Another PC question- backup hard drives

[Thomas Rogers]

Richard L,



Please recognize that I respect your opinion and your experience, and you, and I expect the same FROM you.



With that said, why so argumentative? Note, that I'm making no real assertions here. I'm not trying to prove anything. I'm just repeating information presented by experts. As I do that, I get confronted with words like "so which is it?", which come off a little *********gerie, no? I'm not trying to deceive. I'm not flip-flopping.



2 to 5 years? 10 years? Which is it? Maybe both. Seriously!



To clarify:



The first article said 2 to 5 years. Agreed. However, the Wiki article states what you and I would no doubt agree on; that reliability and life expectancy can vary greatly depending on the quality of the optical disc one purchases. Lastly, for there to be an average of 10 years shelf life, if we agree thats a good number, it means that there will be some that last longer, and some that last less. The question is "how many" last longer, and less, and HOW MUCH longer and less. 2 to 5 years for a cheap CD is not unexpected. I've seen it happen, personally, but who cares about my own experience. Just like you, my personal experiences does not an authority make. So, 2 to 5 years can happen with cheap discs. 10 years on average for all discs (cheap, good, better, best).



I'm sure we both understand bell curves, and life expectancy bathtub curves. For there to be an average of 10 years, it's not unreasonable to expect that some cheap discs fail after only two to five years.



Like I said, I respect you and your advice. I'm not disbuting your general advice. Optical is a fine backup medium...just not a good "decades long" archival medium. That's all I was trying to say. If you were implying that optical is a good decades-long archival medium then that would be improper advice. I never said CDs aren't viable media for backup. By all means, use them. Feel free. As you said, they are better than nothing...and only a fool would argue with that.



The truth is, however, that CDs age. They do have a shelf life, and it is shorter than most think, and that shelf life is greatly shortened if one buys bargain CDs. Most people don't know this. They think they can backup to a CD, then in 2, 5, 10, 20 years they can grab the CD off a shelf and read it. The reality is that just isn't the likely case. The longer that a burned CD sits on a shelf, the more likely it will go bad. That's just the plain fact and very few experts dispute this.



The fact that some data on a backup CD might only be useful for up to 7 years isn't necessarily the main issue. For example, digital photos are timeless, and to many, priceless. The simple fact is that photos in a shoebox in a closet can last for decades (I've got some about 60 years old from when my parents were teens). There is no burned CD out there that I'm going to bet will last that long...not even 1/5th that time.



Lastly, you say:

Most of the experts who are warning us about optical archiving media are simply looking at it from a businesses perspective who are mandated by law, or for possible future law suits, to keep records for many years beyond what the average individual may need for their home PC.

Can I aske where a blanket statement like that comes from? You lose all credibility with me when you claim to be able to say with authority why most experts say what they say. From my POV there is simply no way that you can know what motivates most experts to say what they do.



Bottom line: Are CDs good for backup? Sure! Are they cheap? Sure! Are they better than nothing (no backup at all)? YEP! But they are not the best thing out there for archival storage that has to last decades. They just aren't. There are dozens of experts that will show you this. That's not to say one shouldn't use them. Like you and I said, they are better than nothing.



Backup your data folks. If at all possible, back it up to at least two different mediums: online + optical; optical + external HDD, etc; whatever combination works for you. Recognize that no single backup medium is 100% reliable (optical discs go bad over time; online storage systems can go belly-up when their hosting company goes bankrupt; an external HDD drive can crash). The most important thing is to back up...at very least to ONE medium and even better, to two. If you choose ONE, and it is optical, you might want to check the integrity of the optical backups every couple of years, especially while you still have the source available.



Good luck folks. Backup early, backup often.



TJR

 

[Richard L]

TJR,

I am not being argumentative. You made the statement that experience did not matter, and questioned why I would even consider CD's as a viable data backup media because of their unreliablity. So I told you that I have not experience that kind of unreliablitly that the expert clained in the article.



Like I said, there are all kinds of experts in every field who cry wolf and make exaggerated statements based a what they deem is valid research. My opinions are based on 44 years of computer experience for both the government and corporate world.



I worked on large IBM mainframe computers for the first 25 years and always thought they were a top notch company. Later when I got involved with other brands of computers, it became obvious that IBM makes some great hardware, but their software, and their operating systems are designed to only perpetuate IBM's image. They attempt to make everything proprietary and is what caused them to loose out on the PC market immediately after they created it. Years ago IBM would never even acknowledge that they had any competition for the computer market. If you did not want to buy IBM hardware, they would go to the president of your company and claim that you were trying to destroy the company because you were not buying or leasing the best equipment in the world (at least in their eyes)



In reality, there is no backup media with an average life expectancy of over 10 years. Businesses will often need longer file retentions and so they will transfer their data to multiple medias to ensure they have a viable copy should it be needed in the future.



You are the one who provided the link to the German IBM expert who stated that CD's may only be good for 2 years. I said that my experience was that I had CD's over 10 years old that still worked. Then you provided more experts who now claim the average live of a CD is 10 years. So if the shortest life of a CD is 2 years, and 10 years is the average life, that would indicate that there are CD's that could last for 18 years? And I certainly don't agree that magnetic tape is the best media for backup/archived data. Tape is only used because of it's ability to store very large volumes of data on a single reel or cartridge.



10 years may not be long enough for some critical government or business applications, and perhaps even a few home PC users, but for the vast majority of home PC users, that is more than enough time. And that's what we are talking about here...Home PC users. Most PC users do not even backup files and the few that do probably could not even find a 2 year old CD backup.



You are being argumentive. I just gave my opinion based on my experience. I think that bothered you more than the fact that I mentioned using CD's. You asked me what the life expectancy of a CD was... 2, 5, 7, or 10 years, Which is it? Those were the different answers your different experts claimed, so obviously the experts can't even agree so I feel I don't have to agree with any of them, since I have CD's over 10 years old. I have no scientific evidence to say why my CDs last that long, so I guess I must just be lucky. :grin:



...Rich

 

[Gavin Allan]

I do know that 8-track tapes suck for long-term storage. About a year's worth of life is doing pretty good with them.

 

[Thomas Rogers]

RichardL,



Had a nice long post to respond, but it got eaten, so I will paraphrase.



I'm not being argumentative. Seriously.



Writable CDs degrade. That's really not up to for debate, IMHO. The debate is how quickly. Cheaper writable CDs degrade more quickly. That seems to make sense to me. So, whether it is 2, or 5, or 7, or 10 years for a writable CD shelf life; well that's an interesting debate. But let's agree that that debate really is about WHEN not IF a written CD degrades. The fact that it does degrade, and that there are some very reputable sources that warn against counting on a shelf life beyond a few short years makes CDs not the best long-term archival storage.



That's all I was saying.



I wasn't trying to question why you even consider using CDs for backups. That's not what I was saying, or at least not my initial surprise. Given your self-professed decades of experience I found it surprising that you didn't at least caveat the use of CDs. Most everyone that I talk to that is "in the know" will warn people against the false-sense of security that if something is sucessfully burned to a CD that it will be available for years and years to come. It may be available for 2 years, or 5 years; maybe even 7 or more. But don't count on it being available much beyond. Likewise, don't be surprised if that CD you burned just a few years ago can't be read when you take it off the shelf. It happens. It really does happen.



You also missed my point about the 2, 5, 7, 10 years thing. The life expetency of a written CD varies based on many factors, including storage conditions, the quality of the CD burner used to burn the CD, the quality of the original disc, etc. These things accounts for the differences...the range. I've had cheap CDs that I couldn't read within just a couple of years (I verified their correctness when originally burned). I've had others burned long, long ago that still seem okay today. I don't see the so-called experts disagreeing when I see different numbers quooted. But then, I don't typically (that's a good word, typically) think or debate in absolute terms. I don't need a SPECIFIC number that all agree to in order to take value from the overall message/warniing. I understand that there are different situations that can make all of those values valid.



If people think they can burn a disc and it will be readable in 5, 10, 15 years, I believe they should be warned. Their data simply may not be there then...or even a lot earlier.



Thanks for reading,

TJR

 

[Richard L]

TJR,

I guess we will just have to agree to disagree.



I think the variability of the CD life is based on the quality of the manufacturing process which leads to poor recording/burning accuracy. I have known people who have bought spindles of 50 or 100 blank CD's and 80% were recorded with serious errors that could not be reread within minutes after the CD was burned.



That's whyI said that it is essential to verify the files were burned to the CD without any errors or corruption. Most tape backup systems do the verifying automatically, but CD's don't. To verify the tape data, they reread the tape and compare the tape data to the original data byte for byte to ensure the data was copied exactly as it exists on the original.



I am curious if some of the claims of "Degradation of CD's" is caused by data being recorded with serious errors in the first place? Since the experts never explained how they tested their theorys, we will never know.



...Rich

 

[Thomas Rogers]

RichardL,



Actually, the experts do explain why, how, and when the degradation happens. Its caused by breakdown of the the dyes within the rewritable CDs; those same dyes that turn opague when hit with the write laser during the burning process. These dyes break down and getting cloudy over time. This issue happens for all rewritable CDs and DVDs due to their very nature and physics/chemistry of the composition.



Manufactured (aka pressed) CDs and DVDs don't exhibit this same type of degradation. The pits those discs are actually stamped in place, not completely unlike the physical grooves on an old-fashioned LP are stamped in.



That's why the experts claim the degradation that they claim. Its not because there are errors in the first place (and note, some CD duplication software DOES actually do write verifications, especially those that burn from an ISO image). The issue is that the inherent nature of the technology exhibits for degradation over time. It's a chemical process.



There really isn't anything to disagree with. The dyes break down. The question isn't IF they do. The question is how long it takes, and how long until a burned disc will be unreadable...and that varies based on a multitude of factors.



TJR

 

[Richard L]

I understand what they claim in their theory, I just don't know what kind of tests, how many test, when were the test done, etc. and how they proved their theory is correct. Then how did they determine exactly how long a CD would last? From what I read it only appears to be a theory and does not take into account that many CD's last far beyond their theoretical life span? It could be something as simple as the quality or chemical purity of the plastic used to make the CD and some brands of CD's may use a more optically pure plastic that does not become opaque from the laser heat? Perhaps they tested CD's manufactured in China where quality control is pretty much non-existent?



Simply being a so-called-expert, and having a theory does not make that theory right. There are plenty of doctors with medical theories, and plenty of physists who have theories, and scientists who have theories, and astronomers who have theories. and as soon as one of them discloses their theory a dozen other experts will have a different theory. Even when the experts agree on a theory, they are often proven wrong by someone actually does some valid testing.



I can accept proven facts, but rarely agree 100% on just theories. I have a lot of my own theories on many subjects that I cannot prove they are right, so I feel my theories are just as valid as the experts who cannot prove their theory.:grin:



...Rich

 

[Thomas Rogers]

RichardL,



You are getting it now. There are too many variables to be able to come up with one single number. Also, I think it is you that is calling this a theory. I think there is harder science at play than that, but, heck even gravity is just a theory....but I still tie things down to avoid the damage that theory can produce. Lol



BTW, even the optical storage technology assoc (see link) is non-specific with longevity numbers for the reasons I mention. It is intereting that they state that unrecorded CD-Rs have a 5 To 10 year shelf life. My understanding of CD rot is that burning helps to slow down the dye chemical changing process, again, the question is how much.



The wiki entry is good reading too:



http://en.wikipedia.org/wiki/CD-R#cite_note-5





TJR

 

[Richard L]

TJR,

I never thought CD's were good for eternity and the theory that the laser may damage the plastic is very likely a good and valid theory. I have heard of those theories before but they never provided the proof or scientific evidence that their theories were correct. Providing their exact testing procedures would lend more credence to their claims and might uncover some overlooked flaws in their theory. Sound like a job for MythBusters..:grin:



I just have to question why there is such a wide range of life spans estimated by these experts, and how they derived at those times? That's the part that I disagree just based on my experience.



When someone says that CD's are good for only 2 years and I have CD's that are over 10 years and probably more like 12 years old, I have serious doubts about their claims. Then another expert claims that the "Average Life Span" for a CD is 10 years which certainly agrees with my experience, I have to question those who are claiming only a 2-5 year life span.



The manufactuing of CD-R's has always been a quality control issue. When 80% of some CD brands cannot be read immediately after burning the data, tells me that there are quality issues. I have experienced a number of quality issues with new CD's but no where near the 80% failure rate others have encountered (probably less than 10%). Perhaps it is because I rarely fill an entire CD with data, but typicaly only burn a few dozen MB's at most. However most of my MP3 CD's are pretty much filled with as much music as I can cram on them and even after 5-6 years living in my car they still work just fine?



...Rich

 

[Thomas Rogers]

RichardL,



Don't take this the wrong way, but for someone supposedly learned in this industry I don't understand why you continue to assert that the 2 years versus the 10 years thing are somehow at odds.



Again, if you were following and reading the links I provided, the 2 year failure was attributed to cheap recordable CDs. The generally accepted shelf-life of recordable CD (CD-R) is more conservatively stated at 10 years. Both of these can be true, and the fact that there is a range isn't a reason to discount these assertions.



As for hard science, that does exist.



There are labs that qualify recordable CD shelf life (longevity), and they do so according to a set of ISO standards (ISO 18927:2002, updated by ISO 18927:2008). See the link below. This particualar set of testing just deals with temperature and humidity. There are other factors, including light, and "time-dependent flow phenomena"



Take a look at the link below as well:



http://www.itl.nist.gov/iad/894.05/loc/overview.html



(in particular, take a look at the technical background section, copied below):

Technical Background

Recordable optical disc media contains an organic dye layer whose transparency can be altered either to absorb a laser beam or to allow the beam to pass through to a reflective layer behind the dye [1,2,3]. The nature of this organic dye is such that when the internal energies of its molecules reach a particular threshold, an irreversible chemical reaction occurs, and the dye layer loses its transparency. This property allows a high-energy beam to ewrite' data by burning epits', in the form of dark marks, to the disc during recording. A low powered laser reads the data by either passing through the transparent dye layer (without causing any molecular change) to the reflective layer or by being absorbed by the nontransparent marks in the dye.



Due to the organic nature of the dye, degradation and breakdown of the transparent portion of dye layer will occur over a long period of time as a natural process. This process, which has its roots in chemical kinetics, can take several years in normal environment conditions [4]. Higher temperatures and humidity will accelerate this process by increasing the thermal and kinetic energies of the dye molecules.



The effect of these processes can be modeled using various techniques including the Eyring model, which has a theoretical basis in chemistry and quantum mechanics and can be used to model acceleration when many stresses are involved. For this study, the Eyring equation will be used to model the effect of two stresses, temperature and relative humidity, on the rate of a reaction or degradation, which can be related to the time-to-failure of the optical disc.



The Eyring model has found broad application and shall be the model for estimating life expectancies of Recordable DVD systems. The equation



t50 = AT aeH/Te(B+C/T)RH



where



t50 is the time to 50% media failure;

A is the pre-exponential time constant;

T a is the pre-exponential temperature factor;

H is the activation energy per molecule;

k is Boltzmann's constant (1,38 ~ 10 -23 J/K);

T is the temperature, K;

B, C are the RH exponential constants;

RH is the relative humidity, %;



was derived from the laws of thermodynamics and, in this form, can readily be seen to handle easily the two critical stresses of temperature and RH. For the temperature range used in this test, it is common practice to set a and C to zero [5].



The Eyring model equation then reduces to



t50 = AeH/kTe(B)RH

That's pretty technical, but it does give me some assurances that this "theory" is based in sound science. The dyes degrade. How quickly is a factor of environment and composition. Better construction of higher quality and better storage will mean longer life; less of either means lower life.



TJR

 

[Richard L]

TJR,

That last technical article is the missing link that you never mentioned in your previous posts.



I was not arguing that CD degrade over time, my question was why the experts seemed to disagree as to fast the degradation occured and what the ultimate live span of a CD really was. That's why I suspected that the 2 year life span was linked to the early CD's and that they have improved the quality of the plastic and dyes. That article is obviously over my head but appears to support the theory that there is some finite life to a CD.



Also feel that for the majority of Home PC users that CD's are a safe an viable backup media, but they should also know that there is a degradation of the media over time that should be considered when determining how long they wish to preserve their data.



...Rich

 

[Thomas Rogers]

RichardL,



Yep...that's what I was trying to say too: CDs are a safe, cheap, reliable way to backup your data, but steer clear of the bargain brands and also recognize that even the best brands won't last for decades.



TJR