based on 60% compliance
does not answer the question "what about his disc? ... or
what about hers? ... or what about MINE?" The next visual
addresses this concern. The data is from the same model,
but shows only the percentage of discs in compliance from
95% to 100% ... the very upper portion of the data on
this graph.

CD-ROM Lifetime Study Results

Now, if we select a percentage compliance of 99.5% at the
very top of the graph, we can see that the estimated
lifetime is equal to 100 years, by following the
horizontal red line to the 95% confidence curve. We are
now clearly covering the critical issue of "which disc"
in defining a lifetime. Any lifetime estimate is
meaningless unless this fraction of the disc population
involved ... the percentage compliance ... is clearly
included in the model AND is large enough to include
virtually all of the discs. We said earlier that "all
CD's are not made the same". That comment may also be
applied to all discs made by any one manufacturer, as any
of the media vendors would agree. There is variability in
all things, and we need to address that variability in
defining product lifetimes.

Lifetime Statistics Prediction

The concluding statement that we would now make from this
study is:

     With 95% confidence, we estimate that 99.5% of 3M
     CD-ROMs will last 100 years at 30C (86F) and 90%
     relative humidity without exceeding a mean BLER of
     50."

It has been our intent throughout this program to be very
conservative in lifetime estimates ... thus, we use 95%
confidence limits; we define and include 99.5% of our own
product; we select a fairly warm and humid storage
condition of 86F, 90% RH and we base the lifetime
estimate of 100 years on an END-OF-LIFE equal to less
than 1/4 of the Industry Specification maximum BLER
value. As a practical matter, we would recommend storing
3M CD-ROM's in a cool, dry location rather than the
higher conditions of 86F and 90% RH used for this
particular example.

3M's CD-ROM Warranty

This study resulted in 3M recently providing a 25 year
warranty on all 3M CD-ROM media ... our commitment to
continuing high quality and performance in our products.

Related graphics to this paper

%g SCH01.pcx;
%g SCH02.pcx;
%g SCH03.pcx;
%g SCH04.pcx;
%g SCH05.pcx;
%g SCH06.pcx;
%g SCH07.pcx;
%g SCH08.pcx;
%g SCH09.pcx;
%g SCH10.pcx;