nMP pricing compared to 2012

[slughead]

One thing I wonder… Perhaps the spartan selection and high prices for TB enclosures are more a symptom of lack of demand than anything else?

That's got to be a part of it. However, Intel has created high barriers to entry into the Thunderbolt market which has created an oligopoly. The big companies that have been "blessed" to create thunderbolt solutions have no small upstarts to compete with thanks to Intel.

I also agree there hasn't been a lot of demand due to the current batch of machines that use TB, and that the nMP will increase demand. However, the level at which it will increase demand for TB2 devices is going to directly relate to the amount of devices sold. Apple's not exactly dumping these things on the market with these high-price points. It'll take a heck of a lot of nMP sales before the TB market shifts and more peripheral-makers jump in.

In time, if there's enough money in it, other companies will be dragged in. For now, though, it's not looking good.

 

[iBug2]

When you buy your ARECA TB array, you're buying a TB SATA controller, a PSU, and a box to put it all in. The problem with TB storage is you're locked into these expensive packages. If you could purchase the TB controller separately (and it didn't cost an arm and a leg ($900 for ONE SAS port), or lacked features (Two eSATA II for $200)), it would make it easier. Things will likely get better for the TB market, but don't forget that competition is strictly limited here as Intel is suing the crap out of anyone who doesn't bow to them before releasing a product.

The TB controller adds around 50-150$ to an enclosure so far. Enclosures with USB or Ethernet or Firewire cost 50-150$ less than the same enclosures with TB and that's a quite small fee. Yes, there aren't any cheap ones, ones that cost 25$ like you can find for USB 3.0. But isn't that a good thing? You keep saying that an enclosure has to be built well because you are risking your precious data if it goes bad. And TB enclosures are usually high end ones aimed at video professionals mostly and those guys probably won't go for very cheap boxes to put their data in anyway.

 

[slughead]

The TB controller adds around 50-150$ to an enclosure so far.

I've heard that. Provided that's true, the question is: What the heck are you paying for?

But isn't that a good thing? You keep saying that an enclosure has to be built well because you are risking your precious data if it goes bad. And TB enclosures are usually high end ones aimed at video professionals mostly and those guys probably won't go for very cheap boxes to put their data in anyway.

Just because you pay more doesn't mean you get more. The 3-drive caldigit model you were going gaga over had a freakin power brick.

Any "professional" trusting their data to a power brick has my sympathies.

The only thing "not cheap" about TB enclosures universally is the price. The PSU quality (which, apart from maybe vibration, is the most important thing) is not reflected therein. I'm sure there are a few with good quality PSU, but you pay way more for them.

 

[iBug2]

I've heard that. Provided that's true, the question is: What the heck are you paying for?

I don't know. What are you paying for when you buy a 1350$ 8 bay enclosure with ethernet and USB?

Just because you pay more doesn't mean you get more. The 3-drive caldigit model you were going gaga over had a freakin power brick.

Yeah and it's claimed dead silent, so of course I'm gaga over it. Never heard about an enclosure that silent. I'll let you know if I ever lose a drive in it.

The only thing "not cheap" about TB enclosures universally is the price. The PSU quality (which, apart from maybe vibration, is the most important thing) is not reflected therein. I'm sure there are a few with good quality PSU, but you pay way more for them.

I don't know about the various PSU qualities of Promises or Lacie enclosures but probably they wouldn't get good reviews all over with really bad PSU's in them. We'll see when the Calldigit one is released and what type of reviews it gets.

But to me, noise levels are more important than PSU quality so what's most important is subjective. So even though if someone tells me that a Pegasus has a better PSU than Calldigit and I have 0.1% less chance to lose data with that one, I'd still buy Calldigit because Pegasus has a 48 dB noise level, which is very loud.

 

[slughead]

I don't know about the various PSU qualities of Promises or Lacie enclosures but probably they wouldn't get good reviews all over with really bad PSU's in them.

Of course you don't know about the various PSU qualities of external enclosures, nobody does, including the reviewers!

What are they supposed to do? buy 1000 of them and put them through the paces? Meanwhile companies like the PC PSU manufacturers have to deal with thousands of reviewers on NewEgg and other shops posting star ratings. I'd trust the cumulative data on NewEgg over a paid thumbs up or down review on MacFancy Magazine any day.

But to me, noise levels are more important than PSU quality so what's most important is subjective.

At least we agree on something: If you don't care about data integrity, you have a lot more options with thunderbolt! I very much hope you enjoy your equipment.

Meanwhile, I am about to upgrade the PSU in my array with an 80 Plus certified one with 800 positive ratings on NewEgg, as the data outside my case is as important to me as the data inside.

 

[AidenShaw]

...
Meanwhile, I am about to upgrade the PSU in my array with an 80 Plus certified one with 800 positive ratings on NewEgg, as the data outside my case is as important to me as the data inside.

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops. The power company transformer isn't redundant, though - if the grid goes down the two UPS systems keep everything up for 15 minutes or so - more than long enough for the dirty caches to flush without relying on the cache backup.

Everything but scratch space is on RAID-60.

I like my data....

 

[MattDSLR]

I still do not understand limited expandability in the new Mac Pro minis
Will apple add something before official release?
I will pay extra 500 for extra two slots

 

[nateo200]

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops. The power company transformer isn't redundant, though - if the grid goes down the two UPS systems keep everything up for 15 minutes or so - more than long enough for the dirty caches to flush without relying on the cache backup.

Everything but scratch space is on RAID-60.

I like my data....

That's hardcore.....

 

[VirtualRain]

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops. The power company transformer isn't redundant, though - if the grid goes down the two UPS systems keep everything up for 15 minutes or so - more than long enough for the dirty caches to flush without relying on the cache backup.

Everything but scratch space is on RAID-60.

I like my data....

Wow... Hardcore indeed. . What kind of data warrants that level of protection?

 

[Varmann]

Wow... Hardcore indeed. . What kind of data warrants that level of protection?

Money lost losing it?
Time spent creating it?
For most companies the hardware is a small fraction of the total cost.

(For many consumer as well. Funny to hear someone complaining about a $10 price difference of a game, then playing it for 200 hours. Is $0.05/hour that important?)

 

[thekev]

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops. The power company transformer isn't redundant, though - if the grid goes down the two UPS systems keep everything up for 15 minutes or so - more than long enough for the dirty caches to flush without relying on the cache backup.

Everything but scratch space is on RAID-60.

I like my data....

You're my hero.

 

[iBug2]

I'd trust the cumulative data on NewEgg over a paid thumbs up or down review on MacFancy Magazine any day.

I wouldn't say Anandtech is a Macfancy magazine but anyway.

At least we agree on something: If you don't care about data integrity, you have a lot more options with thunderbolt! I very much hope you enjoy your equipment.

Meanwhile, I am about to upgrade the PSU in my array with an 80 Plus certified one with 800 positive ratings on NewEgg, as the data outside my case is as important to me as the data inside.

All Pegasus PSU's are 80 Plus Bronze certified. And I didn't say I don't care about data integrity. You make it sound like it's binary. Of course if someone said that one PSU has 50 times as much failure as another, I'd choose the better one. But what's the risk of failure you are adding with a slightly worse PSU? Nothing is foolproof in tech world. Everything can fail. The important thing is the actual numerical values of the risk you are taking.

----------

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops. The power company transformer isn't redundant, though - if the grid goes down the two UPS systems keep everything up for 15 minutes or so - more than long enough for the dirty caches to flush without relying on the cache backup.

Everything but scratch space is on RAID-60.

I like my data....

So you suggest that we all get what you have for our home office? You know that's overkill for most small businesses. I'm pretty sure our genetics research facility in the university has some sophisticated data protection like that but that's a research lab.

I have a smart online UPS, connected to it two really high quality 5 plug outlets and then the computer equipment. The UPS takes care of any kind of irregularities with the line voltage due to double conversion, and the outlets are good enough for a computer setup that costs around 5K. And I view even my own setup as a bit overkill for my needs. I want to get rid of the UPS because even through 2 doors, I can hear the fan noise.

And because I also love my data, I take regular backups. The only data I've ever lost are the ones I didn't backup, and they were basically junk I could replace immediately.

 

[slughead]

All Pegasus PSU's are 80 Plus Bronze certified.

And like I said, you're paying for it.

You should also know 80 Plus only has to do with the power consumption and not the reliability, for that you need product testing like the 800 reviews my PSU got on NewEgg. I'll assume anyway that Pegasus uses good PSUs (though I do not correlate price to quality): I could still get better performance (2GBps) and the same (or better) reliability for $75 in my current Mac Pro (SATA III card, alternatively you could do a $75 eSATA with $125 enclosure with a decent box and excellent PSU). I'm sure the TB2 model will be the same price (eventually), but that only brings it up to par with what I just built.

Edit: BTW: 1/4 of the reviews of this thing talk about drive failure, RAID problems, and power issues.

----------

Every array in my lab has dual controllers with dual power supplies, with internal batteries/capacitors/flash for saving dirty cache, connected to two separate UPS systems, connected to two independent power drops.

Your pornography must be very important to you (j/k )

 

[AidenShaw]

That's hardcore.....

Actually, it's not "hardcore" - it's standard practice for enterprise storage.

 

[iBug2]

And like I said, you're paying for it.

The problem with those boxes is that they come with drives in them and the price per drive is much higher than retail. The box is around 500$ and they are charging another 500$ for 4x1TB drives, which costs no more than 250$. Unfortunately Caldigit is also selling their boxes with drives in them, so does Lacie.

 

[slughead]

The problem with those boxes is that they come with drives in them and the price per drive is much higher than retail. The box is around 500$ and they are charging another 500$ for 4x1TB drives, which costs no more than 250$. Unfortunately Caldigit is also selling their boxes with drives in them, so does Lacie.

As a practical matter, if you can't find a good product for TB, nothing else matters. It really doesn't matter if theoretically, somehow, someday, someone will offer a decent product at a decent price. Theoretically doesn't get you a high quality product at a decent price.

After about 10 minutes of googling, I managed to figure out that Pegasus uses regular retail consumer-grade hard drives, not anything special ( I could not even locate information on the drives in the R4 ). In fact, according to NewEgg, 25% of reviewers reported failure with the same HD Pegasus uses in their R6 which retails for $2,300.

Meanwhile, I can throw four WD Black (workstation grade SATA with 5 year warranty) drives for $360 into either my Mac Pro or a $250 home-built array+SATA III controller. Higher quality, lower price, period.

Last I checked, $500 is a lot more than $250. So again: you're paying a high price for limited options with thunderbolt.

 

[theSeb]

Meanwhile, I can throw four WD Black (workstation grade SATA with 5 year warranty) drives for $360 into either my Mac Pro or a $250 home-built array+SATA III controller. Higher quality, lower price, period.

I hope that you're not planning to run them in RAID then.

 

[haravikk]

I hope that you're not planning to run them in RAID then.

Why not? WD Greens aren't much good in a RAID, but the other types are all good options for different reasons; Blues if you just want it to be inexpensive (they're not great, but they're not bad either), Reds if you want energy saving and consistent responsiveness (good for RAID-5/6, particularly when doing stuff other than just streaming big files), and Blacks for a pure performance RAID.

I wouldn't pick Blacks over Reds for a system where heat may be an issue, but basically I see Reds as the option for capacity RAID setups, and Blacks for performance RAID setups (assuming you need more capacity than SSDs can give of course).

I'm probably erring towards Reds myself, but it'll be a while before I can replace all my drives with WD Reds so I'll have a few other drives mixed in for a while.

 

[iBug2]

As a practical matter, if you can't find a good product for TB, nothing else matters. It really doesn't matter if theoretically, somehow, someday, someone will offer a decent product at a decent price. Theoretically doesn't get you a high quality product at a decent price.

After about 10 minutes of googling, I managed to figure out that Pegasus uses regular retail consumer-grade hard drives, not anything special ( I could not even locate information on the drives in the R4 ). In fact, according to NewEgg, 25% of reviewers reported failure with the same HD Pegasus uses in their R6 which retails for $2,300.

Meanwhile, I can throw four WD Black (workstation grade SATA with 5 year warranty) drives for $360 into either my Mac Pro or a $250 home-built array+SATA III controller. Higher quality, lower price, period.

Last I checked, $500 is a lot more than $250. So again: you're paying a high price for limited options with thunderbolt.

Maybe you are having trouble understanding my points on every post. I'm not saying TB is cheaper. I'm saying TB is not as expensive as you try to make it out to be. 1000$ box contains drives. And of course people will complain when the drive fails. Is it because of the box or because every drive can fail? I can find you someone with a failed WD Black in a second. ME! I had one WD Black fail so far, and 3 WD Green's and one WD Velociraptor. All inside my Mac Pro. So drives can fail wherever you put them. End of story.

----------

Why not? WD Greens aren't much good in a RAID, but the other types are all good options for different reasons; Blues if you just want it to be inexpensive (they're not great, but they're not bad either), Reds if you want energy saving and consistent responsiveness (good for RAID-5/6, particularly when doing stuff other than just streaming big files), and Blacks for a pure performance RAID.

I wouldn't pick Blacks over Reds for a system where heat may be an issue, but basically I see Reds as the option for capacity RAID setups, and Blacks for performance RAID setups (assuming you need more capacity than SSDs can give of course).

I'm probably erring towards Reds myself, but it'll be a while before I can replace all my drives with WD Reds so I'll have a few other drives mixed in for a while.

For data integrity one should go with Xe's since they are the enterprise storage WD offers. Blacks are for people who want speed, not protection. RED is a compromise on both accounts but better than Black for RAID.

 

[wallysb01]

For data integrity one should go with Xe's since they are the enterprise storage WD offers. Blacks are for people who want speed, not protection. RED is a compromise on both accounts but better than Black for RAID.

You guys do releaze that for "normal" people with just one relatively small RAID (thinking <10 disks here and often <4), the differences in failure rates are not very real world applicable. The best thing to do is just have some ironclad backup off site and some more regular backup onsite. Once you've done that, small changes in reliability make little difference. These disk reliability concerns are more for that guy running a RAID-60, with maybe 30 disks or so, who's much more likely to see failures and rebuilds. Or for clusters with PBs of storage, who regularly changing disks.

IMHO people paying an extra 50-100% on disks just reliability when they are only using 3-4 of them are thowing money away. At that point, you'd rather just have 50-100% more disks to reduce your data loss chances. Obviously, its a little different to tell someone with 4 disks to do that, than someone with 40, or a cluster 400.

 

[theSeb]

Why not? WD Greens aren't much good in a RAID, but the other types are all good options for different reasons; Blues if you just want it to be inexpensive (they're not great, but they're not bad either), Reds if you want energy saving and consistent responsiveness (good for RAID-5/6, particularly when doing stuff other than just streaming big files), and Blacks for a pure performance RAID.

I wouldn't pick Blacks over Reds for a system where heat may be an issue, but basically I see Reds as the option for capacity RAID setups, and Blacks for performance RAID setups (assuming you need more capacity than SSDs can give of course).

I'm probably erring towards Reds myself, but it'll be a while before I can replace all my drives with WD Reds so I'll have a few other drives mixed in for a while.

IBug2 has already mentioned some of the reasons, but another reason is that WD recommends against using WD blacks in raid.

 

[iBug2]

You guys do releaze that for "normal" people with just one relatively small RAID (thinking <10 disks here and often <4), the differences in failure rates are not very real world applicable. The best thing to do is just have some ironclad backup off site and some more regular backup onsite. Once you've done that, small changes in reliability make little difference. These disk reliability concerns are more for that guy running a RAID-60, with maybe 30 disks or so, who's much more likely to see failures and rebuilds. Or for clusters with PBs of storage, who regularly changing disks.

IMHO people paying an extra 50-100% on disks just reliability when they are only using 3-4 of them are thowing money away. At that point, you'd rather just have 50-100% more disks to reduce your data loss chances. Obviously, its a little different to tell someone with 4 disks to do that, than someone with 40, or a cluster 400.

That's true, I use only Green's now because they are ultra cheap and I can just buy 2 and use one as backup if some data is crucial. I only responded because the poster said his data is important and he cares a lot about drive failure. If that's your main priority Xe's are the way to go. So far out of 35 Green's I bought, I've lost 3. So that's almost 10% failure rate. They weren't DOA but eventually failed.

 

[haravikk]

IBug2 has already mentioned some of the reasons, but another reason is that WD recommends against using WD blacks in raid.

Sure, because that's what they position Reds for. For a shared RAID that's always online and potentially serving a whole load of read/write requests simultaneously then sure, Reds are the best option as they'll save energy and cope well with lots of requests thanks to their consistently low seek time.

But for a direct-attached RAID I think Blacks are the better option, as you're not going to subject it to same simultaneous strain that multiple users will, and you only need it to be online while your computer is (i.e - it's not likely to be in use 24 hours of every day). They'll also perform better for that kind of use-case.

WD will definitely recommend the Reds for any RAID, because that includes a range of different setups and they'll do well in all them. But not every RAID is over a network with multiple users.

Blacks are for people who want speed, not protection. RED is a compromise on both accounts but better than Black for RAID.

I'm not sure that's really true, as WD Blacks now have a 5 year warranty, and accuracy is a big part of the sales pitch, which means for protection they should be just fine. Besides, the point of RAID setups like RAID-5 and 6 is that you can use relatively unreliable drives without fear of a failure destroying all your data. Sure that doesn't mean you should use really failure prone drives, but all drives have the potential to fail after all, so do RAID controllers, PSUs etc.

Point is, I don't believe there's a big difference between how Blacks and Reds perform in a direct attached RAID, so if you're willing to pay extra then Blacks will be that bit faster.

 

[AidenShaw]

Sure, because that's what they position Reds for. For a shared RAID that's always online and potentially serving a whole load of read/write requests simultaneously then sure, Reds are the best option as they'll save energy and cope well with lots of requests thanks to their consistently low seek time.

But for a direct-attached RAID I think Blacks are the better option, as you're not going to subject it to same simultaneous strain that multiple users will, and you only need it to be online while your computer is (i.e - it's not likely to be in use 24 hours of every day). They'll also perform better for that kind of use-case.

WD will definitely recommend the Reds for any RAID, because that includes a range of different setups and they'll do well in all them. But not every RAID is over a network with multiple users.


I'm not sure that's really true, as WD Blacks now have a 5 year warranty, and accuracy is a big part of the sales pitch, which means for protection they should be just fine. Besides, the point of RAID setups like RAID-5 and 6 is that you can use relatively unreliable drives without fear of a failure destroying all your data. Sure that doesn't mean you should use really failure prone drives, but all drives have the potential to fail after all, so do RAID controllers, PSUs etc.

Point is, I don't believe there's a big difference between how Blacks and Reds perform in a direct attached RAID, so if you're willing to pay extra then Blacks will be that bit faster.

The Reds have firmware that's optimized for RAID operation (different error recovery, timeouts, etc). RAID edition drives are also better able to deal with vibration than desktop drives.

 

[slughead]

I hope that you're not planning to run them in RAID then.

Then buy 4x2TB WD Red drives for $440--still cheaper, still better than the regular drives included with these Pegasus rip-offs... And with twice the space.