Under the Microscope by Alex Serpo

A calculated look at the latest trends within technology products.

Five reasons SSDs are great

Posted by Alex Serpo @ 15:55 27 comments

It's rare that we see a genuinely new technological paradigm land on retail shelves. Solid state drives, or SSDs, are one such technology, and have only just begun to limp onto the market at strictly enthusiast prices.

With close examination, it seems that the days of the good old mechanical hard drive are numbered, with SSDs ready to eclipse them in the long run. Here are five reasons to look forward to the ascendancy of SSDs.

1. Speed
First, some background. Unlike mechanical hard drives, SSDs don't contain any moving parts. Rather, they rely on millions of tiny flash memory cells. These tiny flash cells are non-volatile memory, meaning that they retain their data even when no power is present. Today SSDs most commonly utilise NAND-based flash memory.

This means no moving parts, and thus no spin-up speed, which results in shorter boot times. It also means faster read speeds from idle, and faster write speeds for larger files (unfortunately, due to the nature of their design, SSDs struggle with writing smaller block sizes, see here for more details).

Realistically, this results in hard drive access which is about 60 per cent faster — the score we achieved for a file transfer when comparing Intel's latest X-25M SSD with a 7,200rpm Seagate drive.

Or, for a more practical demonstration, here is a video featuring the MacBook Air.




2. Power saving
There is still a lot of debate around this, but it is safe to say that generally, SSDs have a lower power consumption than conventional hard drives (if only marginally so). This varies depending on the test, the SSD, and the mechanical hard drive you are comparing the SSD to.

However, what we can say is that in the long run, flash-based SSDs have one up on mechanical hard drives in the power consumption stakes. Comprehensive tests performed by Tom's Hardware show that while average power consumption is comparable between SSDs and mechanical hard drives, when optimised, SSDs can achieve a much lower average power requirement and performance per Watt figures.

3. Size
This one is a bit of a no-brainer, SSDs can be much smaller than mechanical hard drives. While they are generally packed in similar form factors to mechanical HDDs (1.8, 2.5 and 3.5 inches), they have the potential to be much smaller. MicroSD cards are a good demonstration of this, as is the seemingly endless increase in the capacity of USB keys. The ability to put a huge amount of storage into a tiny space creates new opportunities in gadget design — iPod Pico anyone?

4. Durability
It is said that the enemies of reliability are heat and moving parts. SSDs have no moving parts, meaning that they aren't susceptible to the vibration or drop damage associated with mechanical disk drives.

The nature of flash memory also means that it can survive through a larger temperate range, with some flash drives operating at up to 70 degrees Celsius.

It is safe to say that generally, SSDs have a lower power consumption that conventional hard drives

Write cycles vary markedly depending on the grade of SSD you are looking at. There are two kinds of flash memories available in SSDs: single level cells (rated at around 100,000 cycles), and multi-level cells (rated at round 10,000 cycles). This whitepaper provides more details.

Technologies such as wear levelling can increase durability, as can other improvements. Improvements can allow as many as 1 million write cycles, a value which allows for continuous use over decades. Given that mechanical hard drives are normally rated around 50,000 read/write cycles, SSDs currently have competitive durability and the potential to surpass mechanical hard drives.

5. The best is yet to come
This is the most important point of the five. SSDs are a fledgling technology, and currently aren't living up to their potential. Mechanical hard drive technology, on the other hand, is mature and so is seeing its limits pushed.

The first thing we can expect from SSDs is a dramatic fall in price. Secondly, we can expect steady increases in both storage capacity and reliability.

Finally, current software is optimised for use with mechanical disks, particularly more recent versions of Windows. As SSDs perform better under certain conditions (such as writing larger memory blocks), so software can be optimised to work with SSDs. This means software optimisation can parallel hardware optimisation, providing further headroom for improvement.

Advertisement

Talkback 27 comments

    Potential to be directly addressed by the CPU as non-volatile RAM Patanjali -- 23/01/09

    The ioDrive by Fusion-io shows what can happen when flash bypasses SATA and is put on the PCIe, providing 700MB/s read and 600MB/s write.

    But Flash has the potential to replace DRAM, leading to the scanario where installing programs effectivelly makes them in RAM and ready for immediate running.

    Computer s could have TBs of direct CPU storage. Computer boxes would be completely redesigned as thare would be no HDDs requiring their own cages for vibration isolation and heat dissipation.

    We are on the cusp of a real computing revolution.

    All current SSD problems are on the write side. Patanjali -- 23/01/09

    The large block size means small writes mean reading a whole block, modifying and writing out the whole block. Most seem to become better when the transfer size is 64KB and above.

    The life of an SSD is limited by the number of writes that one memory location can endure.

    Both of these limitations DO NOT apply to reads, so any situations where there is write rarely and read mostly, such as for music instrument sample drives for audio recording, will allow SSDs to be used NOW with no performance penalties, fast access, fully random access (without penalty) and infinaite lifetimes.

    It will get better Mel Sommersberg -- 23/01/09

    Firstly, an observation about the video clip - I haven't owned a PC in many years that has booted as slowly as the three Apples in the video - it says alot for PCs and the Windows operating system.

    Moving to flash drives, much has been said recently about laptops needing to be recalled because flash drives have earned a reputation for failing in large numbers. I am sure that this will pass in years to come. Hard discs were by no means any more reliable when they were first invented but time has sorted out the early problems associated with them and it is not uncommon for a hard drive to last for 15 years and more. How to they make them so cheap?

    One day flash drives will become faster and reliable than they are at the moment and they will take over from discs. It is inevitable.

    Originals had stepper motor head positioning Patanjali -- 23/01/09 (in reply to #320121525)

    that were very sensitive to the orientation of the HDD. Vertically mount the HDD and the wrong data was consistently read.

    Servo head positioning allowed more reliable and faster access times.

    Low price and better random write time will spell the HDD death knell Patanjali -- 23/01/09 (in reply to #320121525)

    It will depend whether the price will drop for SSDs enough to generally replace HDDs before other topologies (like direct attached flash to CPUS) make the whole 'drive' concept obsolete.

    At any rate, 2009 WILL be the year that SSDs will get cheap enough and in large enough capacities to replace HDDs in a large number of computers. I do not think that it will be years.

    Not the OS Anonymous -- 04/02/09 (in reply to #320121525)

    They're Macbook Airs in the video, which have 4500 RPM HDDs---one of the tradeoffs in packaging things that small. First gen MacBook airs even used PATA (2nd gen ones use SATA) MacBooks & Mac Minis use 5400rpm (SATA), and the other desktop machinese use 7200rpm SATA. Bootup speed is mostly down to hard drive speed; my MacBook Air (HDD model) is slow at booting (but then I only boot once a day, and it's faster at the small writes than SSD), but my desktop G5 (7200 SATA) is noticeably faster at booting than my Windows box at work. All the video shows is that that MacBook Air HDD models have slow HDDs.

    correction Anonymous -- 05/02/09 (in reply to #320122287)

    My bad, 4200rpm... as stated in the video even.

    Forget about a lot of HDD maintenance Patanjali -- 23/01/09

    SSDs do not require:

    - Defragmentation, because there is NO difference in the access time whether the data is next to each other or at opposite ends of the drive. If fact defragmentation decreases the life of the SSD because of the number of writes it involves.

    - Separate drives for programs and data where speed is required, because the accesses are quick enough that the advantage of having the separate, independent heads is minimised.

    Note that Some SSDs are beginning to approach the lmitations of the SATA II interface speeds (3Gb/s).
    RAID0 will then be the only way of improving the transfer rate. This often lowers the cost, because at the top sizes, one SSD is more expensive than two of half the capacity.

    SSDs suck balls Anonymous -- 26/01/09

    Firstly, most of them are like 64 gigs at the most compared to 1.5 TB HDDs and they got a limited number of writes to each bit before eventually becoming useless. A million writes may seem practically unlimited, but filetables and logs easily bring down the lifetime to less than a year even with wear-levelling.

    No educated user who does anything more than use e-mail and MS Word would rely on faulty **** like SSDs. It will be many years if not decades until SSDs become as reliable as HDDs.

    'Till then, for all practical intents and purposes, it's garbage.

    are you kidding? Anonymous -- 28/01/09 (in reply to #320121653)

    "...most of them are like 64 gigs at the most compared to 1.5 TB..."

    Yes compare the lowest sized SSD, to the top of the SATA size at the moment, that makes sense.

    What were HDD capacities in their early years? 1 MB? My first PC had a whopping 30MB of HDD space, and that was compressed..

    "...and they('ve) got a limited number of writes to each bit before eventually becoming useless..."

    As do conventional HDD's, read the article.

    As the article notes, this is fledgeling technology. Platter based HDD's have been around for decades. In the first year or so, these drives provide more potential than any storage media in recent years. Why simply dismiss this? Are you have a little bit of an emo day?

    "No educated user who does anything more than use e-mail and MS Word would rely on faulty **** like SSDs."
    Firstly, its not faulty. There are limitations at the moment for write cycles, as with current technology.
    Secondly it is fantastic for reading at incredible speeds, you cannot see good use for that except for using office?

    Maybe the process should start with re-designing current OS methods for writing to drives, File system configuration and removing all the services on platforms that would destroy SSD's quicker than we would like. Right now, OS software is not designed to be used with SSDs. Things like caching, search services etc are redundant with SSD. How many other useless services can be removed which will save CPU time and memory? The development of support for SSD will help with this also.

    You dismiss the entire line of product for decades due to the average write cycle being 10,000 writes comopared to 50,000 with platters. (if the info above is correct)
    I dont see how you even thought through that post before hitting submit

    Dead serious. Mack -- 29/01/09 (in reply to #320121747)

    Lowest sized SSD? The highest are about 1 TB, but are experimental and likely are way more unstable than commercial ones.

    "As do conventional HDD's, read the article."

    No they dont.

    "As the article notes, this is fledg(e)ling technology. Platter based HDD's have been around for decades. In the first year or so, these drives provide more potential than any storage media in recent years. Why simply dismiss this? Are you have(ing) a little bit of an emo day?"

    Because they're unreliable at the moment, hence garbage.

    Re-designing an OS to cater to a faulty, unreliable technology wouldn't be advisable, as it would undermine the professional usability we got today. For example, my workplace requires any document moved to another media to be purged from the departure disk with the Gutmann erase (35 overwrites with variable bit patterns.) If we used SSDs, they would become useless quicker and we would be wasting good dollars on expensive equipment. Or, how would I run a server, with a forum board with millions of people posting content every day utilizing an SSD? It would die in less than a month and my accumulated database would be ****

    "You dismiss the entire line of product for decades due to the average write cycle being 10,000 writes comopared to 50,000 with platters. (if the info above is correct)"

    Pffft, wrong. Typical SSDs have about 500,000 write cycles and HDDs have no limit (I got an HD that I've owned since 2000 and there are no bad sectors as of yet. I don't use it anymore, I gave the computer away to my niece and she isn't reporting any problems. Modern HDDs got a lifespan of about 25 years, so I suspect that my old HD has another decade to go, but it's unlikely that she will continue using it by then.

    And no, I'm not dismissing the technology itself, I'm dismissing the retarded commercialization of something that doesn't work. Especially for an expensive price. Until SSDs are improved to the point where the write cycles are limited to something practically infinite like in the hundreds of millions range or where a solution is found to completely eliminate the problem, it remains useless and should be sustained as an experimental technology, and not commercialized.

    In other words: get **** real, wait a minimum of 5 years before spreading some bullshit hype about an unreliable technology.

    What IS your problem? Patanjali -- 02/02/09 (in reply to #320121842)

    Hissy fit time over what?

    HDDs do not exist in the black and white space you appear to place them. They are subject to a lot of limitations, such needing the Gutmann erase to ensure the magnetic material that is normally outside of the path of the heads, but may be subject to adjacent magnetisation, is completely devoid of previous data residues. This is NOT required for SSDs AT ALL.

    The design of OSs has catered for the limitations of HDDs, including the scheduling of defragmentation. If you want an optimised system for efficient handling of streaming, you have to optimise data placement across the available HDDs and their partitions. Because of the variable access times of HDDs, partial stroking (using small first partitions of a HDD) is often employed to limit the maximum (and hence average) access times. NOT an SSD problem.

    Want a separate partition for ease of data management, but have only one drive? If you have a HDD, you will then force much more head travel across the drive. With SSDs there is no performance penalty.

    So if OSs have been designed for HDD idiosynchracies, why can't they be designed to include handling those of SSDs? Quid pro quo!

    As for write-intensive areas limiting the life, there is nothing to stop the wear-levelling routines from shifting the area used transparently around the SSD as each becomes close to its life's end, JUST LIKE is done with HDDs, though SSDs are more predictable. For most users, the heavily written areas will be those controlled by the OS for caching and logs. Have enough RAM to turn off the virtual RAM and disabling unneeded logs will minimise effects upon SSD lifetime.

    Fusion-io's ioDrive outperforms HDD arrays in IO intensive applications and occupies one PCIe slot. In fact, it well exceeds SATA II's limits. It shows that large Flash storage directly attached to the CPU is not too far away. Do you think that will not completely change the way programs are designed and deployed?

    And all the SSD limitations are on the write side - if you have data that is read only (or has non-critical writes), you can utilise the QUIET, COOL SSDs now, WITHOUT lifetime penalty. Fast, quiet and cool can be mitigating factors against cost for some applications, such as sample libraries for digital audio recording.

    This 'garbage' of which you write IS already making inroads and is just about to 'jump the chasm' (look it up).

    Ppl like you misleading noobs causing them to lose their data later, thats what. Mack -- 06/02/09 (in reply to #320122103)

    "This is NOT required for SSDs AT ALL."

    You mean one overwrite is sufficient to purge? Is that done on a bit-to-bit basis or cell-by-cell?

    "The design of OSs has catered for the limitations of HDDs, including the scheduling of defragmentation."

    Defragging was never a problem. Modern comps complete it in less than 5 minutes and the loss of performance is not apparent until a couple months. Unlimited write cycles are always more desirable than having to waste 5 minutes a month defragging. I don't doubt its suitable for streaming, but its still not suitable for general purpose computing. Choose one.

    "So if OSs have been designed for HDD idiosynchracies, why can't they be designed to include handling those of SSDs? Quid pro quo!"

    Because HDD issues dont compare to the reliability problems of SSDs. Also, me and most users prefer XP over bloated, bugged garbage like Vista and 7. Ain't it weird this direction were going: no real advancement, everything being replaced by new, state-of-the-fart technology?

    If I have to stick to HDDs to avoid trash like Vista, so be it. Actually I'd probably switch to Linux by then.

    "As for write-intensive areas limiting the life, there is nothing to stop the wear-levelling routines from shifting the area used transparently around the SSD as each becomes close to its life's end, JUST LIKE is done with HDDs, though SSDs are more predictable. For most users, the heavily written areas will be those controlled by the OS for caching and logs. Have enough RAM to turn off the virtual RAM and disabling unneeded logs will minimise effects upon SSD lifetime."

    Ok, we'll see.

    "Fusion-io's ioDrive outperforms HDD arrays in IO intensive applications and occupies one PCIe slot. In fact, it well exceeds SATA II's limits. It shows that large Flash storage directly attached to the CPU is not too far away. Do you think that will not completely change the way programs are designed and deployed?"

    Maybe, but that's a decade off before it becomes practical. Not in the realm of fantasy, but not reality either.

    And one more thing: according to wiki, SSDs got slower write speeds due to the large erase cells and take more power than an HDD, even when it's off it consumes energy. How sweet is that?

    I think I'll wait 5-10 years before I start giggling like a schoolgirl over the prospect of SSDs and the terabytes of porn it'll bring me.

    FUD - spreader Patanjali -- 10/02/09 (in reply to #320122466)

    "Ppl like you misleading noobs"

    Noobs = beginner. I have been in IT/electronics for almost three decades and was a Tech Manager at a computer manufacturer, personally building hundreds of computers. Of course, that does not automatically make me an expert in this instance, but I do not accept being likened to a fool by one.

    "causing them to lose their data later"

    How does recommending SSDs lose their data? SSDs have a predictable failure profile, unlike with HDDs that are using probability. SSDs diagnostics are a superset of SMART. When a HDD fails, that is it - you MAY have been told beforehand.
    Of course, what you are doing is spreading FUD based upon a limitation that will probably never be one for the vast majority of users. Saying they will absolutely lose their data is a lie!
    I could say that every HDD user WILL lose their data - but that is only true if they never replace it with a newer one first. Most SSD users will PROBABLY replace their SSDs for the same reason before it dies. Most people do not write that much and completely fill their HDDs, both of which have to happen a lot before the SSD will be write exhausted.

    "Unlimited write cycles"

    I have had a few HDDs that didn't provide this!

    "You mean one overwrite is sufficient to purge? Is that done on a bit-to-bit basis or cell-by-cell?"

    One write is enough. One block is done at once. It is not a fuzzy write as in HDDs.

    "If I have to stick to HDDs to avoid trash like Vista"

    SSDs are not limited to trying to make up for so-called Vista limitations. ANY OS can benefit from SSDs as they are fast, quiet, cooler in use and will most likely not present problems for most users at all. While XP by default is not optimised for SSDs, it only requires aligned partitions to mitigate the write stuttering.
    BTW, Vista works better in many ways than XP. For digital audio recording, which requires very stable systems, these days has no core issues with Vista (other than hardware manufacturers not making new drivers). XP took a very long time to be stable enough for this purpose, identifying which services needed to be disabled to prevent streaming interruptions.

    "take more power than an HDD, even when it's off it consumes energy".

    Que? When its off, its off - no power used!
    Powered up, but without write cables, SSDs DO use slightly more power than HDDs, possibly limiting their use for some laptops users, but start accessing them and they are a lot less power hungry than HDDs. I have four 10k Raptors and they get very warm to touch.
    Wiki isn't the fountain of all truth.

    "wait 5-10 years"

    Obviously a 'laggard' (late adopter) in the technology realm.

    Try debunker Mack -- 11/02/09 (in reply to #320122852)

    "Noobs = beginner. [snip] I do not accept being likened to a fool by one."

    Where in my text of accusing you of misleading newbs implies that I called you a newb??????

    HDD failure is predictable. Clicking is a common symptom that something is wrong and that you should back all your **** up ASAP. Bad sectors are another.

    "Of course, what you are doing is spreading FUD based upon a limitation that will probably never be one for the vast majority of users."

    What about active users? Ordinary users are more likely to have problems with lack of HD space than experienced, since they download a lot of videos and dont know how to use an eraser tool to clean up all their temporary **** I got a 250 GB drive, I use up 250 gigs of my allowed internet bandwidth per month and still manage to have 50% free space. But I move and delete **** FREQUENTLY so my HD is always hot.

    "Saying they will absolutely lose their data is a lie!"

    They will eventually, on busy areas. Wear leveling eliminates that but further complicates interoperability and drive will eventually fail in a couple years. If its predictable as you say, that's cool. One less problem down.

    "I have had a few HDDs that didn't provide this!"

    If those HDDs were not made in this millennium, I wouldn't be surprised.

    "SSDs DO use slightly more power than HDDs"

    And they write slower.

    "Obviously a 'laggard' (late adopter) in the technology realm."

    How is this a negative trait? Experimental technologies, by definition, should never be used for anything serious until they're fully finalized. This was true for color TVs and LCD screens. The first color TVs were extremely unstable, the first LCD screen (actually, all of them until a couple years ago) sucked **** and its quality was comparable to a welfare television from a thrift shop. Low contrast ratios, angle distortion, blurring, only one crisp resolution. I have the same stance towards LCDs as I do for SSDs. I'm using one now as we speak due to my 10-year-old CRT losing its focus and becoming worse quality than a conventional LCD. I've gotten used to it so it ain't too bad but every now and then I gotta turn up the brightness to eliminate chroma distortion while watching movies. Luckily, OLEDs are out and won't have these problems, but due to me being a "lagger" I'm again waiting a couple years for it to develop. You never know what initial problems they could have (like the short lifespan.)
    Or take MP4 audio. In 2001 I converted my entire MP3 collection as well as my custom-made music to MP4, thinking I was leet for being ahead of most and always using the latest tech. Then I find out two years later that MPEG released a new extension which includes two features: SBR and pseudostereo, as well as more modern encoders that improve VBR coding. Now I'm stuck with songs that could've either had higher quality or took up less space for the same high quality if I would've waited a couple years for it to mature. I learned now to just chill, keep myself up to date with articles on the tech and keep track how it progresses and decide whether I should adopt it.

    In short: early adoption is for retards who love to deal with compatibility issues.

    Once bitten, twice shy - I can empathise with that Patanjali -- 11/02/09 (in reply to #320122897)

    Sounds like you tried to be ahead of the curve with MP4s but got caught out. Don't let it make you bitter about change though.

    See www.hightechstrategies.com/profiles.html for the definition of 'laggard' in relation to technology adoption. Your previous comments seemed to imply you being generally in this area. However, you appear to have been bitten by your own enthusiasm. Most who went for HD DVD are probably in the same space now.

    Just to clarify, I DO NOT own any SSDs, but am seeing that for my purposes, their silence and read speed are great, BUT the rate of improvement is making the jump less than straighforward. Like most about computers, it requires research just before the funds are actually available to make the final decision.

    SSDs have not been our highest priority so far . We have had to focus on core functionality, and so incremental improvement are second place.

    My main reason for going for SSDs is for digital audio recording where the silence and lower operational heat will make mixing much easier. It has two distinct drive usage profiles:

    - SAMPLES (libraries of individual notes) are usually written once (at installation or update) and only READ thereafter. Any write problems are irrelevant here. These are usually on separate HDD(s) to ensure they are available regardless of what else is going on. Most cheap, large, high-speed SSDs will handle this easily.

    - PROJECTS are a mixture of multiple streams of reads with a couple of streams of simultaneous writes when recording. Write issues can severely disrupt their smooth operation - write stuttering is obviously a deal breaker here. These are usually on a different drive than the OS and apps for the same reason as samples..

    SSDs are obviously about to go from the 'Early Adopters' to the 'Early Majority' = 'Jumping the Chasm'. It does not look like SSDs will fall into the 'Valley of Death" as they are gathering critical mass.

    That does not mean there are no issues. Early HDDs had several operational restrictions, but that did not stop achieving mass penetration because compared to floppies, they were much better. My first was a 5.24" $600 10MB, with stepper head actuators that became misaligned when the case the drive was in was mounted vertically.

    Now SSDs are not the same quantum leap as FDDs to HDDs, but their essential characteristics make them a significant improvement over HDDs. That single SSDs are now being compared to RAID0 Velociraptors means that they have essentially won the speed race. They easily win the read access time race. Next are price and capacity. Capacities have reached retail 256GB, with 128GB being the current price 'sweet point'. With many write issues being addressed, it remains only for price to drop to near HDD prices to knock out the 'bread and butter' market segments of consumer HDDs. These market segments have usage profiles that will probably never tax the wear-levelled write-life of SSDs (if the issue never gets completely addressed).

    Regarding your use of the 250GB drive, even if you completely erased and rewrote the entire drive once every day, as a current generation SSD, it would last 27 years without wear-levelling. It comes down to the % that is actually subject to writes.

    With current SSDs, providing excess capacity for that % will significantly mitigate the 'transparent' losses from write-life termination. I say 'transparent' because no data is lost, but erased blocks may become unavailable. That is, existing data is not lost, just overused blocks are removed from being available. Therefore, 'losing data' is a myth - it is capacity that will eventually degrade. This is because there are no read lifetime limitations.

    The latest OCZ SSDs have deliberately understated actual capacities to provide extra margin for possible capacity losses. I do not know if the capacity above that advertised is actually available to the user on a new SSD.

    Sorry. I did misinterpret your heading. Patanjali -- 11/02/09 (in reply to #320122897)

    Regarding misleading to lose data:

    As I said in my other response to your post, the write lifetime limitations NEVER result in data loss, just loss in capacity as they are withdrawn from reuse BEFORE being written to again.

    That is, existing data is indefinitely available. A block will be examined for lifetime limit when it is ABOUT to be written to. If it has reached its write limit, wear-levelling will transparently (unknown to low-level disk utilities and functions) allocate a free block instead.
    This is EXACTLY as modern HDDs do.

    Oh, f*ck me... I stand corrected Mack -- 13/02/09 (in reply to #320122935)

    I guess I've been biased, but due to several factors:

    1. I lost my data so many times it ain't funny. So knowing an SSD has limited write cycles made a bad first impression.
    2. There was a shitload of hype about LCDs, plasmas and HD in the late 90s which brought nothing but disappointment when I saw the quality was a series of sh!t stains.
    3. BitTorrent. It brought way more problems than it solved by replacing conventional P2P networks and completely f*cked up the entire meaning and beauty of P2P. This really demonstrated the public's retardation by needing the files they are looking for spoon fed to them.
    4. Vista. What a JOKE, LMAO!

    So, seeing the way we're progressing in some fields made me skeptical of changes as most of them were going forward into the past. (SSDs did remind me of floppy disks as they deteriorate over time too.)

    "My main reason for going for SSDs is for digital audio recording where the silence and lower operational heat will make mixing much easier. It has two distinct drive usage profiles: "

    You mean, no more noise when capturing Directout audio, or microphone?

    "As I said in my other response to your post, the write lifetime limitations NEVER result in data loss, just loss in capacity as they are withdrawn from reuse BEFORE being written to again.

    That is, existing data is indefinitely available. A block will be examined for lifetime limit when it is ABOUT to be written to. If it has reached its write limit, wear-levelling will transparently (unknown to low-level disk utilities and functions) allocate a free block instead.
    This is EXACTLY as modern HDDs do."

    Much better, I guess they are worthwhile after all, and may replace HDs in a couple years. Thanks for the info.

    Oh, P.S. I read the hightechstrategies page and gotta say **** you. I knew local retards in my town that don't even know what an iPod is, but they freely preach their illiterate bullshit as if they're worth something and with HILARIOUS confidence one said "(making a comparison) were hip, state-of-the-art, just like walkmans and CDs are nowadays." WALKMANS. YEAH, okay. And you describe ME as that kinda laggard? I oughta smack you.
    "Early majority" almost perfectly describes me. I need to be assured something will work before I adopt it.

    Technolgy adoption is selective Patanjali -- 15/02/09 (in reply to #320123122)

    "You mean, no more noise when capturing Directout audio, or microphone?"

    Exactly.

    Regarding which stage in the technology adoption lifecycle one is, I think that it depends upon one's:

    - familiarity with the type of technology

    - confidence level about the particular technology

    - confidencee level in the organisations/individuals making the technology

    - previous bad experience withany of the above.

    I think many these days are in the Early Majority, not wanting to be the raw guinea pigs, but want the benefits. There will be some areas where we jump right in almost before the design has left the drawing board, some where the whole concept appears so left field that we would almost have to be tortured before we would consider it and many that just not relevant to us.

    I went to a doctor the other day who expressed his aversion to computers, all-the-while using his microprocessor-controlled, voice-commanded dictation machine - what he doesn't won"t hurt him.

    Cool Mack -- 17/02/09 (in reply to #320123287)

    "Exactly."

    Awesome, I've completely misjudged this technology. I could say that many players come with a built-in noise filter, but meh. I never realized it was cuz of HD background noise, I thought capturing Waveout bits flowing thru RAM was a lossless process?

    Sorry, I was referring to audio noise in same room Patanjali -- 17/02/09 (in reply to #320123391)

    such as when microphone recording, or even listening to mixes without computer fan noise masking the subtleties of the mix.

    I do not know about the 'Waveout' stuff - at least not from what you have written. Could you please elaborate so that I might be able to understand?

    Upon relection of you statement - all digital is noiseless Patanjali -- 17/02/09 (in reply to #320123434)

    but limited by the bit depth and inherent DAC noise.

    Basically, fast HDDs are noisy and get hot which requires fans, adding more noise. Listening for detail in music is made very difficult by a noisy computer.

    HDD Reliability Anonymous -- 20/02/09 (in reply to #320122466)

    I currently have 6 HDDs that just "died" suddenly, taking with them Gbs of valuable data and hundreds, if not thousands of pictures gathered through the years. The problem is that they did not break down due to old age - they were only months old when they did. The funny thing was that one replacement that I got just worked for 4 mos. - just 1 month after the 1 year warranty for the original equipment expired.
    So much for HDD reliability.

    Waveout Mack -- 22/02/09 (in reply to #320123691)

    How do I explain... have you ever come across a song on Myspace or any kind of streamed audio where its hard to download cuz its stored on a temporary file that you can't copy or open due to it being "in use by another program?" And as soon as you close your web browser, the temporary file is automatically deleted. Well, you open Audacity, Windows sound recorder or whatever, press "Record" while the streamed audio is playing and you capture the PCM audio thats running thru your sound card. It's also the way you convert MIDI to wav. Similarly, you convert non-conventional video formats such as Flash by directly capturing the frames on your screen.
    I don't even know if "Waveout" is the official term, but it's what Microsoft calls it on the Sound settings. "Directsound" is used too. But anyway, theres always noise, even if no sound is playing on your computer and you record via Waveout you will capture noise instead of silence (below perceptibility though.) I dunno if this is due to the 16-bit depth, but I ain't no expert on audio engineering so I never understood what noise has to do with the bit depth.
    Oh yeah, and environmental noise has no effect on what gets captured via Waveout unless your mic is on, which would by default disable waveout.

    To the guy who lost 6 HDDs: my condolences, you were one of the unlucky ones. Your HDDs might've been improperly manufactured, or some other external factor could have caused the failure (e.g. extreme heat or cold in your room.) One of my HDDs died when it was running in a -40C room. The condensation caused all the components to shorten out.

    SSDs have much lower access times than HDDs - good for audio Patanjali -- 27/02/09 (in reply to #320123864)

    Thanks Mack,

    I didn't realise that you were referring to the basic digital streaming interface for Windows. For digital audio recording, the interface tends to be multi-channel and uses ASIO as it allows a lot lower latency (minimum delay between triggering a sound and it actually being heard). Latency (buffer with number of samples) is usually set for worst case (pessimistic) so there is no possibility of audio glitches (no sound because the audio data didn't arrive in time.

    This brings up the other advantage of SSDs - access times of .1 to .3ms (constant for reads) instead of the 3 to 10ms of even the fastest HDDs. When streaming multiple audio tracks or samples, the number available within a given time (latency period) is inversely proportional to the access time. That is, more tracks and/or lower latency with SSDs. A HDD playing back a dozen or so audio tracks or samples could get saturated in seeks well before its tranfer rate is exceeded.

    hmmm.... Anonymous -- 29/01/09 (in reply to #320121653)

    Tier 1 storage vendors have been selling SSD in storage arrays for years. If you use SSD for the right application, you wont have a problem with reliability - They are not designed for random IO or heavy write (i.e the logs you're probably talking about).

    How many day-to-day laptop users would generate high IO workload using office apps? none.

    Storage vendors have suggested that with wear-leveling and block lifecycling, an SSD will last over 3 years - This is with SAN workload not your laptop)

    good read Anonymous -- 25/02/09

    This has been a really interesting read. :-)
    well done to the participants.

Add your opinion

Alex Serpo

Alex Serpo

Reviews Editor

[+] Read bio

Latest Videos

Sponsored content

Power Centre - Content from our premier sponsors

Tags

Back to top

Featured