It seems the answer here is that, yes the supply chain may be long enough for untrustworthy entities to be part of it somewhere.
A regular database would be great if it was just farmer sells to Walmart and Walmart maintains the database that tracks it all.
It looks like the block-chain solution is the way to create that good chain-of-custody data. Especially when not all the parties have direct contact or even any direct business relationship.
I went back to read the article (stunning I know). Right at the top, it has this statement: “direct suppliers of lettuce, spinach and other greens to join its food-tracking blockchain by Jan. 31. The retailer also will mandate that farmers, logistics firms and business partners of these suppliers join the blockchain”
It looks like Walmart deals with direct suppliers. Those direct suppliers may deal with logistics companies or other indirect suppliers, and somewhere down at the bottom is a farmer. I can easily picture a bunch of farmers selling to a logistics company that sells to one or many processors, that in turn sell to Walmart. Where the farmer grows and harvests, but then hands off for transport. The transporter specializes in just moving from farm to processor. Then a processor that sorts, cleans and packages. Finally sold to Walmart who ships it to stores themselves.
In this scenario, from the story it sounds like the goal is for a farmer to make and entry about what’s delivered to transporter, the transporter to make an entry about the aggregation of farms and delivery to the processor, the processor to enter updates about how it’s processed and moved on to Walmart then. Today, the farmer may not even know their stuff is going to Walmart at all. With the mandate to make the block chain entry, they’ll know that at least some is going to Walmart, since that’s where the mandate is originating. Presumably, others that could take delivery at any of those steps could use the information too.
I’m usually super sceptical about anytime block chain is suggested to fix something. But, this one, with all the disconnected parties involved that all want a record, sounds like a good fit. Then again, some big player could just become the “master database in the cloud” and provide the same function as a trusted registry for all the data.
I’m guessing today it’s all done disconnected from each other between each direct party only, some of them probably only on paper. That’s why it takes so long to track back. An investigator needs to go to each step one at a time and trace it back. This solution, or the central database, would both help with the tracking speed.
Walmart is famous for its existing supply-chain and logistics management system (that tracks numerous international parties and middlemen in any given transaction) and is notorious for using that system as one of its cudgels against suppliers who displease it or who it wants to beat up on regarding price.
Put another way, they already keep a very close eye on everything that goes into and out of their stores from point of origin all the way to customer returns and have done so for decades. Adding a blockchain ledger doesn’t add much value to what these heavy-handed control freaks already have.
A few years ago my FIRST LEGO League team was competing in the “Food Factors” challenge, which explored safety/risks in the food supply chain. Without me claiming to be any kind of an expert, I’ll note that the lettuce supply chain sounds somewhat similar to the milk supply chain (what my team focused upon), and perhaps other supply chains share similar characteristics.
With milk, a number of original suppliers each contribute to a larger and larger collection of material. At any point in the chain contamination may be introduced - by an original supplier, by an intermediate processor, by a mistake in handling, etc. Tests to detect contamination are made at regular points in the chain; if a test at any point reveals contamination then a) the current batch of material must be destroyed, and b) an investigation going backwards in the supply chain (branching out to all of the suppliers) begins so that the original source of contamination can be located and corrected.
Does the blockchain provide a good way to capture this branching path? I don’t know, not having studied the blockchain myself. However, I can easily imagine that there might be some suspicion and distrust between the organizations involved, at least once contamination has been detected and the original source is being traced back.
No, it doesn’t but if you can add it without it causing you much cost or hassle, why wouldn’t you.
Seems to me this is exactly the sort of thing that blockchain is actually useful for.
Basically, it’s a PR move in the aftermath of the earlier contamination – “it’ll never happen again because blockchain.”
Looking at the article, it’s that and also “we want to be able to trace every leaf of lettuce from the farm to the consumer even if it’s been trucked across the country three times, conmingled with lettuce from all sort of other farms and turned into a ready meal” so that when there’s an outbreak at Farm A, we can assure everyone that these lettuces we’re selling are fine because we can prove they never came near those lettuces.
“Blockchain can and should be used to promote transparency around food safety,” said Natalie Dyenson, vice president of food safety and quality at Dole, which, with Walmart and eight food companies, founded the Food Trust group.Retaining consumer trust in product safety is valuable, she said. Beyond public health concerns, incidents like this year’s romaine problem can erode sales across the industry.
“We didn’t have any growers that were part of that outbreak. But when one fails, we all fail because customers don’t look at a head of lettuce and say, ‘Dole didn’t have a problem.’ They just don’t buy lettuce,” she said.
“Every single lettuce processor in the U.S., Canada and Mexico took a financial hit,” she said.
Somebody is using blockchain??? Where do I invest?
I think you misspelled ‘derailing’ there. It is spelt ‘enhancing’. 
Chipotle is another example where supply chain issues contributed to outbreaks of e coli.
It’s because the restaurant wanted to foster a fresh & healthy image that they moved most food prep into the local restaurants but this actually lead to unsanitary conditions as workers did not follow proper food handling guidelines.
Ironically, they were forced to prep more food ahead of time at central distribution centers and ship it pre-packaged to stores. During the outbreaks, they weren’t able to easily track down the source of contamination because ingredients were “pooled” centrally and they were unable to discern exactly which supplier was at fault.
I can see how blockchain can potentially offer some solutions once you get past all the “Bitcoin” hype.
https://www.bloomberg.com/features/2015-chipotle-food-safety-crisis/
Insert joke about $10,000 heads of lettuce here
See, this looks like one of those rare instances where blockchain as a technology might be useful, rather than in the production of Dunning-Krugerrands and inefficient space heaters.
Proving supply chains back to source, in a world where globalisation has stretched supply chains across the world, and into some places that are difficult to monitor seems to be a useful application of this.
We already know that establishing trust in your supply chain is non-trivial, and somewhing that adds value to the business as a whole:
Once you have a whole lot of people getting sick, a lot of inventory to destroy, a lot of fines to pay, and a class action lawsuit on the horizon, yeah, you can’t trust your partners.
This raises two questions for me:
- What blockchain properties does this system have
- Can I be a Miner and get free veggies?
Blockchain doesn’t magically make the data entered valid.
This is another variation on the computers make it better magic trick.
Signed entries in a distributed database with a terribly inefficient consensus algorithm.
Let’s be correct here.
I clearly don’t understand blockchain. I may never understand blockchain. How does someone adding a cryptographic number to a head of lettuce (or a group of lettuce in a pallet) help this situation? Does the number get added onto or changed as it passes through other hands? I mean, I totally understand how an TCP/IP stack works and how extra bits are added to a portion of the addressing to translate into “this packet was sent from here, went through here, and also here, and is heading there… when there replies back, please send back through here, and also here, so it gets to here.” Is that how the blockchain works? Because if not, this makes no sense to me. If the suppliers in between the farmer and the ultimate destination aren’t adding to that blockchain, this is just a database issue that doesn’t resolve anything and sounds like technobabble to try and improve consumer confidence, not resolve a difficult problem. You still need everyone in the chain of delivery to record “this pallet with this block chain came through here.” That’s just database entries, and you can do that with any number, you don’t need blockchain.
Again… I clearly don’t understand blockchain, so maybe I’m missing some magic that someone can explain to me.
Hmm. As I understand, unlike a Distributed Hash Table, each broker/node has to have a copy of the full blockchain*. And the Bitcoin blockchain is 173 gigabytes-ish now, but if it was seriously used as a currency, it would get bigger real fast. (I have no idea how the copies are all kept current. Not my problem.)
522 billion transactions a year translates to 1.4 billion transactions a day.
This is equivalent to 9,722,220 transactions every 10 minutes , which coincides with how often blocks are published to the Bitcoin blockchain.
Assuming transaction sizes stay around the same size, at 250 bytes, this means that every block would hold about 2.4 gigabytes of data .
This transaction volume would generate about 350 gigabytes on the blockchain every day, or 127 terabytes every year.
For lettuce tracking, I guess they could fork the blockchain every year and drop the old moldy records. (Archived, not lost, but not part of the current chain.)
I can see that blockchains are going to be driving the storage and transmission speed technology.
eta: (reads on further in the article) and processor crunching tech, oh yes!
eta2: * for a full node. Lesser nodes can get by with part of it and go fish for rest via a DHT.
If they are that untrustworthy I’m a trifle unclear on how Walmart intends to make this work.
If you want multiple mutually untrustworthy parties to be able to pass around abstractions in such a way that double spending/replay attacks aren’t generally viable that’s a thing you can do. (And an interesting one; preventing double-spending of things that copy with the pristine perfection of digital data is neat.)
The half of these omg-blockchain-logistics arrangements I’ve yet to hear anyone explain is how you deal with the fact that preventing tampering is at best incomplete (at worst actively unhelpful if it lulls people into a false sense of security) if people putting lies (or sincere beliefs that have been tampered with downstream) into the system is a risk.
If people lying to you isn’t an issue you only need to make sure that your DBA isn’t in the pay of a ruthless lettuce cartel and your problem is pretty much solved. The only other vaguely novel tool you might use is GUIDs, so that you can serialize the produce without coordinating blocks of values among your vendors ahead of time.
If people lying to you is an issue, then enshrining their deceptions in future immutability is of limited benefit; especially in a case like this where there is a fairly clear authoritative party that trusts itself and whose suppliers are obliged to cooperate with.
