The SSD Review The Worlds Dedicated SSD Education and Review Resource | 
It seems that the world of technology has stopped with allegations that some SSD companies are pulling the old ‘bait and switch’ routine in their SSDs by switching off components that many had recognized through initial SSD reviews. We have read several reports and forums, most of which simply repeat the original information, and finally have decided to clarify things just a bit from our perspective. Get ready though as many may not like our viewpoint; it goes against the grain somewhat.
RECENT EVENTS
Some time ago, Kingston released their V300 SSD and it contained synchronous NAND flash memory, only to be recently updated to asynchronous memory that may be identifiably less powerful than the original release. Similarly, the PNY Optima was originally released with a SMI 2246en controller where a product change has it now being shipped with the LSI SandForce SF-2281 controller, the latter being an 8 channel controller vice 4. In neither case did companies notify SSD buyers of this change. Do we like what has occurred…definitely not. Is it as bad as many have made it out to be ? We don’t think so. What we do find concerning, however, is how so many jumped on the band wagon that these companies were providing better SSDs to reviewers for review, and then pulling the ‘bait and switch’ in retail sales. This is not, and has never been the case.
BAIT AND SWITCH EXPLAINED
Bait and switch is the ‘intentional’ advertising of a product with a specific component or feature that would attract a customer, only to switch this product in the sales channel with a lesser performing and cheaper product, increasing the profit to the manufacturer. This hasn’t occurred in either case as the original memory in use by Kingston in their V300 was due for a refresh (Kingston loves refreshing the SSDNow family) and, as for the PNY Optima, the mid-class SMI controller was switched for the upper-class LSI SandForce controller. If anyone in the industry tries to explain that a 4 channel controller (SMI) is more capable than an 8 channel controller (LSI SandForce), simply walk away and laugh. It is not worth the argument and those straggling few anti-SandForce war horses aren’t worth the listen as their one sided view destroys any knowledge or credibility they might hold.
The PNY Optima issue doesn’t exist. As for Kingston, let’s make it very clear that we didn’t like the switch, but the issue lies in whether they tried to pull a fast one, and more importantly, whether the new components matched their listed specifications for that SSD. Kingston was very frank in stating that maintaining the original memory was not value smart for that particular product. It is a given that asynchronous memory has lower write transfer speeds when moving incompressible data, such as media, but does this alone make the case against Kingston. We don’t think so. Today’s asynchronous NAND flash memory has come a long way and, as long as that memory meets the listed specifications of the product, they are in the clear, even if we don’t like it. What’s more; we don’t get a say as to whether they test with ATTO or AS SSD to determine their listed specifications.
HOW DID WE GET HERE IN THE FIRST PLACE?
You can blame us reviewers, you know. Years ago, when reviewing SSDs became popular, the manufacturers all put up a united front to stop our opening of SSDs in order to identify components. After all, it is not done in so many other areas of the industry, so why would we do this? Imagine buying a…oh I don’t know…any Apple product and learning that many parts within were actually Samsung manufactured. The world would stop, but nevertheless, this is the reality of cost and value. You might see Samsung and Apple fighting like cats and dogs daily, but the reality of Apple products are that Samsung plays a vital role in their production.
SSD reviewers are a very particular bunch and probably the most competitive anywhere. In our efforts to provide information about SSDs that wasn’t evident on every other site, we have poked and prodded to the most intricate details about SSDs and their components. How much did you know about a hard drive? Not much…it just worked. With SSDs, we speak of 4 channel vice 8 channel controllers, synchronous vs asynchronous memory, encryption, SATA vs PCIe, DRAM cache, and even our performance is broken down into consumer vs enterprise testing, throughput, IOPS, and even drive latency. Quite frankly, what the #$%^ would a person buying a value priced consumer SSD ever need to know about IOPS or latency for? Actually…let’s go one step further.
UNDERSTANDING SSD PERFORMANCE
Both the Kingston V300 and the PNY Optima SSDs are value driven products, products that are described as a ‘cost-effective way to revive your computer’. If you bought one and don’t like the components, your only recourse is to rely on specifications. Do the tests that you are conducting match that as advertised? Are you then looking at this from a more advanced test methodology where incompressible data results show the visible difference and, according to all the published reviews, this isn’t correct? Well, first, if you bought a value SSD for more intense media work where you would need the higher performance, take a seat. One needs to consider the listed specifications and not what the reviews report. How many actually understand SSD performance in any case?
If you are buying an SSD that has been marketed as a value solution (which these both have), we are going to believe that you are looking for that upgrade that the SSD has over a hard drive. Would you believe that, regardless of what SSD we put side by side, there are no visible differences in the visible performance of any when completing typical consumer activities. That’s a fact Jack! The visible performance boost we see from SSDs comes entirely from disk access and has nothing to do with the specific components we are looking at in these two specific cases. If I put the Kingston V300 with synchronous or asynchronous side by side, and I put the PNY Optima with SMI or SandForce controllers side by side, even the experts can’t determine which is which in the ole ‘Coke/Pepsi’ taste test, that we need to term as typical user activities for safety sake here. There is absolutely no visible difference. How can that be?
In SSDs, visible performance is the result of disk access time, these times being as fast as .01ms. Disk access time is the time it takes from a request to a return of information. More importantly, it is the time it takes from the systems demand on start up to operating system files to be executed. In the hard drive, it is around 9ms, or about 90 times than the SSD theoretically. This is where we see such a visible upgrade in the SSD compared to the hard drive. Do we think we might see a difference when comparing SSDs that all have the same super-fast disk access? Of course we won’t. To see the differences, we have to conduct activities that stress the SSD, the most obvious being the movement of media files that are highly incompressible. How often does the typical consumer do that, and more importantly, would they even recognize the difference of a few seconds in the transfer?
HAVE WE SEEN THIS BEFORE?
Of course we have. In fact, we expressed some pretty blatant views when Asus switched off the LSI SandForce SSD in their ZenBook for an older and much less performing SSD. As a consumer, the two things that were important in that case were the understanding that, in typical use, one couldn’t tell the difference between either SSD. In more demanding business use, however, this ultra book just took a dump on the reality of enhancing work productivity and it affected many media professionals. As much as we hate to say it though, the key here is still whether the manufacturer falsified their specs in any way; they didn’t then and to our knowledge, they haven’t even today. Correct me if I am wrong please.
IS THERE A SOLUTION?
For all of those stating that you will never buy Kingston or PNY products again, take a seat because we can easily rhyme off dozens of others who have done similar in countless industries. The consumer will always lose out on the manufacturer cost/value consideration because, if a product won’t make money, it won’t be sold. If you were looking for anything but a value SSD, you should have purchased just that. I was amongst the first to tell Kingston that the memory switch to asynchronous didn’t provide for good optics and I can guarantee they will heed that advice. As for PNY switching off the SMI 4 channel controller for the LSI SandForce 8 channel, they should be given credit quite, frankly. In the end, blame the reviewers for giving you a look inside SSDs as you never saw with hard drives, and most likely won’t for most other tech items. It’s our fault after all!
KEEP UP WITH THE ‘LEARNING TO RUN WITH FLASH’ SERIES!
Sometimes the changes are baffling. Intel very recently changed the block size on their SSD DC S3700 from 512B to 4K, which caused my sysadmins no small amount of heartache.
Just punting by pointing to the specs is not acceptable. Most SSD vendor specs are utterly useless. There is no mention of latency or guaranteed write times, things like “95th percentile write latency is XXX microseconds and sustained long-term throughput is XXX Mbps”. In the absence i=of industry standards, if the vendor makes a BOM change, they should also change the product model number and provide release notes. The GM power switch recall fiasco shows the dramatic consequences of what happens if /you don’t. Clients may have tested a given configuration, and changing the product mid=stream is deceptive at best, fraudulent more often than not.
Fazal,
We have a totally different mindset when one considers your scenario, and just about any drive other than the typical value drive we see here. The industry should have mandated an across the board spec standard years ago. Even at the lower levels, there shouldn’t be sales based on ATTO tests. In this scenario, however, we are speaking of consumer value marketed SSDs where the manufacturer didn’t guarantee components whatsoever, just performance based on an industry accepted poor performance scale. These SSDs are both value SSDs yet the argument is that users are not getting performance from them, that was never guaranteed by the manufacturer.
Well, when I buy something I want it to be what I was told it was. If you think a 2281 controlled drive is superior to a SM controlled drive then I couldn’t dis-agree more.
Ummm the SF2281 is far superior to the SMI 2246en, by more than listed specs of throughput, IOPS and build (8 channel vice 4). If you were to buy these items, you woiuld be told by the company that listed specs would meet such and such a speed. Do they? We aren’t stating that we like it; we are bringing forth the reality of what it is.
Yeah, while the controller is faster, it still has tons of issues though. Performance over time still drops and BSODs are still not completly away (can confirm this on 3 different drives)
For a storage device, that holds my data, i kinda want to know what controller i’m getting (as its the most important part of the drive) even if its faster than its suppost to be.
I’ve never seen BSODs caused by an SSD and I’ve tested about 9 different SF drives using an extensive bench. Could you detail your test methodology so that we can try to actually reproduce this? I’ve never heard of such an issue on modern SF or Samsung controlled drives.
Samsung drives are not the issue here. But sandforce sf2xxx drives are. I’m usually getting (but rarely nowdays) bsods while running some IDEs. Some other times they are completly random (eg, while watching youtube or drive even dissapears when i try to boot).
Granted, they are much less frequent, now that i’m on 5.x.x firmware. But still they are there and i’m not surprised, since this is really an issue of hardware (power logic) rather than software. They really just patched things, so they are much less frequent.
Although one thing is is possible though; they could have fixed this completly (in hardware) when the new revision was release (which dropped the power consumption quite a bit).
Correct me if I am wrong but after reading this article three times I still have a problem with the underlying logic which appears to be: non-disclosed changes in the specs are ok as long as there is any kind of metric that can possibly forced to justify the means. Asynchronous flash is not the same thing as ONFI synchronous flash or toggle mode flash and snatching up a few SF2281 controllers from a fire sale doesn’t cover up the fact that this controller is 5 years old now, regardless of the number of channels that are available.
If you want to sell something, have a data sheet and then please stick to it, anything else should be considered at least borderline fraudulent.
I don’t see any non-disclosed logic in the specs. The ONLY consideration is the fact as to whether both past and persent specs are met. There is no necessity for any company to tell you anything about the internals and this has only come into play because of reviews.
We don’t like that companies set specs through ATTO, but it is an industry standard. Also,weI didn’t see a data sheet that was not followed, but even in the case of the data sheet, they change as well due to components not being obtainable at a healthy price any longer.
The fact that Kingston would not address the issue directly when confronted with the early “firmware issues” says everything to me about their intent. The NAND switch information was pulled out of them whilst they kicked and screamed that there was no issue. As usual, the cover up/obfuscation is worse than the original problem. To this day their tech support is pretty much telling customers that it is our fault which I think is really bad form.
What firmware issue and why would they have to address it directly. The key here that we seem to be missing is that this is a value SSD with no guaranteed performance, other than the basics as listed, and certainly no guarantee as to components. There is no obligation whatsoever to address anything, so long as the original specs are maintained in accordance with how they put them forward. None of us like this. It’s business though and isn’t any different with SSDs than any other retail item in the world.
“What firmware issue” Exactly. There were numerous theories about it possibly being a firmware issue but they took a VERY long time to disclose that it was not the issue people were inquiring about. They only pushed the “value” tagline after this information became more widely disclosed to cover themselves. I bought my first V300 based partially on published reviews. They were not pushing the value message when I bought my first one and I expected my second one to also be a V300 and not a V300″A”.
Regardless of whether they changed the firmware or components, why is it necessary to disclose it if the specs published by the company haven’t changed? The marketing of this drive being a value drive has been on the wall for some time, and even if it was marked as a premium and then changed to value because of the change of components….what is not correct in what they did?
A better question might be “What is not proper in what they did?” and my answer is that when you make such a substantial change in the product also change the product name. The 240 certainly has become more of a value play since it has been dropping at about $10 a month since this information became more widely available.
So my take on this is that if I am purchasing a “value drive” I should not worry my pretty little head about what any of the internal components are. I’m looking forward to the review of their new FURY value drives since it ought to be a very quick read. Just publish the current selling price, a couple of stock pictures along with some ATT0 data and it is done.
Fury drives are utterly pointless, now that cheaper and faster drives like mx100 are available.
Kingston really dropped the ball, when it comes to SSDs. Unless they bring something new to the table, they might aswell exit the consumer ssd business entirely.
Seriously? Do you employ thousands of employees worldwide? Do you take risks in buying flash and playing in a market with razor thin margins? Comments like that just show how much of a crucial hoser you are.
Nope, this is the reality. Fury is both more expensive and slower than competitive drives. So unless kingston significally drops the prices, its really no point in buying them. They really have no edge here; mx100 for example has topal encryption, powerloss capactiors, reliable controller while kingston has nothing.
Its a fact, that if you want to be competitive you either need NAND fab or inhouse controller, neither of those kingston poseses.
If a car company made a car that could go 160 mph and
advertised it to do 160 mph.
Made it have good handling, great suspension, great tire
grip. Then sent it to be reviewed by a car magazine. Which gave it great reviews.then decided it cost too much money to keep producing it after a year, took away the good handling, great suspension, great tire grip. But said we did nothing wrong it still goes 160 mph, take a look at our data sheet if you like.
Most people would not drive that car around any curves!!! Would you?
In the end the people that buy the SSD will decide what is relevant. There are several of these reduced performance SSD’s on the market and they all sell for such reduced prices that I have a hard time believing anyone is making a profit. So maybe it was a technically OK move but as a business practice it was, is and will be a loser.
Value product doesn’t mean that the specs or the performance can change at will. You can express this opinion as many times as you like but I think we both know that this is plain, well, sorry, plain BS.
Maybe the industry standard needs to change.
So as long as the atto benchmarks are the same its all good is that right?
Well then can Mushkin sell their Mushkin Enhanced Chronos as a
Mushkin Enhanced Chronos Deluxe I think they have the same atto speed?
That would be bad business.
Agree totally….why can everything related to standards be agreed on but not a standard test mechanism???
Great point thank you.
If you want a locked BOM, stick with enterprise drives and get on the manufacturer’s ECN list. Hard drive manufacturers make running changes on consumer-grade devices all the time.
This is an odd article.
The concerns aren’t that the product doesn’t meet specs, it’s that the product doesn’t meet reviews. Specs are great, but most SSD buyers are going by benchmarks, not specs. They’re going to a site, such as your own, reading reviews, and buying the fasterst (which isn’t always the best specced.)
With this product, people are NOT getting the reviewed product. They won’t be able to hit the benchmarks they’re seeing in reviews. That is absolutely a problem. It’s so odd to me that you’re defending it. In that case, what’s to stop Kingston from putting the top of the line product in a mid-tier package, sending it to reviewers, then claiming it didn’t meet their margin goals so they changed it to a lesser performing one? The reviews would all claim the product was an immense value, and the product would sell, but it would be a lie.
Let’s be very clear in stating that we don’t like it, don’t support it and addressed this some time ago first hand. What is being written, however, is the result of anger and frustration where we see someone cry wolf with words like ‘fraud’ and accusations that companies are actually putting some special sauce in reviewers drives. It gets to be out of hand and all angles need to be explained, whether we like them or not. In this case, product was changed but still matched the specifications of the manufacturer. Like it or not, that’s a fact. This is simple business where components either become impossible to get or harder to get, forcing companies to change what they have installed. So long as these do not go against the original specs, there is no reason to voice a change. This is business and it occurs at every level. If you buy a car today, the chances of it not having different components next year, for the same purchase, are one in a million. As long as the posted specs don’t change, they are in the clear. The truth remains; we don’t like it but there is nothing that can be done. Although grey, the company was fully in their right to do that. Personally, I just hate how everyone grabs a hold of something that will gain attention, posts and reprints without voicing all angles We attempted to display the other angle.
pcper.com and gamersnexus.net are both playing that same different angle today.
I wrote the GamersNexus article and was alerted to this one within a few minutes of publication. Looks like Les and I were writing at the same time.
The accusations of ‘damage control’ from the ‘net are getting tiring, but it is comforting to know that Les & Hellstrom are of the same opinion.
Just read this this morning and yes, comforting to know that all aren’t simply blinded by the initial shock. Thanks for this post and feel free to include your link.
Why buy Kingston SSDNow V300 with asynchronous flash? When you can get a
crucial mx100 or m500 synchronous
pny pro synchronous
muskin deluxe synchronous for the same or less
if these companies can make value ssds with synchronous
nand so can Kingston.
Stop the madness already!
I am the owner of the original v300 and its great but if I bought one when they did the nand switch I would be extremely upset!!! Consumers try to spend their hard earned money on the best they can afford your site and others are helping us do that. the specs , reviews , tests all help us to buy the best we can afford . keep up the good work . and tell these businesses to do business the old fashion way with some integrity.
Les I don’t mean to put you on the spot, but it would be helpful
if you could review the new kingston v300 with asynchronous flash, and compare it to the first review of the v300 with toggle flash, benchmark, specs, recommendation.
when companies change something as important as flash or controller the ssd should be reviewed again to make the review current and accurate. It seems no one wants to do that.
Looking into it.
So if AMD tomorrow replaces the Hawaii core with a Tahiti core and starts selling it as 290 for example, it is going to be OK? No one will like it, but it will be OK.
If I am not mistaken because SSDs are not exactly my specialty, this is the same like taking an Agility ans selling it as a Vertex. But it will be OK. Right?
And also, am I reading correctly? The same point of view in three sites almost at the same time? Wow. This screams “damage control”, only doing more damage.
Companies are starting to look like politicians, and hardware sites are starting to look more like political newspapers trying to make blakc look like gray or even white.
” So if AMD tomorrow replaces the Hawaii core with a Tahiti core and starts selling it as 290 for example, it is going to be OK? ”
Good catch !
I don’t think so really. The Core is advertised. Now if a component of that changed, but didn’t affect the listed specifications, that would be your match.
Well you are wrong because not all customers know about cores and Hawaii core and Tahiti core and stuff. But anyway.
Specifications. You are trying really hard to defend Kingston here.
I could bring other examples, but forget it. I will be losing my time.
What your main argument here is, is that SSD manufacturers don’t give enough info about the specification of their products. So it is logical and ethical and fair and justified and whatever, specifications that change performance heavily, but are not specified on the box or on the official page of the product, to change at will.
What makes an argument like this valid…. it’s on 4.
JMO
Well, perhaps from your side then, we should ensure that consumer SSDs are marketed as validated enterprise products where they have to stick to the BOM, this meaning that the price goes up significantly for all. I am not defending Kingston or PNY, but rather, explaining the other side. Was this wrong? I think so in consideration with the level of transparency we have in the industry today, but was their anything fraudulent? No. I might suggest that we agree to disagree but there really isn’t any disagreement here, only the picking apart of specifics.
You want to change the materials for X, Y reasons? Does this affect the product and how it performs? You want to be fair to your customers? Change the model number because after swapping the materials the new product is NOT the old product. Products that are 99% the same and they are only different in 1% could have huge difference in price and/or performance. You don’t sell them under the same model, in the same box.
It is so simple that I will agree with you, that we agree to disagree in this one.
BTW, my side is the customers side. What is yours?
Our side is as it should be, to ensure all angles of a subject are covered. There was a strong lack of leverage in this matter that we felt needed to be stated. You, on the other hand, approach this from an angered view and similar to that of a horse wearing tacks so he can only see straight ahead. Thanks for the input.
Reading the comments here indicates that just too many horses with tacks have a different opinion than you who, of course, can see the whole picture.
Thanks for the nice article Les.
Meanwhile, Kingston has also changed HyperX 3K’s NANDs too. But this time for a good reason maybe:
https://www.ssdlabs.net/incelemeler/kingston-hyperx-3k-120gb-incelemesi-biraz-daha-sandforce/
Smackdown from Tweaktown:
“As an SSD reviewer, it’s my job to sort that out and recommend products based on the needs of our readers. Instead of giving every product that has an Amazon link an award in order to profit ourselves, we actually investigate the products and fulfill our commitment to the public.”
i could tell the difference between coke and pepsi. coke tastes a bit like the water that ws used is full of dust (all the coke band drinks do) my wife was able to taste what i taste when she was pregnant and could only drink pepsi.. now shes not pregnant she can drink coke again. but i just cant drink the stuff. even orderd pepsi at a resteraunt and they gave me coke and i knew they had from the taste..
so your paradighm is flawed 🙂
Great post. I realized access time was most important back when we used actual ram setups pre-ssd era. I have always found it odd reviews and specs list read/write speeds but almost never list access times. Takes more research to find the access times which is the most important.
In all fairness, 99% of SSDs have the same access time which is why emphasis isn’t placed on it. There needs to be a way to differentiate and, unfortunately, the common method is by high sequential read and write speeds, that of which is rarely ever used by the consumer anyway. Thanks for the comment!
The problem is that this article downplays this behavior by manufacturers as “ok” by providing justification for it. “They have to do it to make a buck” is practically as good as anything written here. So why not just come out and say that?
I hate to say it, but its almost like the author does not want to get these companies riled or something. The sad fact is replacing components that have a significant affect on performance without rebranding them is farcical and IMHO wrong. And this article is “poopooing” the behavior, but also saying… “meh its ok” THAT is BAD for both the author and the industry.
Switching components is very deceptive when it does have an adverse effect on the end user. The author should really be hammering these guys instead of providing a muttering excuse.
I am the author and see you view completely, although I don’t necessarily agree. It is not ok and I have had my own opinion of this practice all along. It is not by any means new. I explored this from the perspective of the completely logical viewpoint, however, and as a consumer SSD, the are no bill of materials that tie any consumer SSD to the parts within. The only reason you know of those parts is because of the reviewers. The only requirement to the manufacturer is tyo meet posted specs, whether using compressible or compressible data samples to test. That is the only requirement. Morally, do I think it is wrong to change the components of an SSD after the world starts purchasing as a result of reviews. Absolutely!….that is if you hadn’t made some type of mention of this occurrence.