OS hard drive - SSD, mechanical or hybrid?

onebuck · 09-02-2015, 09:31 AM

Hi,

Quote:

Originally Posted by wroom

Not at all.
A mechanical drive with a faulty reader arm can be repaired to read out the data.
But how do you mend a silicon chip that has come to be deteriorated by EOS?

https://en.wikipedia.org/wiki/Failur...of_electronics

Not a head crash! Forensics are limited with this type of damage. Media is beyond reading when a head does crash. No way you are going to read data from a damaged cylinder track.
If you have an actuator failure then that could be repaired as long as the disk media is not damaged. Actuator/ARM failures will usually create damaged media surface via a head crash. The flying head is not in contact with the media. One good example, is a 747 is flying inches above a runway and a rabbit attempts to run under the plane. Proportionally this would cause a crash by a impurity that cause contact with the head & disk media. The head will no longer fly but will crash into the media thus damaging any data on that cylinder. Even vibrations from a failure in the mechanics can cause head crashes that damage media surface beyond chance of any data recovery.

Quote:

Originally Posted by wroom

What will happen to the files stored in silicon charge tunnels when there is "Hot carrier deterioration" going on in the chip?
The same physical phenomenon used to store the data in FLASH, but going rogue.
Ever thought about the name of the "FLASH" technology being an engineers pun? Well, it actually is.
The silicon business is full of acronym pun. Like the "EOS" digital photography. Being that EOS in the silicon die may have been the biggest obstacle when developing the technology.

FUD! Impurities can cause issues but QC does weed this failure out by not allowing the use of material that is inferior as Semiconductor device defects. As to your use of 'EOS as meaning 'Electro-Optical System' is mixing technologies and not really relative directly to 'FLASH memory', each type will use different control interfaces;

Quote:

From FLASH;
Flash memory is an electronic non-volatile computer storage medium that can be electrically erased and reprogrammed.
Introduced by Toshiba in 1984, flash memory was developed from EEPROM (electrically erasable programmable read-only memory). There are two main types of flash memory, which are named after the NAND and NOR logic gates. The individual flash memory cells exhibit internal characteristics similar to those of the corresponding gates.
Whereas EPROMs had to be completely erased before being rewritten, NAND type flash memory may be written and read in blocks (or pages) which are generally much smaller than the entire device. NOR type flash allows a single machine word (byte) to be written—to an erased location—or read independently.
The NAND type is primarily used in memory cards, USB flash drives, solid-state drives (those produced in 2009 or later), and similar products, for general storage and transfer of data. NAND or NOR flash memory is also often used to store configuration data in numerous digital products, a task previously made possible by EEPROM or battery-powered static RAM. One significant disadvantage of flash memory is the finite number of read/write cycles in a specific block.[1]
Example applications of both types of flash memory include personal computers, PDAs, digital audio players, digital cameras, mobile phones, synthesizers, video games, scientific instrumentation, industrial robotics, medical electronics, and so on. In addition to being non-volatile, flash memory offers fast read access times, as fast as dynamic RAM, although not as fast as static RAM or ROM. Its mechanical shock resistance helps explain its popularity over hard disks in portable devices, as does its high durability, being able to withstand high pressure, temperature, immersion in water, etc.[2]
Although flash memory is technically a type of EEPROM, the term "EEPROM" is generally used to refer specifically to non-flash EEPROM which is erasable in small blocks, typically bytes.[citation needed] Because erase cycles are slow, the large block sizes used in flash memory erasing give it a significant speed advantage over non-flash EEPROM when writing large amounts of data. As of 2013, flash memory costs much less than byte-programmable EEPROM and has become the dominant memory type wherever a system requires a significant amount of non-volatile, solid-state storage.

Controllers for each type will handle the cells and how presented to the end user for reads & writes.

Quote:

Originally Posted by wroom

Did you know that radio interference can fry a silicon chip?
The EM pulse from a nearby lightning strike can easily take out a silicon chip that is powered on.
The magnetic media is much, much more resilient to such failures.

The power that is radiated will dictate damage. If reflected and not filtered/shielded then certainly the chance of damage is probable.

Quote:

Originally Posted by wroom

I'm saying i prefer a brick house before a straw house - Even if the proverbial wolf may be able to attain mortar artillery. Main reason being: The bricks won't catch fire by themselves.

Simplistic!

Quote:

Originally Posted by wroom

I fully agree. That was my main point. Use at least two drives in a raid mirror.
If making an array of SSD's, then use two different brand drives. It will lessen the probability of the two drives failing at the same time.

Different manufactures will not insure that failure will not occur in any configuration for storage of data. FUD, no real data to support such a configuration.

Quote:

Originally Posted by wroom

The failures of magnetic media has more locality, which limits the damage to some of the data.
The failures of silicon flash on the other hand, is more of chip global.

A scratched magnetic disk can be read. But a fried FLASH chip cannot be powered on.

I have repaired Hard disk and can say the repair of the disk media is near to impossible. If the problem is the disk interface then yes the likely repair can be done. To read a damaged media you will get boundary information but not in line track reads from that head crash on the media. Yes, Solid state media that has been charge damaged is nearly impossible to read. That is why you should protect any solid state media, most damage is poor handling or power filtering issues. Backups! But that can be said for the computer overall that one should protect the system from power fluctuations or potential signal noise. PM is something that can cause more damage to the system as a whole if the signal magnitude is sufficient. One reason our satellite systems are shielded to hopefully prevent this type of damage(solar Flares/PM created from solar emissions or intentional human created PM). One can shield their system but not everyone can afford to have a Faraday cage.

'SSD' technology is improving and surpasses earlier technology;

Quote:

From http://www.zdnet.com/article/samsung...ces-16tb-ssd/#
And the heaviest company in NAND flash is Samsung, who demonstrated their leadership in a keynote address. They announced several breakthroughs:

The 48 layer, 256Gbit (32GB), vertical NAND flash chip.
The world's largest - almost 16TB - drive.
The world's fastest - 1,000,000 IOPS - drive.
A 2U box with over 700TB of capacity - roughly 15PB in a single rack.
And a continuation of their intelligent drive initiative.

That last is the most important. Modern drives contain powerful CPUs that can do much more than manage blocks of bits.
Rather than shipping raw data to hosts, these CPUs could perform local processing, giving hosts only the data they need while reducing network loads and latency. Samsung engineers are working with standards bodies to develop APIs to enable device computation.

Talk about changes in technology! Please note the above concerning the CPU controller usage. To me that is the way of the future to present data to and from a storage device. Nice!

wroom · 09-02-2015, 12:40 PM

Quote:

Originally Posted by wroom

What will happen to the files stored in silicon charge tunnels when there is "Hot carrier deterioration" going on in the chip?
The same physical phenomenon used to store the data in FLASH, but going rogue.

Quote:

Originally Posted by onebuck

FUD! Impurities can cause issues but QC does weed this failure out by not allowing the use of material that is inferior as Semiconductor device defects.

FUD? Impurities?
I'm not talking about impurities. I'm talking about "Hot carrier deterioration", and it's not caused by dust mites.

Quote:

Originally Posted by wroom

Ever thought about the name of the "FLASH" technology being an engineers pun? Well, it actually is.
The silicon business is full of acronym pun. Like the "EOS" digital photography. Being that EOS in the silicon die may have been the biggest obstacle when developing the technology.

Quote:

Originally Posted by onebuck

As to your use of 'EOS as meaning 'Electro-Optical System' is mixing technologies and not really relative directly to 'FLASH memory', each type will use different control interfaces;Controllers for each type will handle the cells and how presented to the end user for reads & writes.

I think you misunderstood my writing. I was not clear enough in what i said.
EOS is a term in the silicon industry that is spelled out "Electrical Over Stress".
It happens when the frequency*voltage in the conductors on the chip is more than what the fine geometries on the chip can handle. A combination of temperature and electromagnetic force starts moving the etched geometries around, and finally it fries.

The engineer pun i refer to, is for FLASH technology, that some engineers found it to be a given that it should be called FLASH, since they had seen so many flashes in the lab during the development of the technology. Same pun with EOS. Maybe a very dry sort of humor in the silicon business?

Quote:

Originally Posted by wroom

If making an array of SSD's, then use two different brand drives. It will lessen the probability of the two drives failing at the same time.

Quote:

Originally Posted by onebuck

Different manufactures will not insure that failure will not occur in any configuration for storage of data. FUD, no real data to support such a configuration.

The same advice goes for magnetic media too. How probable do you think it is that two differently designed units manufactured at different times and places will fail at exactly the same time?
How possible is it that a raid of only WD green drives manufactured right after the tsunami happened will completely fail, botching all the data? 100%?

It is a good advice. And it didn't come from me originally.

BTT

onebuck · 09-02-2015, 03:11 PM

Hi,

Quote:

Originally Posted by wroom

FUD? Impurities?
I'm not talking about impurities. I'm talking about "Hot carrier deterioration", and it's not caused by dust mites.

'HCI' is hot carrier injection and is not a deterioration but is phenomenon when a hole or electron has sufficient energy to break a potential barrier. This is due to poor design and usually occurs in MOS based solid state devices. And is the means to destroy a solid state device via the excess current that will break the potential barrier down.

When I was speaking of impurities it was when the substrate is created during the doping of the die. One must keep these impurities within dopant to a minimum or the die is useless when creating a solid state device be it a transistor or multiples to create integrated circuitry.

Quote:

Originally Posted by wroom

I think you misunderstood my writing. I was not clear enough in what i said.
EOS is a term in the silicon industry that is spelled out "Electrical Over Stress".
It happens when the frequency*voltage in the conductors on the chip is more than what the fine geometries on the chip can handle. A combination of temperature and electromagnetic force starts moving the etched geometries around, and finally it fries.

Quote:

Originally Posted by wroom

The engineer pun i refer to, is for FLASH technology, that some engineers found it to be a given that it should be called FLASH, since they had seen so many flashes in the lab during the development of the technology. Same pun with EOS. Maybe a very dry sort of humor in the silicon business?

Familiar with EOS in those terms. I mistook your reference to FLASH with digital photography and mistook it as relative to your "Like the "EOS" digital photography" meaning 'EOS as meaning 'Electro-Optical System'. Not as now reference of Electrical over stress/EOS that you now use as the term in reference to the solid state devices..

Quote:

Originally Posted by wroom

The same advice goes for magnetic media too. How probable do you think it is that two differently designed units manufactured at different times and places will fail at exactly the same time?
How possible is it that a raid of only WD green drives manufactured right after the tsunami happened will completely fail, botching all the data? 100%?

It is a good advice. And it didn't come from me originally.

BTT

Not with 'SSD' since each device may use a different controller type thus a mismatch in operation therefore the efficiency may fall since one must/should configure the 'SSD' to the system for effective efficient performance. If you feel that the use of different manufactures is a doable function then do so for your systems. Personally, I would not since performance may be something that I prefer to have my systems function without issues. I would suggest everyone to read manufactures data for the 'SSD' to allow proper setup for their system and configuration. The likely failure for a 'SSD' is not from proper functions within the device but likely from external issues. Failure rates for 'SSD' are low and most devices will not fail over the manufactures life of the device. Earlier 'SSD' had issues with multiple writes but the new generation 'SSD' controllers alleviate this problem thus extending the life. Again read the manufactures spec sheets.

Do I favor 'SSD' over 'HDD'? Yes! My primary device(s) will be 'SSD while a 'HDD' will be for secondary storage. I am not saying that 'HDD' does not have a use in the system but performance can be improved with the use of a 'SSD' which can be optimized for a system.