Hacker News Reader

Comment by rkagerer | Hacker News Reader

Inferior endurance is the "dirty secret" of the consumer-level SSD industry. You and I know better, but the average consumer doesn't pay attention to write endurance or under-provisioning.

While I'm no proponent of storage mining, I do feel a little empathy. First they sell products that don't stand up as well to wear, then they wash their hands of it when you actually want to use the thing. Feels a bit like when stainless steel appliances gave way to plastic ones.

That's why I stopped buying entry-level, and prefer enterprise SSD's or at least good brands that publish endurance figures and have held up to third-party testing proving them out.

Here are some fun experiments writing to SSD's for days/months/years on end to see how long they'd last:

[1] https://techreport.com/review/27909/the-ssd-endurance-experi...

[2] https://forums.whirlpool.net.au/thread/9rl1lp19?p=9#r179

[3] http://www.xtremesystems.org/forums/showthread.php?271063-SS...

I have no problem with a manufacturer publishing a lifetime write rating and denying the warranty once you've exceeded it. But unless they were up-front about it, amending the terms afterward by saying "well, we never intended you to use it like that" seems a cop-out.

View on HN

Consumer 100Mbps internet will also not match datacenter 100Mbps internet. Consumer coffee maker wont match industrial coffee maker when you try to make coffee continuously, it will overheat the sensor area and start making bad coffee and eventually fail much sooner than the industrial one.

It all comes to the use case scenarios. It's not really a dirty secret but cost issue where you can pay up more for robustness and buy the industrial use one if the consumer grade stuff doesn't fit your needs. On 99.99% of its users it's fine, almost everyone is happy with a consumer SSD.

I’m down with a consumer coffee maker not working as well as an industrial coffee maker, but 100Mbps is a well defined thing, and I don’t understand how it isn’t criminal that my internet speeds never even approach the advertised rates. This isn’t an issue of my “use case” not requiring honesty, this is an issue of me paying for something and simply not getting it.

I've never heard or experienced anyone not getting their advertised speed (apart from DSL that varied depending on quality of the cables etc. or wireless services). There are indeed laws that guarantee that, depending on where you live.

But you can't really expect to max out 100/100 24/7 all year long on a consumer pipe. It is overprovisioned and assumes that most won't be utilizing theoretical maximum all the time. Just as the coffee maker.

It's not only about the bandwidth but other factors like latency between regions etc.

Also, the prices given for consumers assume that they are not going to max out the bandwidth all the time, so the ISP having 1000Gbps line to the rest of the internet can sell it to 1000 customers, giving them each a 10Mbps. It will work fine up until a lot of them are watching Netflix at the same time or torrenting.

The moment you have guarantees on the speed, your prices and speeds are adjusted to make it feasible(the same hypothetical ISP can have 2 customers at 100Mbps, 30 at 10Mbps and 500 at 1Mbps with guaranteed speeds and vastly different prices). In many places you can definitely pay a lot of money and have a connection that is not pooled with your neighbour.

It's the same difference as with the shared hosting and a dedicated machine hosting.

> the prices given for consumers assume that they are not going to max out the bandwidth all the time

Consumers should be free to violate any and all assumptions with no negative consequences. Especially unstated ones that only people familiar with their business model would know about.

If an ISP advertises 1 gigabit/second internet, consumers should be able to fill that pipe with 1 gigabit of data every second. In practice only a minority is going to do that but it should be possible with no punitive throttling of bandwidth whatsoever.

Right, if they think customers need less then they can just advertise less. They shouldn't get the benefit of advertising a big number while providing a small number just because most customers can't tell. That will make it impossible to fairly compare them with their competitors.

Sure, customers can do that. That's how you get quotas, fair usage limits, higher prices when the average usage pattern diverges from the intended use cases before the technology or infrastructure upgrades catch up.

The one caveat is that today all ISPs advertise their plans as "up to" a certain speed very prominently in all marketing material. It's no even fine print or byline. The "fair usage" clause is the kind of thing that's buried deep to hide that they may employ throttling.

Think of a "service" you provide. If your employer was expecting you to deliver peak performance 8h per day, 5 days per week, for the rest of your employment with no exception, would you still accept the salary you get now, or would ask an order of magnitude more? It's just that your employment contract and law introduce some "fine print" to allow downtime, and the chance to make mistakes or run at lower efficiency with no (immediate) termination.

Why is the same so unacceptable when it comes to a product or service you pay for?

> Why is the same so unacceptable when it comes to a product or service you pay for?

...because it could be described and measured precisely. And it's not.

My friend, there are very few things that we can describe precisely, from a legal standpoint or even for technical specs. It's all about "ranges". And the higher the precision and the guarantee, the higher the price, or the converse. Which was my point, in case I couldn't manage to make it clear enough even when putting it next to something an average person could relate to. There's leeway built into any legal or technical framework and the contract terms should follow.

Your "precisely defined and measured" SATA3 6 Gbit/s interface never delivers the 6.000Gbps to you by design (8b/10b encoding). So are all SSDs and motherboards in breach? SSDs rarely hit any of the advertised speed. By comparison their endurance is defined far clearer for the consumer.

Whether it's an SSD or an internet connection they come with clear predefined conditions that are legal* and the customer agrees to. Everything else is high consumer expectations meeting low consumer willingness to pay, or even to understand the relationship.

*Of course there are plenty of illegal and abusive clauses but giving a stated 900Mbps average speed with a 1Gbps contract, or selling an SSD with stated low DWPD don't really fit that description.

Don't shift the goalposts. You started from a very broad claim:

> Why is the same so unacceptable when it comes to a product or service you pay for?

...and now you are talking about "ranges".

Datacenters and ISPs for commercial use can provide very detailed SLAs.

Consumer ISPs refuse to publish metrics or have 3rd party services monitor them to create accountability.

They refuse to provide the most basic SLAs and refuse to adopt any user-friendly quality rating.

Meh:

> The one caveat is that today all ISPs advertise their plans as "up to" a certain speed very prominently in all marketing material.

I literally started how describing something as "up to" [ostensibly a range] is not outright misleading, if at all. It was also strictly speaking of marketing material and how well you can convey info in one banner.

Seems you're not a stranger of moving goalposts and dodging question. You're selectively arguing some of my points, while so elegantly avoiding others because "oh that's different". Which is a very immature way of pretending to have a meaningful conversation, and a good way of hinting you have no counter for that one.

You're doubling down on your misconceptions. Whatever "they" do it's because it's legal. And I'll be blunt: your definition of what is moral and what things should mean is irrelevant. If you want to be right at any cost you can just keep inventing personal definitions to apply to others but not to you because "that's different".

I made my point at quite some length. If that didn't do it, nothing will.

> "If your employer was expecting you"

Firstly you are comparing performance guarantee of a person vs a router. Secondly employment law is a different from consumer law. It's a separate can of worms, with specific clauses for discrimination, etc.

The reason this is important is to limit fraudulent claims, if my business is selling a 1 GB/s internet connection that actually only reaches the speed once a month, and someone is selling 1 GB/s internet connection that works reliably, our marketing material and contract might look the same. I would be defrauding the consumer and my competitor.

I am comparing 2 situations where laws set the framework and contracts specify the relevant details within that framework. The contract or terms of use usually contain the exact definition of the service and the conditions it's offered in, and the SLAs. So I'm not entirely sure how these being "different" changes anything for the purpose of the discussion. The contract sets some terms within the framework set by the law. This is valid whether it's a router or a person. The only thing you could point at is that the marketing material is misleading. But...

GP's claim was based on the same misunderstanding you make now regarding that marketing material:

> if my business is selling a 1 GB/s internet connection

You're doubling down on the same mistaken assumption. I randomly browsed over ISP websites just now and they all sell "up to xGbps" connections. Marketing material anyway can't cover the breadth and scope of an actual contract. So it's the job of the law to decide what "up to x" can mean (you deliver that speed at least 60% of the time for example), and for the terms for the specifics of that particular contract.

Just because we don't like something doesn't mean they're breaking the law or that there's a moral obligation on one side to satisfy all claims for the other.

No marketing material is 100% accurate because you can't cram all the things that need to be said in one punchline. But if you enter an agreement based on that one marketing claim then surely the problem isn't with the provider.

>"contract or terms of use usually contain the exact definition of the service and the conditions it's offered in, and the SLAs"

I would love to see this fantasy world - I have been with 5 different residential broadband providers in the past 8 years, not a single contract contains SLAs nor do they give any indication of how their speeds or latency stack up in practice. Calling the Virgin Media customer line does not produce any usefull information either, and their sales staff cannot answer basic questions.

Even worshippers of 'mah free market' should be able to see the problem: I was looking for the best provider money can buy, and I ended up stuck in year-long contracts with companies that actually provide a crap service, because the only way to know is trial and error. My money should have gone to the best company instead.

>"Just because we don't like something doesn't mean they're breaking the law or that there's a moral obligation"

I think we have a moral obligation not to defraud others. Casually I define it as misleading people into parting with their money or causing them to suffer losses. To me that includes selling herbal cures that you know are fake, homeopathy, or inviting people to a restaurant telling them you'd pay and then taking off as the tab arrives. None of these activities are actually crimes.

> I would love to see this fantasy world

There's no need for condescension especially in this case. I live in that fantasy world where I get exactly what I pay for. It's not a guarantee that every country has the same consumer friendly laws but even here it's perfectly acceptable to sell an "up to 1Gbps" connection that falls shy of the marketing.

> not a single contract contains SLAs

If they didn't it's because you chose to pay less for one that has no SLAs. I have a consumer and a business contract from the same provider at my home address. The consumer one offers higher speed for lower price but offers no guarantees except the legal baseline. The business connection on the other hand lists for example minimum, maximum, and "normal" speeds, clearly states that there is no data cap or throttling, and the SLA for restoring connectivity is within 8h, and more.

The only way for marketing material to tell the truth, the whole truth, and nothing but the truth is to be the contract itself plus the relevant law. Anything else will be concealing some information.

> My money should have gone to the best company instead.

The problem is that you have a low limit for what you're willing to pay for, but a high limit on what you are asking for. It's a typical consumer education issue.

> I think we have a moral obligation not to defraud others.

Look, no offense but this is rant territory. You use your personal definition of "defrauding" to make the case that suits you best. Your SATA interface never actually reaches 6.0Gbps, your SSDs almost never reach the advertised speeds, your electricity isn't exactly 120/240V or exactly 50/60Hz. Your steak isn't exactly 200g. Your car doesn't consume exactly 7.9l/100km. and not all McDonalds chicken nuggets are the same size.

If it helps you understand the point, try to apply the same standards to yourself to see why guarantees cost.

Show me one marketing banner that you consider valid and let me nitpick. We'll either find out that no banner in the whole world is honest, or that you can nitpick anything.

"The problem is that you have a low limit for what you're willing to pay"

You are missing the point: money was never the problem. My current provider is miles better than my previous one, but you could not tell that by comparing contract or marketing material. So I could not select the best firm, so they lost out on business, so the market is not functioning as it should. You brough up SLA, I don't need an SLA for SLA's sake.

I think it is unfair to compare a different market entirely

>"your electricity isn't exactly 120/240V or exactly 50/60Hz"

Your residential internet provider can just throttle you willi-nilly. On the contrary, electricity gris has to follow very strict rules, and if the voltage drops by half we call ot a brownout. The entire country comes to a halt and executives have to answer some tough questions.

Similarly car milage cones from tests inaccordance with specific methodology, they cant just slap a number that's wrong by a factor of 2 and call it a day, but Virgin Media can.

> On the contrary, electricity gris has to follow very strict rules

The electric grid is regulate as such, to define the acceptable ranges. ISPs are regulated differently but the law still defines acceptable ranges for their service. They either don't prohibit, or even explicitly allow saying "up to 1Gbps" while offering lower for example. If you find the practice bad look to change the law. Internet should be regulated like a utility.

You make a case for having a crappy ISP experience but not for them deceiving you because you bought an "up to 1Gbps" connection and get less than the upper bound. You could get some guarantees but you didn't want to pay. Most things you pay for deliver lower than the upper bound, by how much determines if it's fraud or not, depending on the law.

> car milage cones from tests inaccordance with specific methodology, they cant just slap a number that's wrong by a factor of 2 and call it a day

It's very ironic that until just now you insisted "the consumer is being defrauded with fraudulent claims" and now you justify equally unrealistic numbers in other scenarios like car fuel or pollution numbers (don't get me started on how they measure PHEV fuel efficiency, factor of 2 would be great). So you willingly accept equally fraudulent claims if they're obtained in an irrelevant standardized test. Do you really think Virgin Media couldn't pass a "standardized test" with flying colors? By your own admission it's no longer fraudulent because they did a speedtest against their own servers and it came out great.

How many times will you switch tracks like this and just throw stuff against the wall to see what sticks? It looks to me like your argumentation is very poorly thought out, and you wobble between conflicting ideas, or misunderstood ones. You moved the goal posts around so much neither of us knows where they'll pop up. I get you have a bad ISP but their problem isn't the marketing slogans unless you go for the most superficial of reasoning.

> The business connection on the other hand lists for example minimum, maximum, and "normal" speeds, clearly states that there is no data cap or throttling, and the SLA for restoring connectivity is within 8h, and more

What stops them from providing this information to consumers? They can just measure their network performance and provide these statistics.

> today all ISPs advertise their plans as "up to" a certain speed very prominently in all marketing materia

Where I live they don't.

> Why is the same so unacceptable when it comes to a product or service you pay for?

Because consumers deserve more protections in general. ISPs know exactly what they're doing, they shouldn't be able to mislead consumers by promising performance they can't deliver.

Then expect to pay significantly more.

TANSTAAFL

I think this is unrealistic. Very few consumers are going to have high utilization of their bandwidth. The connections are going to be mostly idle. It makes sense from an efficiency standpoint to then share that resource among more customers. That way you can bring down the price.

Look at the cost of business internet. It's guaranteed. If it says 100 Mbps then you're going to get it with effectively no interruptions. But you're also paying 5-10x more than a consumer connection, because the ISP can't resell some of that connection that you don't use.

all that may be true. but capturing reality and providing sales terms that reflect it is the responsibility of the seller. one that they have clearly abrogated in the interest of extracting more money from the market.

> Very few consumers are going to have high utilization of their bandwidth. The connections are going to be mostly idle. It makes sense from an efficiency standpoint to then share that resource among more customers.

That's okay. What's not okay is suddenly punishing users who for some reason decided to use a lot of bandwidth because they were under the impression that they could.

>Also, the prices given for consumers assume that they are not going to max out the bandwidth all the time, so the ISP having 1000Gbps line to the rest of the internet can sell it to 1000 customers, giving them each a 10Mbps. It will work fine up until a lot of them are watching Netflix at the same time or torrenting.

I don't think this is universally true. I have a gigabit line via a residential ISP ($60/mo) in Europe and it will max out 24/7 without throttling.

What I don't have is access to the same support and service guarantees that would be included in a business plan.

Question: Will it still not throttle if all the other customers of your ISP are also trying to max out their gigabit lines at the same time?

I can't say for certain because I do not work there, but in practice even at the busiest times I've never seen speeds drop.

The big difference between the consumer and the industrial grade product of the same specs is usually in the guarantee of reaching those specs consistently, and in more extreme conditions.

A consumer laptop screen might do just well in the dashboard of a car but it might also turn yellow. You can't complain because it wasn't meant to be used like that (like operating in 80+C temperatures). Your vacuum cleaner would certainly not last long if you started vacuuming water and construction debris, the kind of stuff an industrial vacuum is built for.

A consumer SSD is not designed to be used for this kind of mining. Regular consumers are more price sensitive and ignore other aspects because they're not tech educated. Many don't even understand what this "mining" entails, just that they could make some money. There's a minority of consumers who went for the cheap stuff and now want to make money off of it by pushing it beyond the stated limits. I honestly don't think they have a leg to stand on.

It's a good thing that not all products are made to meet all expectations because they would cost all the money, all the time.

Agreed for SSD, but some consumer 1Gbps broadband never reaches 900Mbps. Its spec should be corrected rather than "best effort".

This is why the law is there. In many countries the advertised speed has to be within a certain range of the average, or hit that speed a certain percent of the time. So 800-900Mbps average, or hitting 1Gbps 60% of the time could very well fit a "1Gbps connection" tier from a legal perspective.

This is why ISPs advertise "up to" a certain speed and then go on to detail the exact conditions in the terms.

It's ironic OP doesn't complain about SSDs never reaching the theoretical and advertised interface speed. Which is no different from internet speeds not reaching the network interface speed as many others here call an example of being cheated. SSD endurance is far better described than it's performance for the average consumer: warranty expires after "this much" data written in "this much" time.

> It's ironic OP doesn't complain about SSDs never reaching the theoretical and advertised interface speed. Which is no different from internet speeds not reaching the network interface speed as many others here call an example of being cheated.

This comparison is not exact. I haven't seen SSD manufacturers advertise transfer speeds of the interface. They say they support this interface, but the transfer numbers advertised are different. Whether those actually happen is a different story.

In the case of internet connections, they don't usually advertise the "interface speed" either. They advertise actual bandwidth[0]. Yes, nowadays, a lot of providers will advertise 1 Gbps connectivity, which coincides with the interface speeds with which people are familiar. But for example, in my case, advertised speed is 400 Mbps. Over a PON (passive optical network). My transceiver (optical interface) will support much more than that. I've never had a 400 Mbps network interface in my computers. Should I expect them to give me 1Gbps? Turns out, they usually do, but I guess that's more of an exception than the rule.

The issue is when a provider will advertise say 400 Mbps, like in my case, but in practice I'm never able to reach more than say 40 Mbps. In this case, yes, there may be some kind of law based on which this behavior is "lawful" or not.

---

[0] Which may be "wire" bandwidth, as opposed to "useful data" bandwidth, so they'll happily count encapsulation, etc, but I'd argue that's somewhat different, although still scummy.

> This comparison is not exact. I haven't seen SSD manufacturers advertise transfer speeds of the interface.

I went to the first consumer grade SSD manufacturer I could think of, Corsair, and lo and behold, the description of the SSD advertises the interface speed [0] as "up to" 1,950MB/s write and 4,700MB/s read.

[0]https://www.corsair.com/us/en/Categories/Products/Storage/M-...

I get the feeling we might understand the term interface differently. What do you take it to be? To me, the interface is PCI-Express 4.0 4x. The "1950 MB/s write and 4700 MB/s read" are the specifics of the drive's performance.

According to Wikipedia, PCIE 4.0 4x can sustain more than the spec sheet advertises [0].

The spec sheet advertises 4700 MB/s write and 1950 MB/s read. The interface can handle more than 7000 MB/s in each direction. That's more than twice the advertised bandwidth of this drive.

If what you mean is that the MP600 isn't able to handle the claimed 4700 MB/s "most of the time" or something, than that's fair (I have no idea if that's the case or not). But the "interface" doesn't really come into play, here. The drive is advertised as PCIE 4.0 4x because the 3.0 revision couldn't handle the advertised maximum read specs.

Just like it would seem strange for an internet provider to sell a 400 Mbps (advertised) connection over a 100 Mbps line.

---

[0] https://en.wikipedia.org/wiki/PCI_Express#History_and_revisi...

> There's a minority of consumers who went for the cheap stuff and now want to make money off of it by pushing it beyond the stated limits.

Wasn't the GP's whole point that those limits aren't stated? Or that they are deliberately misleading, e.g. consumer internet sold as "up to 10 Gbps", but would more accurately be labeled "10 Gbps once in a blue moon".

I touched on this point in the comment right above yours: https://news.ycombinator.com/item?id=27005552

TL;DR version: This should be dealt with in the law.

So, all we have to do is defeat regulatory capture, and then we can have meaningful advertised speeds. That also might first require getting money out of politics.

If you’re in the US, fortunately, Larry Lessig has been working on fixing both of these issues so we can have meaningful advertised internet speeds:

- rootstrikers.org

- equalcitizens.us

I wonder how often that is the case (the limits not being stated).

I've looked a bit for an nvme drive for an older computer, so I wasn't looking for the absolute best performance, but rather larger capacity and better endurance for a low price. And all the reviews I have seen had rated endurance numbers.

I've just had a look at WD Blue drives [0], and the spec sheet is easily accessible on their site and gives the endurance in GB written, not some timeframe based on an unknown "normal daily usage".

---

[0] I've chosen WD specifically because of the recent shady PMR / SMR drives.

You should contact your provider. My ISP has always provided at least as much bandwidth as advertised.

Household internet speed are the maximum possible if at 2am nobody is downloading anything in whole building for example. Company pay for guaranteed bandwidth, and premium level troubleshooting from provider. The prices are astronomical.

My employer with 400 onsite employees (in noncovid situation) has provably worse transfer speed & latency than my household with our consumer 1 Gbps connection. I expect them to pay at least 100x more for it. But its a bank, they can't do differently - they never had connection issue in last 10 years I work there. If they ever had, the team of experts would have fixing it as their top priority.

This.

We had a consumer-grade coffee machine at the office once, along with free coffee capsules for the employees.

Whereas in somebody's home that machine would have made 4-5 cups of coffee a day, that machine was making in the order of hundreds a day.

Of course it broke quite often.

But having worked at a restaurant as a waiter in my youth, I can tell: industrial (even entry-level) professional dish-, washer are on another league, completely different. If anything, for having the dishes done in 40 seconds (!)

> ...for having the dishes done in 40 seconds (!)

While my next home dishwasher will be an industrial / commercial-grade dishwasher after one too many services calls for a supposedly premium residential make (Bosch, and I've heard rumblings Miele fares no better), I also have to acknowledge there are trade-offs.

1. I need to supply higher voltage and amperage than most homes plumb for dishwashers. Fortunately not three-phase for the smaller units I'm looking at.

2. I cannot put my more delicate dishware into them, and hand wash them.

3. While there is more energy and water use, I'm okay with that. That marginal inefficiency is more than made up for by keeping a residential dishwasher from entering the landfill in just a few years. Under residential use, I can reasonably expect to run it until parts run out, which for the better brands like Hobart, does not appear to happen for decades.

4. While I have to use detergents that are not available at the grocery store, I'm okay with that.

I sometimes wish there were open source reference blueprints for the big ticket household appliances built around a design dimension that will not change for the foreseeable future: human planiform scale. And manufacturers plug into that frame so the basic core like the tub of a laundry machine or dishwasher doesn't change common mounting points and dimensions while everything around it changes and swaps Ship of Theseus-style over the decades, radically reducing the footprint of these machines entering the waste stream. What I'm really wishing for is more cradle-to-cradle design however, and I should leave the actual implementation to the domain experts.

There's actually not a lot of parts in a dishwasher, I'm curious what gave you trouble with your unit.

Since I have a parade of technicians coming through to fix the issue, I had the opportunity to talk with them. The high-end units of all brands these days have sensors everywhere to save water and energy, and simultaneously monitor cleaning effectiveness as they are all competing on advertising the least utilization with the most cleaning power (monitoring cleaning power and cleaning results is another aspect of saving water and energy). It reminds me of Apple's "you can never be thin enough" focus.

In my case, it has led to an issue that is still stumping the techs, where something is causing the unit to believe it has finished the session well before it even gets under way with the hot water cycle sometimes, and other times to not even start after the initial tub fill. The diagnostic code thrown indicate a malfunctioning sensor they've already replaced twice. Maybe if I call them out enough times, Bosch will just offer to replace the entire unit for me. While I don't get charged labor for subsequent tech visits for the same issue, I still pay for parts, so maybe Bosch's dastardly plan is I'll Ship of Theseus pay for a new unit part-by-part and end up paying 10X what a normal unit would cost before it works again.

In the meantime, it has been fascinating getting an inside look at the design decisions. Lots of incredibly clever engineering has gone into these beasties, which screams to me as a software guy of a ton of refinement over decades of accumulated experience, much of it learned the hard way in the field. Kind of like when I see a software system where the sharp edges have been burnished away into a smooth patina of refinement, and there are still odd bugs that pop up now and then.

Tips for a long lasting Dishwasher:

1: Demand that it's attached to the hot water (machine need less power to heat it up, less chalk...AND Cheaper(your boiler is better at producing hot water))

2: Clean it once a month, especially the pump intake (has often sand like stuff on the ground)

3: DONT use tabs but powder.

However, Industrial grade stuff is more expensive (from the power standpoint), but better for the water (not stupid aromas)

Can you elaborate on tabs vs powder?

Something about pre-wash routines not getting soap to clean some of the grimes.

Technology Connections has got you covered:

> Ever wonder how dishwashers work? Are you ever bummed by the performance of yours? Well, this video can answer your question and possibly provide you with a solution!

* https://www.youtube.com/watch?v=_rBO8neWw04

Oh Technology Connections. What a good channel!

> 1: Demand that it's attached to the hot water (machine need less power to heat it up, less chalk...AND Cheaper(your boiler is better at producing hot water))

Most machines actually specify that the water entering the unit has to be at least 49C. To meet this you usually have to run the hot water on the kitchen sink—which then wastes water. (And may be where people get the mistaken idea for pre-washing: they're running the sink anyway.)

AFAICT, the only brand that can officially be connected to the cold water (10C) supply is Miele, which have an internal heater. And the official specs show that they actually use less overall power with cold water. This is probably because (a) some of the cycle can use cold water, and (b) a smaller quantity overall needs to be heated up compared to running the tap. It does take longer, time-wise (~30m), though; see page 45:

* https://www.miele.com/pmedia/ZGA/TX2070/10544640-000-03_1054...

BSH says cold water is okay, warm water only up to 60 °C (!).

I'm under the impression most dishwashers follow pretty much the same design. There's an inlet valve, usually just the aqua-stop itself, which feeds into the water/ducting plate, which also acts as a heat-exchanger between waste water and water intake. Another two valves control the water softener (to switch between normal operation and regeneration using salt water, which is what the salt you put into it is for). Fresh water runs into the pump sump, which contains the waste water pump, which is controlled by a water level switch connected to the inlet (which also turns the aqua-stop off). The pump sump has a third connection, going into the pressure pump, followed by the water heater, that then goes to the washing arms. The water-clearness sensor (which is just a light barrier) is located in this path.

So overall there are remarkably few actuators and sensors:

    - Three valves
    - Two pumps
    - One heater
    - One solenoid to release the powder / tab latch
    - Outlet temperature sensor at the heater
    - Thermal cutoff integrated in the heater
    - Water clarity sensor
    - Level sensor

I have never had a dishwasher connected to the hot water (which at 60C would be too hot).

I've never heard of a dishwasher that did not have a water heater.

They all worked very well for years. I've had Bauknecht, Siemens and Miele.

AFAIK commercial dishwashers are very diverse: some are rugged equivalent of a consumer dishwasher, some can only sterilize pre-washed dishes, and some do the quick 40 seconds wash... but use really a ton of electricity.

Yes, for smaller cafe-style commercial establishments, I've seen the "ruggedized consumer" type models, but chatting with repair techs, I got the impression from them they feel those units are nearly as corner-cutted-into-not-BIFL as their consumer brethren. I did forget to mention that another tradeoff others might not be aware of is having to scrape off food before it enters these commercial units. I already do that.

My style of cooking is wash as I prep and cook, so I'm constantly in motion while prepping and cooking. I dynamically figure out ahead of time whether what I'm using at that moment is needed the rest of the session, and wash it the moment I get a "let simmer for N minutes" break, and everything that is dirty that doesn't land on the meal table lands in a sink of hot soapy water. I scrape for a compost pile as I go as well.

This minimizes clean up time and food waste at the end of the meal, but I have since found out few people do this. I learned it when I worked back of house. While I had dedicated team members for various clean and prep activities, there was still plenty to do with not a moment wasted, so building that habit was the only way to keep up with the service rush. Having just learned queuing theory gave me a leg up over a couple old salts who learned the same through brute force intuiting the habit, and I taught the less-experienced team members by giving them rote expert-system-like if-then procedure.

It was a glorious feeling when my entire shift team had only a few minutes of cleaning and tidying at the end of the shift well before the cleaning crew arrived (thank goodness we did not have draining and cleaning the fryer on our duty list), and we could relax a few minutes over a snack and shoot the breeze just before clocking out.

Our office has worked out that it's still cheaper to buy replacement consumer-grade applicances than one really expensive industrial-grade and pay for servicing and maintenance. It's really wasteful but made more financial sense.

I don't follow, I think?

The tests that you link to suggest that SSDs meet and exceed their manufacturer's durability claims as measures in petabytes written.

The problem appears to be that manufacturers want to translate that number into something consumers understand, and farmers now want to take them up on that for their industrial use of the devices.

Manufacturers use "lasts for 10 years" as the label for consumers. Which seems entirely reasonable - the disks really will last for 10 years in typical usage, and this phrasing seems useful for consumers.

Now we could fix by either requiring that any warranty standard must hold under any imaginable usage scenario, or manufacturers could resort to a somewhat squishy "reasonable use".

I think the "any use case" requirement is not practical, and we have a long history of "reasonable use" being used, eg for cars.

The problem here IMO is that "reasonable use" is a loophole that can be used to justify anything, or at least it doesn't tell the customer how to use the device before expecting trouble.

If the limitation was also expressed in objective, preferably quantitative terms, it would be clear. Internet conections usually have that: 1Tb transfer at 100Mb/s speed, then 64kb/s. 50% of the time, the speed may be limited by congestion, etc.

I think they reference a "standard daily use"?

But even if that's not strictly quantified, I think the manufacturer position is justifiable, in particular when advertising to consumers.

It's perfectly fine if they want to claim:

> Lasts for 10 years!*

> * Based on a total lifetime of X PB and typical usage of Y GB/month.

But if they claim "Lasts for 10 years.", then their implied warrantee is "Lasts for 10 years.".

As long as "typical usage" fits reality and is not blatant lie like "serving size" on many cookies.

(lets say that there is package of snack that everyone eats at once - package label claims that it is expected to be not eaten at once, what allows them to make less horrifying nutrition label)

Part of the point of how I prescribed it is specifically that you can take X and divide by your own value of Y to get useful expected lifetime under more rapid writes.

That is helpful, but primary visible message still should not be actively misleading.

On further consideration, "perfectly fine" was a poor way of saying something more along the lines of "not fraudulent".

AFAIK most drives have their TBW buried in the specs somewhere

> Inferior endurance is the "dirty secret" of the consumer-level SSD industry. You and I know better, but the average consumer doesn't pay attention to write endurance or under-provisioning.

Besides Sandforce SSDs I've never have had any SSD fail, including cheap SSDs with 300+ drive writes, which is an insane workload for anything consumer (in my case, being used as a scratch disk) -- you are literally writing dozens of terabytes to an SSD that only holds twohundred GB or so.

None of my normal-use SSDs approach any amount of wear like this, and despite some being used daily for almost ten years, they generally have less than 50 drive writes.

So I really don't see why any FUD is warranted here.

I did, it was Intel 330.

It was in a machine running Ubuntu and manifested in such a way, that during kernel update (linux-headers, i.e. writing lot of small files, to be more specific) it just gave up and became unresponsive until power cycle. It still did work during normal, user usage when the demand on it was light.

FWIW I observed similar behaviors, but it was caused by the file system (btrfs), not the underlying storage.

> First they sell products that don't stand up as well to wear

They stand up to the wear of a certain duty cycle. If you have a heavier duty cycle buy a product designed for it (which will probably cost more).

There's a reason why Toyota's Corolla sedan (or even their Supra) costs less than their race car that won the 24 Hours of Le Mans.

* https://en.wikipedia.org/wiki/Toyota_TS050_Hybrid

Bad analogy. In case of Toyota sport cars its the refinement you are paying for. There wont be much endurance difference between LFA vs RC, and especially against 20 year old tuned Supra mk4. Toyota had a bad habit of making bulletproof engines like 2jz or 1UZ-FE.

It's not really a secret. If you get a consumer KitchenAid mixer and use it for baking in an industrial kitchen, KitchenAid is probably going to reject your warranty when you send it in for a burned out motor after 6 months.

Warranties cover defects, not wear and tear far beyond the advertised workload.

You might want to revisit Magnusson Moss act. No such thing as 'advertised workload'.

If advertised duty cycles aren't a thing, then every SSD, printer, and paper shredder is breaking the law.

I'm not aware of anything in magnusson moss prohibiting you from indicating a devices designed workload-- whether it's pages shredded per minute, pages printed a month, or TBW / DWPD. It is well known that flash has a limited lifespan and no court is going to rule that burning through those cell writes with a commercial workload somehow entitles you to coverage under an implied warranty.

>Inferior endurance is the "dirty secret" of the consumer-level SSD industry. You and I know better, but the average consumer doesn't pay attention to write endurance or under-provisioning.

The hope is that capacity increases outrun the loss in endurance by changing the way the SSD is used. If you have a large capacity SSD the assumption is that you often just write a large file onto it once and it stays there for a long time, resulting in a net gain in endurance.

If you had the opposite, like an SSD cache integrated into an HDD where the cache is being overwritten all the time you run into a problem and your best choice would be to use a smaller capacity SLC SSD with more endurance.

Pretty easily solved imo by ending the warranty on the rated TB written (they all have ratings for this and they vary wildly - faster drives can have 1/10th the endurance of others) - in addition to the time limit.

I do have a problem with someone selling me a product and then telling me what I can and can't do with it.

I would think there is a write limit in the warranty.

If that is the case, it's more a question of whether the advertised lifetimes are dishonest marketing or not. I guess most people shopping the warranty would understand "years or write limit" well enough.

I would say they aren't telling you what you can do with it, they are just telling you what uses they will cover under warranty. In this case, it's likely that a legal requirement to cover the drives under warranty for this use would cause the price to go up or the product to be pulled.

It was truly problem in early SSDs, but now normal consumers never exceed rated TBW since 3D NAND is used.

Would Samsung 970 EVO Plus NVMe drive be considered consumer grade or enterprise? Also, are NVMe drives susceptible to faster degradation due to higher throughput?

> Would Samsung 970 EVO Plus NVMe drive be considered consumer grade or enterprise?

This is a consumer SSD, by Samsung's own classification: https://www.samsung.com/semiconductor/minisite/ssd/product/c.... The 970 Pro goes into prosumer territory. The DCT range are their enterprise drives: https://www.samsung.com/semiconductor/minisite/ssd/product/d...

> Also, are NVMe drives susceptible to faster degradation due to higher throughput

Only if you use the time saved to write more data than you would have with a slower drive. e.g. for 1TB both the 970 evo (nvme) and 870 evo (sata) are covered by warranty for 600 TBW.

The 883 DCT on the other hand offers 0.8 DWPD for 5 years or drive writes per day. So 5 * 365 * 0.8 = 1460TBW, or about 2.45x the consumer drive

>Only if you use the time saved to write more data than you would have with a slower drive

Well, NVMe SSDs do get quite hot due to much higher data transfer speeds, and higher temperatures lead to faster flash memory degradation, although I have no idea how important it is in practice (probably on the order of an SSD dying in 10 years instead of 20)

Consumer NVMe SSDs only get to those higher temperatures if you're writing a lot more data than a SATA SSD or hard drive could handle. And such higher temperatures might at a stretch have a meaningful impact on the data retention time for data currently stored in the SSD, but won't meaningfully shorten the program/erase cycle count.

Flash has (much) worse data retention at higher temperatures, but endurance actually increases with temperature. So the optimal way to handle flash memory is mildly hot when written to (to increase endurance), but stored cold (to increase retention).

Watch out for the fastest Samsung and WD drives. You want to find TB written (TBW) rating for the drive, which is in the specs but takes a little digging usually to find.

Consumer models will often have about 300-600 TBW of life. Data Center hardware designed for continuous use will have 1000-3000 and up.

Watch out for the FASTEST drives like Evo and WD Black, faster drives (PCIe 4.0, 7GB/sec read) can have much lower endurance.

Evos are consumer.

Nvme drives degrade at the same rate as measured in data written. It's a protocol difference, not a difference in the flash quality or construction.

It's funny that people here are talking about mining as a morally wrong activity, earning income from wasting resources as if it is somehow unique.

The whole mining waste pales in comparison with what financialization of real estate and other markets has done to our economy.

Reply on news.ycombinator.com