Jump to content
IGNORED

Solid State Hard Drives


DEANO2

Recommended Posts

1 minute ago, beerandmusic said:

well google then and pick your own......any reliable source will tell you ssd are more reliable...especially in the case of stored media that is not constantly being overwritten (e.g. music files)....

 

The percentage of failure is much less with SSD, and is true in my own experience as well....but any authoritarian on the subject will confirm.

I have HDDs that are over 20 years old and still work without a fault (I've had a fair few fail too). SSDs simply haven't existed long enough for their real-world life span to be assessed.

Link to comment
1 minute ago, beerandmusic said:

I am confident that i have played with more hdd than most here.

Have those 20 year old drives being spinning and performing writes 24x7?

The average hd life just spinning without doing anything is under 6 years.

And what did SSDs look like 6 years ago?

 

Anyhow, the way storage technology is progressing, nobody really needs lifetimes of more than a few years since devices are replaced more often regardless.

Link to comment
34 minutes ago, sandyk said:

In other words, your recommendation here is for HDD, as  SSDs are highly unlikely to be competitive on price per GB in the immediate future.

That suggestion is valid whatever other parameters (e.g. SSD) have been decided on.

Link to comment
33 minutes ago, nbpf said:

Interesting but why would access times matter in the specific context of music playback? I have compared playback from internal SSD, external SSD and external HDD with and without transcoding and I could not hear any difference in my system. It is of course well possible that my system + ears are not very revealing. But I would like to understand the rational behind your argument. I am actually not surprised by my observations: the data are buffered in RAM before being sent to the USB output, I understand. Thus, I would expect that, as long as the buffering is not compromised, access times and throuput have no bearing on the outgoing data stream. I might be missing something, of course.

Your assessment is spot on.

Link to comment
9 minutes ago, jabbr said:

I think an explanation could be the high level e.g. 5V/12V signals, and power spikes from the different electronics. I would expect such differences to be fragile and vary from system to system, cable to cable, cable routing etc, be sometimes present and othertimes not. I run CPUs connected to my DACs from RAM, just because.

How would a computer send data from a storage device to a DAC without passing RAM?

Link to comment
4 minutes ago, nbpf said:

I understand the point but, if power spikes from read operations can have a negative impact on sound quality, then SSDs could perhaps be actually worse than HDDs. If I was concerned about the impact of the electromagnetic field caused by read operations, I would tend to prefer devices that load data to memory at relatively slow, continuous rates to devices that  read in high bandwidth, intermittent bursts.

The data is always sent in bursts at the SATA link speed. When the host wants to read from a SATA device, it sends a request (FIS) for some block of data. The drive then fetches the data from the storage medium into its RAM buffer and signals the host. The host responds by checking which FIS the drive is responding to (multiple requests can be queued) and starting a DMA transfer from the drive to a previously designated destination in system RAM. A slow device will take longer to respond, but once it does, the transfer rate is the fixed. Some host controllers allow forcing a lower link speed than the maximum supported by the drive. If you believe it might make a difference, and your system supports it, you might want to try dropping the SATA link speed to 1.5 or 3 Gbps.

Link to comment
37 minutes ago, darkmass said:

But what may bear some discussion in this thread could be the failure characteristics of spinning drives versus SSDs.

 

Spinning drives are generally reputed to fail with some advance warnings, and even after failure the full data can often be recovered.  On the other hand, SSDs seem to have the reputation of failing like someone flicked a light switch--instantly gone--and with no data recovery possible.  As always, of course, backing up important data is necessary no matter what type of drive.  But it strikes me that SSDs generally have a much more catastrophic way of failing.

This aspect has me wary of SSDs as well. My OS and important data are on a RAID-6 of enterprise HDDs. I also have an NVMe SSD for throwaway data that just needs to be fast.

Link to comment
4 minutes ago, darkmass said:

As a thought experiment, it seems to me that using SSDs for music server storage has its particular risks.  The music would generally be "write-once", to get an item into place, then selections would potentially be read over and over from those fixed storage locations.  The reading part of that seems to me to be just the thing to electrically wear an SSD.  And SSD storage locations can only survive a finite amount of wear.

NAND flash is tricky stuff. Blocks have a limited number of erase cycles, and stored data fades over time. Even reads cause incur some wear. For reliable operation, writes need to be spread evenly, and static data needs to be refreshed. Unlike HDDs, SSDs have no such thing as write-once blocks. Any block can be relocated at any time. It's all down to the controller, of which you have no knowledge.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...