Running cost of 2TB SSD SHR vs HD SHR how does the high electricity cost impact running costs?

Currently reading
Running cost of 2TB SSD SHR vs HD SHR how does the high electricity cost impact running costs?

953
318
NAS
DS620slim, DS415+
Operating system
  1. Linux
  2. macOS
  3. Windows
Mobile operating system
  1. Android
  2. iOS
Following a recent post on cost of HD vs SSD, I checked the current cost, taking into account the high energy costs that we currently pay.
I took SHR 3x2TB for SSD and HDD for people like me that use max 3TB of disk space.

I compare WD red Pro 2TB vs MX500 2TB SSD Amazon price by today, electricity costs by today in Europe.

WD RED Pro
3 x 2TB hard disk: 3* 106 euro = 318 euro for the 3 disks / 5 year lifetime = 63 euro/year on disks
electricity: 3 disks * 6 watt * 24h *365 Days /1000 * 0.5 euro/kWh = 78 euro/year
Total HD cost 63+78 = 141 euro/3 disks/year.

SSD
3x MX500 2TB; 3*170 euro = 510 euro / 6 years lifetime (MTBF is almost twice as high as HD, I will add 1 year of lifetime) = 85euro/year on disks.
electricity: 3disks * 0.5watt * 24h*365days /1000 *0.5 euro/kwh = 7euro/y
Total SSD cost 85+7= 92 euro/3 disks/year

For sure with bigger disks the SSD costs are not competitive, but I was surprised that electricity has such a big impact on disk running costs.
Also regular WD red have lower price and electricity cost is about half of the Pro, but still you will notice that electricity costs really impact the running costs of disks a lot at the moment.

Another reason to switch off the nas at night :)
 
Last edited:
Another reason to switch off the nas at night
So this got me thinking 😜 Synology's power off/power on feature... since it's capable of restarting the NAS, then it must not fully power off. So... is the "power off" state, the same as when the NAS goes into "safe mode" under a UPS provision... or... if it differs... how so? Saving some shillings seems right for basic home use.
 
My router has USB 3.0 "NAS" capability. I have a 3TB WD Red sat in a caddy attached to it, containing my essential files.

It works rather well actually. Shared with my Windows machines using SMB and the router tells me it is using 14w.

I turn the NAS back on when it is needed. Which hasn't been the case since I turned it off! :D
 
Last edited:
WD Red Pro 2TB, PN: WD2002FFSX

AVG R/W: 7.8W
Idle: 6.0W
Sleep: 1.4W

but that's energy consumption for direct current (DC). You have to use alternating current (AC) consumption, which happens in the NAS transformer. That transformer has efficiency (losses during AC/DC conversion). You can safely calculate with a factor of 0.7.

Then from the invoicing point of view of your consumption, you need to calculate with:
AVG R/W: 11.14W
Idle: 8.57W
Sleep: 2W

However, your disks are not constantly in:
R/W or Idle or Seek mode.
So you have to find a formula to smooth this because you won't find that data in smartctl tables, except PoH.

Note:
Crucial does not have official data on environment data on its website (strange).
But they did not forget to write:
45x more energy efficient than a typical hard drive.
Explained by them:
Active average power use comparison based on published specs of the 1TB Crucial MX500 SSD and the 1TB Western Digital® Caviar Blue™ WD10EZEX internal hard drive. All other capacities of the Crucial MX500 SSD have comparable active average power consumption specs.
With all due respect, these Western Digital® Caviar Blue™ WD10EZEX had End of production a long time ago. Their environment (consumption) was about:
AVG R/W: 6.8W
Idle: 6.1W
Sleep: 1.2W
link here

but I discovered the real measurements for The MX500 1TB:
AVG R/W: 2.4W
Idle: 0,102W
Sleep: doesn't exist for SSD

Then (Edit: corrected table):
1664526082357.png


Then the MX500 is just 13x more efficient. (Edit: corrected value from 14 to 13)

Maybe it would be good to see the lifespan of that SSD vs HDD because that could show another parameter = TCO over 10 years.
Because 2TB MX500 is about 700TBW
and WAF = 8.57!! (3k P/E cycles)... it is better to find by smartctl data written to flash / data written to the host


However,
I think that HDD is the winner there in the case of 10y TCO period. Because you need to take into account the durability of such SSD.
But I understand that everyone has his own point of view.
In any case, I left full-time spinning in all of my NASes 2y ago, except the Primary site.
 
I'm not a fan of turning the NAS off. These drives are built with the expectation of running 24/7, they reach an "operating temperature" and should remain at that temperature. Like anything made of metal, they expand and contract with heating and cooling and even a slight change can make data unable to be accessed. Also, powering up sends a power surge that wears on the circuitry over time causing potential failures. I do all that I can to insure power is always on, UPS and backup generator. The cost of drive or NAS replacement far outweighs the cost of electricity to me.
 
@jeyare - I think you mean 'sleep' rather than 'seek' in your msg & table above. HDDs certainly do 'seek', but this isn't usually a low power operation. Also I think the blue heading in your table is for the MX500 rather than the WD Red Pro.

When the differences in power draw are laid out like that and calulated over a year it is certainly an eye opener.

Personally I think the 'wear & tear' argument about HDDs stopping & starting are likely overstated; NAS drives are designed with a number of start/stop cycles in their specification (my Ironwolf Pros are specced at 600000), and an operating temperature range (5-70degC) just as much as they're designed to be run 24/7. I think modern good quality NAS type HDDs are just as happy with 24/7 operation, frequent spin down to save power or anything in between.
 
When the differences in power draw are laid out like that and calulated over a year it is certainly an eye opener
OFC the Power consumption is dependent on the Operation variables:
  • more idle time = more savings or vice versa
  • you need to know that these numbers are defined by lab environment e.g. 25C ambient temp, the tested drive is not located in a closed cabinet (e.g. in NAS), ...
  • idle time is sometimes occupied by services like SMART, then the consumption is increasing (up to your setup)
  • ...

Finally - there is not a single static number in real operation for R/W or another indicator. Don't take the vendor's number as a single source of truth. There is just one way how to check it - measurement outside NAS = DC.

I have been pointing out for a long time that the best specification kind of doc for drives (HDD, SSD, NVMe) are made by Seagate. This cannot be compared with the absence of such data from WD. Look at the data on the consumption of a specific PN:
A recommendation for search:
use " Seagate <PN> product manual" and you will get a similar PDF
I stopped buying WD HDD more than 12 years ago. And the only drive brand I would buy from them is HGST.

Here is an old (2008) analysis, but still valid (principles point of view) - An Analysis of Hard Drive Energy Consumption:
in short version:

Also, used RAID defines the consumption growth:

enjoy
 
I have been pointing out for a long time that the best specification kind of doc for drives (HDD, SSD, NVMe) are made by Seagate. This cannot be compared with the absence of such data from WD.
I concur; Seagate spec data is generally v good...Ironwolf (standard or Pros) HDDs are generally my goto models these days, though I have recently bought some Toshiba N300s to try in my ZFS box. They're a fair bit cheaper than Seagate / WD in the UK.
Here is an old (2008) analysis, but still valid (principles point of view) - An Analysis of Hard Drive Energy Consumption:

in short version:
Interesting to note from this:
"The power consumption of reads and writes are noticeably asymmetric. For 1 GB of data, reads cost at least twice as much as writes in 75% of our tested drives".
I think most people would asume that writes consume more power than reads; I certainly did. I wonder if this is still the case with current modern HDDs?

Thx for posting
 
Here is also another parameter: WORKLOAD RATE.

And WD is as well unclear in this math:
Workload Rate is annualized (TB transferred ✕ (8760 / recorded power-on hours))

More clear math is about:
Workload Rate = (Lifetime Writes + Lifetime Reads) * (8,760 / Lifetime Power On Hours)
look at the Seagate explanation:
Just to be sure:
Read or Write operations from the host computer.


Then for the mentioned WD Red Pro 2TB, PN:WD2002FFSX , Workload Rate = 300TB
it means:
1664534013451.png

Description:
PoH = [Operation mode hours/day] x [Operation mode days] .... cumulated for each year in operation
Write = [Drive capacity] x [Capacity utilisation] x [utilised capacity overwritten per year] .... cumulated for each year in operation (Lifetime Writes)
Read = [Drive capacity] x [Capacity utilisation] x [utilised capacity reader] .... cumulated for each year in operation (Lifetime Reads)
Note for the Read operation: Every Scrubbing and every SMART check affects your Total Reads. If you run it once a week, you will read the used part of the disk 2 x 52 weeks a year. That's a total of 104 times read used disk capacity during the year and you haven't read any of the data you have there for your needs.
Transferred = (Lifetime Writes + Lifetime Reads) * (8,760 / Lifetime Power On Hours)

If we change the capacity of the drive from 2TB to 6TB and leave all other parameters, the story about PRO Workload will crash. The reason why I often read about drive damages:
1664535238977.png


However, if you use Enterprise level drives with the Workload rate = +550TB, the situation changes:
1664536460733.png


See that point?
Here is clear proof of why some HDDs have been running without problems for 10 years (my case).
I always wonder what the drive is (vendor and PN) and how it was operated and I need to see "smartctl -x". Otherwise, it's just a matter of feeling for me.
Or here is clear proof that underestimation during selected drive Workload and its overloading can cause damage too soon. Especially for tireless downloaders of video content from the Internet.
PRO label has not yet made anything that people imagine under it.

Another point
However, if we are talking about home NAS, then we are talking about the drive utilisation in minutes or max. tens of minutes. Because the biggest obstacle to such utilization is the CPU and RAM of such NAS. Not to mention the parallelization of other processes that properly overwhelm the CPU, another impact on the network speed.
This is the real reason why most NAS users cannot utilise their 1Gbps network. Even a non-performing HDD can have a write throughput of 150MB/s = 1200 Mbps = and this is 20% more than the maximum throughput of the 1G network. This is the biggest reason why you are experiencing transfers of around 50MB/s on your 1Gbps network.
Then why buy more powerful drives? The answer is the WORKLOAD RATE = longer lifespan.
People are very happy to succumb to inscriptions like "PRO". They often internally compare it to 24/7. Unfortunately, this is self-delusion.
If they were to compare the enterprise segment and the "PRO" or even the "NAS" HDD segment, they would find out that they get more WORKLOADS for the same cost from the Enterprise segment.
I've already written a lot about it here.

All this has an impact on the selection of HDD/SSD for the required type of operation. That is, also on costs.
-- post merged: --

"The power consumption of reads and writes are noticeably asymmetric. For 1 GB of data, reads cost at least twice as much as writes in 75% of our tested drives".
I think most people would asume that writes consume more power than reads; I certainly did. I wonder if this is still the case with current modern HDDs?
nothing fundamental has changed in the field of basic HW principles for PMR, i.e. CMR technology.
 
Last edited by a moderator:
Here is also another parameter: WORKLOAD RATE.

<snip>
it means:
View attachment 10852
<snip>
Transferred = (Lifetime Writes + Lifetime Reads) * (8,760 / Lifetime Power On Hours)

See that point?
I think you're missing a key phrase in the Synology definition of workrate: "
Annualized Workload Rate = (Lifetime Writes + Lifetime Reads) * (8,760 / Lifetime Power On Hours)

8,760 are the number of hours in a year. The Workload Rate becomes an annualized average expressed as TB/year."

ie, the 300TB workrate is an average, spread out over a year; your calculations for 10 or 5hrs/day are not yearly averages. They show the calculated 'workrate' for (eg) those 5hrs when it is reading/writing, but they then ignore the other 19hrs/day when the HDD is doing nothing. If you include those 19hrs day etc for a year then you get back to the annualised average workrate of eg 156 TB.

By specifying the Workrate as an annualised average, Seagate are implying that this is the most granular level of HDD activity that is significant. ie tehy dont care if you read+write your 300TB in 1 month and then do nothing for 11months, or spread it over 6 months or 12 months; the significant figure is the annual average.

In practice, if you were to make a warranty claim with Seagate, they would likely calculate the average workrate over the period of time the drive has been in use and use that figure.

Does this make sense?
 
Last edited:
It will be a long shot.

To answer @Fortran consideration, it is necessary to describe the issue in more detail. So, let's start with the basics of Workload Rate.

Already in 2013, Andrei Khurshudov from Seagate R&D described it nicely (still valid + I put some important notes in there to be more clear):

By definition, the primary function of HDDs is to store and retrieve data, storing hundreds of Gbits of data in every square inch of storage surface. They are capable of recording and retrieving data at sustained data rates on the order of 200MB/s or more (My note: over 260 MB/s in 2022).

In order to achieve this high recording density and high data throughput, magnetic read and write components are kept at physical separation of several nanometers (1nm = 0.001μm) from a fastmoving rotating media. This is a complex technical design task, necessitating that drives are designed, tested and classified for a specific work environment characterized by the range of usage time and customer workload, among other factors.

My note: Since 1983 we have seen a strong race for Areal density = RAW capacity per square inch of the platters in the HDD. This RAW capacity has a direct relation to net storage capacity of the HDD (storage capacity is always smaller than the RAW, based on many factors). Density also generally affects performance. Just imagine in 1953 the Areal density was just 2 000 bit/in2. Since then, the increase in density has matched Moore's Law in 2014, reaching 1 Tbit/in2. In 2015, Seagate introduced a hard drive with a density of 1.34 Tbit/in2. More here.


Workload is an engineering term used to define the amount of work stress the drive is exposed to during normal operation.

For example, Drive A could be reading and writing several GBs of data every day while another drive of the same design, Drive B, could be reading and writing several hundreds of GB of data per day. In this case, we would say that Drive B is operating under much higher workload stress.

My note: this is the point - for the work stress, it is not interesting how much the disk rested after work, but what demanding work it performed during work stress. It is the same as with people who can "smash their body" in the gym in an hour and then stay in "normal mode" the rest of the day. It is the immediate/sudden/short load that affects our organism. The more load in a shorter period of time, the more the body is "destroyed".

In order to get an idea of how much workload is too much, let’s review three typical scenarios (drives A, B and C):

Let’s consider a 4TB Seagate® Enterprise Capacity 3.5 HDD. This drive is capable of a sustained data transfer rate of about 175MB/s (My note: it was 2013). Let’s imagine three of these drives all operating in similar conditions (and assuming the same server):


  • The first drive (Drive A) is consistently transferring 5MB/s (or transferring an annual average of 158TB/year).(My note: conversion to TB in decimals)
  • while the second one (Drive B) is transferring 10MB/s (an average of 315TB/year).
  • Finally, the third drive (Drive C) is, in this example, transferring 100MB/s (an average of 3150TB/year).
It is easy to see from the above scenarios that Drive B is exposed to 2× higher workload stress than Drive A, and that Drive C has 20× higher workload stress than Drive A.

Assuming linear dependence, the next reasonable conclusions would be to presume that Drive B will have 2× higher failure rates than Drive A, and Drive C will have 20× higher failure rates of Drive A. However, Seagate data suggests that assumption of linear scaling of failure rate with workload is incorrect. Years of research and experimentation has allowed Seagate engineers to understand the complex effects of workload on drive reliability and to come to the following conclusions:


  • Every HDD type has some safe threshold of workload that is now defined as the workload rate limit (WRL).
  • As long as the workload doesn’t exceed the WRL, workload stress has very little to no impact on this product’s reliability and failure rate.
  • When the workload exceeds the WRL, the reliability of this product will begin to decline.
Therefore, it is very important to understand the workload stress of an actual data center and select HDDs accordingly. Table 1 gives a summary of Seagate recommendations for selecting the most appropriate drives for different data center environments.

Table 1. HDD Recommendations by Workload
1664719579304.png

My note: Re Table 1 - As you can see, the PMR technology (otherwise called CMR) has not changed from the point of view of principles during that entire period. What is valid in 2022 was also valid in 2013 (even much later) - HDDs that will be exposed to larger transfers (R+W) should meet the criteria for Workload +550TB/year (call it what you want, Workload Rate Limit, Annualized Workload Rate, or just Workload). The calculation formula is still the same.

Assuming that drives A, B and C were all nearline drives, we’d expect on average drives A and B to have very similar reliability (both workloads are below the WRL of 550TB/year). Drive C, on the other hand, with its average workload rate of 3150TB/year, will significantly exceed the recommended WRL for a nearline drive and will be exposed to higher risk of failure.

The table allows data center operators to select the right type of HDD for the right workload. Following the recommendations should ensure the highest possible reliability of HDDs used and lower the long-term TCO.

In Figure 1 (below), one can see that drives A and B belong to the same safe zone and have no failure acceleration due to contributions from workload. Alternatively, Drive C operates well outside the recommended WRLs and could show decline in reliability.

1664719621793.png



End of Khurshudov's explanation part.

----------------------------

So, as long as we are united in theory, we can move deeper.

Some inputs to the considerations:
  • only SAS HDDs are expected to have duty cycles of 24×7
  • whereas SATA (broader range from consumer to enterprise levels) drives are typically rated from 8×5 to 24x7 (exactly defined models)
  • Therefore, it is much more important to read the Workload rate than the "Pro" label. Because "Pro" is the marketing component of sales and Workload rate defines what drive we need for our purpose. Of course, a drive with a low Workload rate will also work. No doubt. However, you should expect your TCO to go up because those drives just can't handle it.
And now the answer to the @Fortran consideration that was the purpose of this long exercise:
  1. Workload e.g. 300TB/year is a figure spread over one year. It is based on the components used in the construction of this drive.
  2. If we want to define resistance from the point of view of workload stress, then we NEED CARE if we constantly transfer (R/W) the same data capacity (MB) to/from these HDD components during a short time (large instantaneous load) or during a long time (load distributed over time). Remember the story of the drive: A,B,C. Or my tables in the previous post.
  3. That is why we CAN'T and MUST NOT COUNT the time when the drive is out of that operation = then, these drive components aren't stressed = the drive durability is not reduced from the point of view of Workload at that time.
  4. The annual average mentioned by @Fortran, therefore, is not significant at all. You can significantly reduce the disk resistance even in a single year in case of a heavy Workload.
  5. In practice, if you were to claim an HDD (warranty), you should know how to download smartctldata (PoH, LBA W, LBA R) and calculate the Workload (previous post); they will (vendors) only use the calculation I mentioned. Do not confuse it with the data "Total Bytes Written" = a number of sectors written to the NAND media = valid only for SSD. And don't use the generally assigned SMART ID at all - it is essential to know how the manufacturer for a specific PN records these IDs. The differences are even at the firmware level of the same devices!!! Therefore, to get such ideas, the following conditions must be met:
    • The drive is not dead, and it is possible to read at least SMART data.
    • The drive recorded these three SMART IDs set (LBA W, LBA R) at the Firmware level. And that's not common (especially for WD, but I've seen Seagate HDDs that didn't have this). The PoH SMART ID is common.
Time for a coffee.
 

Create an account or login to comment

You must be a member in order to leave a comment

Create account

Create an account on our community. It's easy!

Log in

Already have an account? Log in here.

Similar threads

Oldest continuous running drive. LATE 1980’s PATA Quantum Quickview 330GB used in an LG 3410a OTA/cable...
Replies
0
Views
539
Thanks, I bought the same part from startech (or a very similar one) and with this unit the SSD boots on...
Replies
7
Views
2,500

Welcome to SynoForum.com!

SynoForum.com is an unofficial Synology forum for NAS owners and enthusiasts.

Registration is free, easy and fast!

Back
Top