Drobo

HD cache access

with different HD cache sizes, what is the measurable difference in read/ write times in the Drobo?
This was asked before in this post here, http://www.drobospace.com/forums/showthread.php?tid=923&pid=6660#pid6660, it is stated “no noticeable difference”.
I am geeky and want to know what the difference is. Not knowing another person’s reference level of noticeable, I would rather have data, vice observation.

Anyone have any input on this?

thanks

considering that they do quite extensive benchmarking with their products probably means that there was no statistical significance and any advantage/loss would be down to random fluctuation.

Yes, I am sure that DRI does extensive testing. Docchris, not trying to be dismissive, but I am not asking if they tested the product.
I am asking about REAL NUMBERS that either DRI or customers can reference.
The Drobo is not a hairdryer. I ask how hot it gets and reponse is that it won’t get to hot to burn my skin.
thanks

There was no measurable difference between the two speeds of drives when tested.

Jennifer, I know that you are attempting to be helpful, but please may I ask that you not answer with a non answer to my post.
And it would help if you read my post correctly. I am asking about cache size of a hard disk. Not speed!?
I do not appreciate it, and it does nothing to further the general knowledge.
Especially, when I created this thread because I was sorely disappointed by the response in the other thread. Thanks.
I am a paid customer and hoped to learn more about the Drobo and share knowledge with other users.

Respectfully,

yym

Umm… I don’t understand what you’re after if you’re not after speed.
Cache size doesn’t affect acoustics, and speed and acoustics are the only two measures that come to mind.

USB2 and Firewire connection speeds are both slower than a typical hard disk directly connected to your PC via SATA. More cache memory on a hard disk won’t make a difference since your can’t feed it data fast enough with your USB2/Firewire connected Drobo.

Of all the Drobo models the only one where hard disk cache may possibly make a difference is on the Drobo Elite model with its dual gigabit interfaces designed for access by multiple computers at the same time. Even then you will get better performance if you add more disks to spread the load over, rather than getting disks with lots of cache memory.

i agree, jennifer’s post seems quite clear, the size of cache on the hard disks makes no difference to drobo. what metric are you after if not that?

you wanted to know the difference between hard disks with large cache and small cache, and you asked for the difference in read /write times (your exact wording from your first post), and you want ni actual measurements not smoe hand waving vagueness, well you are in luck, here you go… the difference in read/write times is:

0

thanks for the serious reply.[hr]

You’re as bad as Jennifer.

yoyoma,

Respectfully, you’ve asking an impossible question to answer. Jennifer has already stated that the difference is not measurable. But you’ve asked them to measure something that they have said is not measurable.

Since you found ajspencer’s answer more appropriate, the better way to phrase your question would have been to ask “Is there any benefit to using a drive with larger cache in Drobo?”

What? Impossible question? I am flabbergasted at what you wrote.
I have asked them to measure something that is not measurable?
So are you saying that they cannot test the read/write access speed of a disk1(8MB cache) in the drobo. And that they cannot then test the read/write access speed of disk2(16MB cache) in the drobo. And that they then cannot compare the two and come up with a number?

Or are you saying that they CAN do that and DID do that the read/write speed for disk1 was identical to disk2?

And if you say that they CAN do this test, and DID do this test, then I am asking to know what those numbers were. It is interesting if they are identical, but please stop telling me that is identical and get on with posting the ACTUAL speeds. [or please point me to where that has already been posted on this board, thank you].

actually you asked for the difference. they said there was none. that was a perfectly clear answer to your original question.

if you want to know drobo’s speeds - they are on their website.

dri have just told you that those speeds are not affected by disk cache.

so with :

  1. the actual speeds published on the website
  2. the knowledge that those speeds do not change in relation to the size of the disk cache

what information do you feel you are missing?

So first you asked…

[quote=“yoyoma, post:1, topic:1102”]
with different HD cache sizes, what is the measurable difference in read/ write times in the Drobo?

I am geeky and want to know what the difference is.[/quote]
Then Jennifer answered…

[quote=“Jennifer, post:4, topic:1102”]
There was no measurable difference between the two speeds of drives when tested.[/quote]

At least we share the same feeling on something. :slight_smile:

[quote=“yoyoma, post:11, topic:1102”]
I have asked them to measure something that is not measurable?[/quote]
Based on Jennifer’s answer, they measured, and there is no difference.

If the difference is not measurable, the difference is, for all intents and purposes, zero. It’s like asking someone to measure the difference in length between two pieces of wood. They pull out a tape measure and get 360mm on one piece, and 360mm on the other piece. There is no measurable difference.

You are asking for what the difference is, and the difference is zero. Not close to zero, not almost zero, but zero. So your question should be answered but you have a new question, “What is the read/write speed of Drobo?”

Well, that one is tricky.

ajspencer has already explained that the drives themselves are not the limiting factor in Drobo read/write speed (save for possible DroboElite), so cache size doesn’t matter.

Effective (from the client side) Drobo read/write speed varies depending on the conditions - whether there is simultaneous access, what interface is being used, whether the data request is large or small, how full the drive is. However, given all other variables are the same, the cache size of the drive does not affect the speed.

So far, to recap, the questions and answers on this thread so far are:
[list=1]
[]Is there a difference in read/write speed between different drive cache sizes?
Answer: No. (Jennifer says “no measureable difference”)
[
]Does drive cache size affect Drobo performance?
Answer: No. (ajspencer says drive cache is not the limiting factor)
[*]Did DRI test performance of drives with different cache sizes?
Answer: Yes, otherwise they could not state that there is no measurable difference. (A difference of no difference still requires measurement)
[/list]

If, in Drobo world, f_8MBcache(a) = f_16MBcache(a)
Then for any a (being the same make/model drive), the result of f_8MBcache and f_16MBcache are equal.

Hahaha, Docchris types faster than I do. Or maybe he’s just more concise.

I am not looking forward to searching the “knowledge base” or the website in order to piece the information together.

Thank you both for your responses. Much better responses than Jennifer’s.

you do tend to waffle a bit :stuck_out_tongue: and i type horribly fast which coupled with my dyspraxia usually means its littered with typos so i have to spend 5 minutes going back and correcting them all! and even with all that, i still beat you :stuck_out_tongue:

and yoyo, click here:

http://www.drobo.com/products/drobo.php

then click on “view performance charts”

i found that in about 30 seconds flat

Docchris, you beat me to the punch, I was just about to post that.

:wink:

incidentally the site has changed a lot since i last looked at it, its much nicer (the products page with them listed in that nice table).

Since i’ve got my pro (and for a home user its very expensive) im not planning on upgrading my pro anytime soon so havent bothered checking the website

@Yoyoma, that is exactly why we have the Knowledge base and the website. So you can search for additional information on your free time.

Honestly, if you needed an immediate answer, you can always contact Tech Support. We are more than happy to answer any questions you may have concerning speed and performance of the drobo. We also probably would have directed to you the website link that Docchris posted as well.

I have seen the chart before and it is boilerplate to show difference between Gen 1 and Gen2. So I assume it is for optimum conditions. It is more informative than nothing, I suppose. I will understand that some of you feel the question answered. You claim that DRI has responded. So, if anyone reading this has any findings to share with their own experience, I would welcome that.

Your core question seemed to be whether to get an 8MB cache drive or a 16MB cache drive, and the answer is it doesn’t matter.

If you have another question, fire away! :slight_smile: