NETGEAR kicks loose Six-Bay SMB NASes

[claykin]

Bring on the Readynas Pro test Tim...

My apologies. I missed that in the Netgear spec.
I take back my comment about the QNAP TS509 PRO and Sans Digital MN4L+ being indicative of the ReadyNAS Pro performance. They use entirely different CPUs.

 

[beq]

I see thanks for the update.

beisser on the ReadyNAS forum helpfully mentioned that:

"the pro uses a Intel(R) Pentium(R) Dual CPU E2160 @ 1.80GHz"

I replied:

"Interesting that the Pro uses a desktop, dual-core Intel Pentium CPU (even if it is an older 65nm pre-Core 2 Duo part), whereas both the Thecus N7700 and QNAP TS-509 Pro use mobile single-core (65nm) Intel Celeron M CPUs (with the Synology DS508 trailing behind). Methinks the ReadyNAS Pro should be faster, even if it consumes more wattage (which I'm not that worried about)."

Anyways, I believe many of the ReadyNAS Pro beta testers like him have had to resort to benchmarking the Pro from RAM disks because their mechanical HDDs were too slow...

 

[beisser]

well that or you resort to benchmarks which dont use the local harddisk as source. some of those benchmarks are iometer, atto disk benchmark and copying stuff from a ramdisk to the pro.

i dont know if iozone uses the local harddisk as source, but i hope it doesnt otherwise the benchmark-results would be flawed

 

[beisser]

I see thanks for the update.

beisser on the ReadyNAS forum helpfully mentioned that:

"the pro uses a Intel(R) Pentium(R) Dual CPU E2160 @ 1.80GHz"

I replied:

"Interesting that the Pro uses a desktop, dual-core Intel Pentium CPU (even if it is an older 65nm pre-Core 2 Duo part), whereas both the Thecus N7700 and QNAP TS-509 Pro use mobile single-core (65nm) Intel Celeron M CPUs (with the Synology DS508 trailing behind). Methinks the ReadyNAS Pro should be faster, even if it consumes more wattage (which I'm not that worried about)."

Anyways, I believe many of the ReadyNAS Pro beta testers like him have had to resort to benchmarking the Pro from RAM disks because their mechanical HDDs were too slow...

the pentium dualcore e2160 has a conroe-core. it was actually released after the core2 duo. the only difference is slower fsb and less L2 cache.

see http://www.techspot.com/review/53-pentium-e2140-e2160/

 

[drbiggly]

I see the results of the ReadyNAS Pro has placed it pretty high in the speed category. Awesome is all I have to say; I think that the X-RAID/X-RAID2 is really the type of technology that is going to be the ease of use that most folks who need something at home that is expandable would want to go for.

Once they can certify the 1.5TB Seagate drives for them, that makes for a very formidable amount of space.

I'm anxious to see the 'Pioneer' (diskless) model specs and what works with it. That one is the one that interests me the most.

-Biggly

 

[corndog]

Yes, good to see it ranking so high. I have to admit though, that with all of Netgear's claims about "we couldn't work this NAS fast enough to find it's limits" it didn't manage to clearly beat the QNAP TS-509Pro. I was thinking it would.

The Infrant, with X-RAID is definitely a level of ease and reliability higher than most of the competitors out there, but their products were getting to be a lot slower in comparison, so it is good to see them coming back to the head of the charts.

One caveat though - I remember reading in their forums that there are some limitations in the "Pioneer" (diskless) model - some functions or capabilities that are not supported. So I would strongly suggest you spend the effort to read through their pre-sales forum before doing your purchase, just to make sure.

 

[thiggins]

A couple of points to note:

- Remember that the iozone test machine is a 2.4GHz P4 machine with a PCI gigabit NIC running Win XP2. So it's not a current generation machine. But it is the same machine used to test the Qnap and all other products currently in the charts. So while the results won't represent the maximum performance of some of the products, the results can be fairly compared.

That said, corndog, I was also surprised that the ReadyNAS Pro didn't rank closer to the Qnap. Note also that jumbo frames don't provide any boost.

- Netgear has updated a comparison matrix of the ReadyNAS Pro Business and "Pioneer" models. You can download the PDF here.

- "Pioneer" pricing hasn't been officially announced, but NETGEAR said that I could tell you that it will be $1950 MSRP, with street around $1200. Still not cheap. I think we'll be lucky to see it by the end of the year, since NETGEAR can apparently sell all the Pro Businesses that they can make right now.

 

[chirpa]

I have to admit though, that with all of Netgear's claims about "we couldn't work this NAS fast enough to find it's limits" it didn't manage to clearly beat the QNAP TS-509Pro. I was thinking it would.

Well I'd like to see the two NASes compared on a recent decent system. I can get 102MB/s on the ReadyNAS Pro. Benchmarks I see around the web show 55-60MB/s on the QNAP. Including the local review here "The 509 Pro has exceptionally flat performance out to 512 MB filesizes, with speeds exceeding 55 MB/s."

According to http://www.smallnetbuilder.com/content/view/24310/77/, tests here are done using a P4 2.4 with 504MB (shared video), with 80GB 2MB cache, which is a sad test PC. Even the $300 Dell entry level PCs would blow that one away.

 

[thiggins]

Well I'd like to see the two NASes compared on a recent decent system. I can get 102MB/s on the ReadyNAS Pro. Benchmarks I see around the web show 55-60MB/s on the QNAP. Including the local review here "The 509 Pro has exceptionally flat performance out to 512 MB filesizes, with speeds exceeding 55 MB/s."

According to http://www.smallnetbuilder.com/content/view/24310/77/, tests here are done using a P4 2.4 with 504MB (shared video), with 80GB 2MB cache, which is a sad test PC. Even the $300 Dell entry level PCs would blow that one away.

Yes. It's an old PC. But, until recently, more than adequate for testing the NASes that were being produced.

I will be changing the testbed at some point soon. But then most of the NAS Chart data will become not directly comparable.

What is your test setup for your "102MB/s" test, i.e. what filesizes, # of drives, mode and especially is that read or write?

 

[beisser]

if thiggins could get me his test-specs (im a complete noob ob iozone and dont have much of a clue on how to use it yet) i could run a benchmark on a q6600@3ghz with onboard nic, 4gb ram and a wd raptor system disk to compare it to the readynas pro benchmark on that not so up to date system used here.

this should provide at least testresults that can be compared.

 

[beisser]

i can get 95 mb/sec using iometer with the profile supplied by infrant/netgear. the only change i needed to make is to adjust the filesize to be bigger than the ram of the nas

i can get 67 MB/sec running bonnie++ (linux benchmark) on a nfs-share using a 9gb testfile.

here is a snippet to compare:

root@beisser-laptop:/mnt/test# bonnie++ -s 9000 -u 1111 -g1111
Using uid:1111, gid:1111.
Writing with putc()...done
Writing intelligently...done
Rewriting...done
Reading with getc()...done
done
start 'em...done...done...done...
Create files in sequential order...done.
Stat files in sequential order...done.
Delete files in sequential order...done.
Create files in random order...done.
Stat files in random order...done.
Delete files in random order...done.
Version 1.03b ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
beisser-lapto 9000M 41465 99 69055 14 33283 12 29456 78 63843 13 332.2 1
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 1815 11 4066 15 1885 10 1839 9 4604 13 1897 8
beisser-laptop,9000M,41465,99,69055,14,33283,12,29456,78,63843,13,332.2,1,16,1815,11,4066,15,1885,10,1839,9,4604,13,1897,8
root@beisser-laptop:/mnt/test#

edit: bah.. quoting this makes it hard to read, but i think you will see what i mean

 

[beq]

I can appreciate Tim's dillema. It would be a monumental task to reacquire and re-test all of the other leading NASes on the market on a faster testbed (though this is what I'm very interested in).

beisser, any chance you can also get a hold of (any of) the QNAP TS-509 Pro, Synology DS508, Thecus N7700, etc?

Perhaps the Infrant/Netgear staff already have these competitor products in their lab just to see how the ReadyNAS Pro fares, but unfortunately any comparisons from them won't exactly have the appearance of objectivity. What say you, chirpa?

P.S. I'm also curious though why the ReadyNAS Pro performance curves are on average beaten by the likes of the QNAP TS-509 Pro on the existing SmallNetBuilder testbed?

 

[beq]

Oh thanks for the pricing information on the ReadyNAS Pro 'Pioneer'. $1200 expected street pricing is a lot lower than the $1950 MSRP, which is good news and should make it more comparable to the Thecus 7-bay and not too much more than the QNAP/Synology 5-bay ($800-1000).

Edit: I didn't see that Infrant/Netgear had updated the comparison feature list for the ReadyNAS Pro Pioneer edition. It now seems to have all the features of the Business edition, except snapshotting, Active Directory, and VLAN.

 

[thiggins]

if thiggins could get me his test-specs (im a complete noob ob iozone and dont have much of a clue on how to use it yet) i could run a benchmark on a q6600@3ghz with onboard nic, 4gb ram and a wd raptor system disk to compare it to the readynas pro benchmark on that not so up to date system used here.

Here is the iozone command line (all on one line)
iozone -Rab readynaspro_pti_r0_3drv_1g_4_2_1.wks -i 0 -i 1 -+u -f i:\temp.tmp -y 64k -q 64k -n 64k -g 4G -z

You need to map a share with read/write privs as the "I" drive.
Then open a Command Prompt window, cd to the iozone program directory and enter the command. You can name the readynaspro_pti_r0_3drv_1g_4_2_1.wks output file name to whatever you like.

I am running a number of tests with the Core2Duo system used for testing in the FastNAS series and will have that data as part of the review. I am not seeing 100MB/s speeds, even with Vista SP1. Best case numbers are around 70MB/s reads (up until 1GB RAM size), then falling down to 40MB/s
and about 67 MB/s writes from 256MB filesize up.

 

[thiggins]

thanks for the data beisser. Can you simplify it for a simple mind like mine:

What is the filesize tested and results for write and read speed?

 

[thiggins]

I can appreciate Tim's dilemma. It would be a monumental task to reacquire and re-test all of the other leading NASes on the market on a faster testbed (though this is what I'm very interested in).

Frankly, that's simply not going to happen. Either I will start another set of charts or fold the data into the existing ones, using a different bar color.

At any rate, it's only these really expensive NASes that are pushing the envelope right now.

One difference between the Qnap TS-509 Pro and ReadyNAS Pro is number of drives, which Netgear is indicating makes a difference, at least for RAID 0.
There are only 3 drives in the Pro I tested, but Netgear sent 3 additional drives. I will be loading them up and rerunning RAID 0 with 5 (same as TS509 Pro, then 6 drives. Maybe that will make a difference. But I have yet to see # of drives make a difference in RAID 0 results n my Fast NAS series testing.

 

[yoh-dah]

Tim, I would just concentrate on RAID 5. No one should be using RAID 0, especially in a 6-disk system. RAID 5 also is more stressing on the NAS, and that should definitely be taken into consideration.

 

[beisser]

thanks for the data beisser. Can you simplify it for a simple mind like mine:

What is the filesize tested and results for write and read speed?

with iometer the filesize i used was 6gb (my nas has 3gb ram). with bonnie++ i used a 9000megaqbyte testfile. bonnie++ is the only unixbased diskbenchmark im aware of so i used that when testing via nfs. if anyone has a better linux-diskbench let me know

 

[beisser]

btw tim, all my tests were run using x-raid2

and yoh-dah is right.. whoever uses raid0 on a 6 disk system deserves to have one of the drives die

 

[thiggins]

Tim, I would just concentrate on RAID 5. No one should be using RAID 0, especially in a 6-disk system. RAID 5 also is more stressing on the NAS, and that should definitely be taken into consideration.

I will do both. RAID 0 should provide the highest throughput, shouldn't it?