Backing up onto physically isolated and protected data storage works. Nothing else ought be trusted.
Waaay back in the 1970’s some arrogant clowns with rich contracts from an important USA national security agency came up with what they claimed were “mathematical proofs of correctness” that a very clever form of electronic communication via linked computers (the ARPANET) was foolproof and unhackable. (“To hack” as meaning “subvert the security of digital or other electronic systems” was at that time at least 15 years old, I believe originally used in the jargon of the pioneer madmen who “blueboxed” ATT touch tone phones so as to make free phone calls.)
I was part of a team hired on the cheap by a few sane skeptics at that same security agency to actually test those “proofs of correctness”, and we suceeded in demonstrating that the Arpanet (soon to go public as the Internet) had extreme security issues.
A decade later I was working as both a tech adviser and a gay rights activist when the vicious evil immoral self-righteous religious bigots of the Reagan era controlled Justice Department took deadly aim at us, with concentration camps as part of the agenda (really!). Partly because of my digital technical skills, and partly because of love, trust, and friendship, I managed/controlled the largest politically usable digital mailing lists of gay people in the USA (primarily the membership rolls of the Metropolitan Community Church [ironically by far the largest gay lesbian organization in the USA at that time], a variety of gay Atheletic Leagues, and the subscribers to the activist newsmagazine The Advocate).
We NEVER trusted the data to net communications, but built and updated the data bases by floppy drives (yes, that long ago) carried by volunteer private couriers.
I had the hard drives with the data fully disconnected and physically stored in a locked vault, for use only with constant human presence of a security guard and a lawyer specialized in privacy law during each step from attaching and using the databases to updatge and print the addresses until the mail was taken to the USPS where a “friend of Dorothy” postal employee quietly received them and physically mixed them with other mail at 2 AM in the morning at the Los Angeles central postal facility.
Worked great. We now live in wilder times, and anything other than my level of paranoia ought to be suspected of being unprofessional if not crazy.
OK. I asked an AI generator to think about this, in one case for one days receipts for VISA.
**Answer:**
To store all VISA transactions on paper for a day, it would fill approximately 2,500 file cabinets.
To calculate this estimate, we need to consider the following factors:
Number of VISA Transactions Per Day: According to Visa Inc., there are over 150 million daily transactions globally.
Size of a Typical Transaction Record: A single transaction record may vary in size but can be estimated at around 1 kilobyte (KB) to account for transaction details, timestamps, and other necessary information.
Paper Size and Storage Capacity: A standard file cabinet drawer can hold around 2,500 sheets of paper, and a typical file cabinet has four drawers.
Calculating the total amount of paper needed:
Total Transactions Per Day: 150 million transactions
Size of Each Transaction Record: 1 KB
Total Data Generated Per Day: 150 million KB = 150,000,000 KB
Converting KB to Megabytes (MB) for easier understanding:
1 MB = 1,024 KB
Total Data Generated Per Day (in MB): 150,000,000 KB / 1,024 = 146,484.38 MB
Now, let’s convert this data into the number of file cabinets needed:
Number of Sheets per File Cabinet Drawer: 2,500 sheets
Number of File Cabinets Needed: 146,484.38 MB / (2,500 sheets * 4 drawers) ≈ 14.7 cabinets
Therefore, to store all VISA transactions on paper for a day, it would fill approximately 2,500 file cabinets .
Simple multiplication puts a year’s transactions at:
2500 x 365 = 912,500 file cabinets.
I suppose I could ask the machine how many warehouses that would fill - just for ViSA, and then extrapolate to include MasterCard, AmEx, Discover, and then all the hospital records, Amazon purchases, real-time inventory transactions by private companies etc, but I suspect the answer would be “You need to cover every square inch of Montana, every year, in file cabinets to accomplish your goal.”
While some might think so little of Montana that would be a worthy goal, it would seem to me that the idea of “file cabinets” should be voted off the island, and other ideas entertained.
Absolutely, at least until China or Russia or Iran in a fit of pique erases Visa’s electronic database and shuts down the US economy for a month.
Wake me when you get over your snark and actually suggest a solution to the cybersecurity problem. I’m not holding my breath, but stranger things have happened.
I only reserve my snark for idiotic ideas, of which “file cabinets” is certainly one.
I have already listed proposed solutions, which are as simple as air-gapped and redundant backups, and heightened security such as is provided by any number of firms, both public and private. The costs of insurance and/or negligence will force companies into taking the problem more seriously or into going out of business.
Writing everything down on paper is as likely a solution as telling everyone to use flint and stone to make sparks every time they want to light their stove.
On the other hand, having old technology like matches and candles isn’t a bad thing when the power goes out.
Your argumentation style is to take an extreme example (those billions of Visa transactions) and assume it applies to everything. It doesn’t. Take your average university department. It could just keep a paper record of its students’ social security number, courses, and grades as back up (like what was the norm 50 years ago). It could then assign each student a unique number for online data storage and transactions. If the online stuff gets stolen or erased, the students’ SSNs aren’t compromised and their academic records can be rebuilt. I bet that ends up being cheaper than developing and continually upgrading a triple redundant online backup system.
I think institutions should identify the minimum information required to keep their core business running and then as much as possible back that information up on paper.
I bet that all those clinics dependent on UnitedHealth to store their data are right now looking for all the paper records they can find to keep their business running.
UHC serves 6500 hospitals and 1,300,000 physicians, with over 27,000,000 customers. I’m not sure you are visualizing how many “paper records” that would be over the course of a week, a month, a year. You are thinking about the thousands of ones they might need in the next month or two, I am thinking about the billions of pages that would have to be generated over a few years just in case something like a data breach would happen.
Stored on paper, even on microfiles or microfiche, it’s a prodigious amount of data we generate every day , and we’re long past the point of being able to cope with it in hard copy format.
There are some areas where that’s not obsolete. Perhaps car titles or land titles which change only every 5 or 10 years, but for hospitals with data streams on every procedure of every patient going through every doctors’ office or urgent care center or hospital every day? You might as well go back to delivering freight by Wells Fargo Stage Coach.
The right answer is: harden the new technology against the threats, just as we hardened railroads against roaming outlaws, airplanes against skyjackers, and stock transactions against swindlers. It won’t always be perfect, but at least we didn’t load the improvements down with things that made them impossible to use.
It is pretty much the same number of paper records that the same number of hospitals and physicians dealt with before electronic health records (EHRs), which became mandatory in 2009. How was your health care before 2009? Perhaps my memory is bad but I don’t recall hospitals failing because they were overwhelmed with paper records in the 20th century.
A consequence of EHRs is that it became cost-effective for hospitals to outsource data management. This resulted in patient data becoming concentrated in a smaller number of very large databases, creating high value targets for bad actors.
You say that UHC has 1.3M physicians connected to their record management system serving over 27M customers. That’s 1.3M potential entry points for malware and over 27M valuable data points to sell on the dark web. Good luck keeping that secure for any length of time.
Would also note that in the aftermath of a ransomware attack, health centers have to use paper records (like the old days) for medical transactions until the EHR system is restored. This means that prudent health facilities need to maintain a paper record infrastructure just in case. Florida hospital cyber attack: Good Samaritan, St. Mary's return to paper charts
In short, since health facilities worked just fine with paper records twenty years ago and must maintain the ability to use paper records in case of a cyberattack, I don’t really see any unresolvable problem with using paper records as a last resort back up system.
Is it possible you might not have an accurate view of the world?
This is article from a medical provide whose Electronic Heath Record (EHR) provider was hacked by Russians. He was able to stay in business but many of his colleagues were not. One of the elements that helped his practice survive the ransomware attack was an arrangement with…
“…a data extraction company which pulls data from every patient record each night and prints a paper copy of each patient’s International Classification of Disease (ICD-10) diagnostic codes, recent laboratory work, gaps in care, and the most recent updated list of patient medications. This document is known as a point-of-care (POC) report. The data extraction company had a server onsite which was not connected to the cloud-based provider and therefore was inaccessible to the Russians and their ransomware. As a result, we had accurate information on patients dating back to 1 day (October 30, 2021) before the ransomware attack. This proved invaluable as we had a mini version of each patient’s chart in paper format.” You Have Been Hacked! | Annals of Family Medicine
It was connected and that’s why they made a printed copy. Anything connected to a compromised system is vulnerable. Why else do you think this data extraction company made a paper copy rather than just store the data on a hard drive?
If you read about these ransomware attacks in many cases the malware was in the system for months before the threat was known. That means that malware was probably backed up to all those backup systems. So sure, you could back up your electronic data on some flash drive and store it offline in a bank vault, but if the malware has been backed up as well, you are still screwed when you restore the backup.
So what you need is a backup system that you know with certainty can identify the ransomware and not back it up. Maybe one can do it by backing up data on excel files with no macros. But at that point you are getting pretty close to paper records.
And again, if you look at how all these institutions are recovering from a ransomware attack, particularly those that bravely do not pay the ransom, they all go back to a paper record system until the get their electronic stuff cleaned up. So as I mentioned before, a prudent company will still have to keep in place the resources and infrastructure to use paper records if they want to be prepared for the worse case scenario.
Most of that was fiction. If you ask him for a few corporate names that have used paper to reconstitute their records that would be interesting to see.
The bigger issue in “medicine”, his expertise, decisions can not wait. Not that he has ever made any of those decisions.
Never wrong, conversation killer. This can now drag out for weeks.
There is an economic problem with blockchain. Without mining it to offset costs there is no point economically to do it.
It would be a good idea. If the economics were picked up by a clearinghouse it would fly with fees to hospitals and doctors.
That does not mean blockchain can not be broken. Which means there is more cost to running the blockchain.
I do not know the economics of it played out. I doubt many experts have planned it out well. It is an engineering problem that would take a lot to measure out.
Forgot blockchains have low request levels. That stymies a lot of things. IPFS is not good with larger files and many requests.
I don’t think anyone disagrees that a doctor’s office can store paper records. My own doctor, until 4 years ago, stored everything on paper, in files, on shelves all over his office (including 2 exam rooms), and then “went digital” at the start of COVID. A think a typical doctor might have 1000 or 2000 patients.
I think the assertion is that a health insurer, with 30 million customers, can’t store everything on paper. Certainly not reliably and in a timely manner.
I bet Blue Cross used mostly paper records in 1969. Was health care timely and reliable for Blue Cross insurance holders in 1969?
In any case, you have to remember that the proposal is to use paper only as a backup of last resort, as a means of regenerating enough of a lost database to stay in business without having the bad guys win by paying a ransom.
Let’s say from that info that one can assume data stored for a year are clean of malware. Suppose your 30M customer insurer creates a paper copy at the end of each quarter of all transactions made during that quarter. It also recycles paper records older than a year on the assumption that the hard drive backups of this data are clean. Under this system the company would have the most recent year backed up by paper and would minimize losses due to a worse case online disaster to the data from the current quarter.
Ah, this is the meat of the matter. I asked my doctor why he is making the huge effort to “go digital” (and it was indeed a huge effort) and he said “well, at this point I can’t submit most things on paper anymore so anything on paper has to be typed in anyway before submitting”. And the difference between 1969 and now? Well, the sheer volume of stuff that is recorded and stored. I’ll give you a trivial example - in 1969, after falling and ripping my chin open, I needed stitches. My mom took me to the docs house/office (it was one and the same at the time), and my mom and the nurse held me down while the doc stitched me up. Took a few minutes and that’s it, it was over. Nobody wrote anything down other than a slip of paper with the bill on it. Probably 10 bucks or so, because it was less than 5 minutes of work (even including coming back in 10 days to remove the stitches which the nurse did on her own). Blue Cross had nothing to with it, BC was limited to “real” major medical stuff. Not a single form was submitted to BC for my stitches. Today, on the other hand, if I take one of my kids to urgent care for a fall, there are 20 “codes” that are submitted, from the visual evaluation, to the ubiquitous X-ray, then the measurement of the cut, then the stitching (the needle, the thread, the labor, etc), then the antibacterial ointment, then the bandage, etc. There are form to fill out, and there are all those codes submitted to insurance. Every one of those steps has a document associated with it, and every one has to be stored. And if any of the codes are entered incorrectly, it’ll be denied, and then a whole new round of submissions will occur. Then, when I bring the kid back a week later, there are a whole set of codes for removal of stitches (“Suture removal in a primary care office following an ED visit”, code 15853, and maybe 15834, etc). And ALL that stuff has to be stored.
