Running NAS Performance Tester Vers 1.7 on default setting (400MB with 5 loops) from my WIN 7 64bit PC over wireless gives these very slow results. Ave Write in MBps 5.65, 4.22, 5.33 and 4.46. Ave Read in MBps 2.28, 2.85, 3.16 and 2.38.
I know the wireless is slow, BUT I cannot understand why the Read is significantly slower than the write consistently. My RN516 is setup as 6 WD 4 TB disks in RAID10 configuration, giving 10.9 TiB available.
The connection is made with a “Netgear WINDA3100v2 N600 Wireless Dual Band USB Adapter” plugged into a USB 3.0 port on the PC to a “Netgear N900 dual band WNDR4500 Router”. The ReadyNAS RN516 is directly connected to the router, and all firmware is latest on all 3 devices.
Yet over 3 days and 4 separate tests the read speed is typically only 54% of the write, and in the best case 68%. Can anyone please explain this?
Would same results came up whether the wireless network is either secured or unsecured? Is it possible that you run the test using a different computer with either a built-in wireless adapter or using the same WNDA3100v2?
The network is secure with "WPA2-PSK[AES]". The Router is set for Max of 415Mbps on the 2.4GHz which is the connection 99% of the time.
InSSIDer shows that (currently) my network is a Max of 216Mbps. Netgear Genie shows that (currently) the PC is communicating at 104Mbps (was 144Mbps a little while ago... probably the fastest I have seen). Very occasionally it will drop down to 59Mbps, but mostly sits at 78Mbps, 104Mbps or 117Mbps.
I have about 20 networks locally, of which 5 others are on the same channel and Win 7 indicates mine between 50% - 70% signal (using Network Meter). More than 1 other network has a stronger signal than mine (I say that since it MAY be significant).
I am very reluctant to even temporarily make the network unsecured, but might try a variety of other options. (I assume that you wish to eliminate the NAS as the constraining factor).
However having said all the above, my problem is slow reads as apposed to writes on the NAS. (I accept that my wireless speed is not the best around by any means.)
I believe you have already thought of this.... What if you try to run the same test again using the same PC but it should be directly connected to the WNDR4500 Router where the NAS is also connected to? - Lets see if it will make a difference on the results...
Although I would like to do this, it is NOT an easy thing to do (almost impossible). The PC (a free standing medium tower) is 5 rooms away and difficult to move, and the Router and NAS is in the garage room with very little space. I will try tomorrow with an older PC upstairs (my MAC users - daughters will be unavailable and their MACs have passwords), but that is Vista and may be difficult to achieve. I also have 2 other machines, which are XP laptops and I am not certain they do USB 3. Actually USB 3 or 2 shouldn't make much difference, I suppose. I WILL try something tomorrow.
My Cable Modem (ISP connection) is also in the garage, along with the ReadyNAS, router, a UPS and other bits and pieces.
Spent the weekend setting up a Vista laptop (old machine) with direct connection to the router that has direct connection to the NAS. Using the same NASTester software the results were somewhat staggering, although NOT unexpected.
Running NAS Performance Tester Vers 1.7 on default setting (400MB with 5 loops) from my Vista gave this result. Ave Write in MBps 11.84. (All 5 iterations returned 11.84) .Ave Read in MBps 11.82 (Very consistent at 11.83, 11.83, 11.78, 11.83 and 11.84)
I imagine the write to the NAS is cached and the constraining item is the cables and connection protocols with Max 100Mbps (around 12MBps). The absolutely minor difference is in the overheads. The reads vary a little, and the dip in the 3rd iteration is a minor blip, so not to worry about it. From this, I read that the ReadyNAS is performing as well as can be expected with the setup I have.
Obviously it would be nice to improve the WiFi from front room to garage, which is the culprit here (I knew this was slow anyway). Although this does not explain why the read speed is almost half of that obtainable on the writes. Does WiFi have different speeds in the 2 directions? If not there is still a mystery and the title should be "Read much slower than Write over WiFi from Win 7 PC to ReadyNAS machine."
Any comments welcome.
P.S. I cannot get the Vista machine working with WiFi on my local network, otherwise I would 'test' using that where the Win 7 PC is situated.
WiFi network interference may be your biggest issue which will result in unpredictable performance. Not sure how you selected your current channel. You may want to set your router's channel manually to either channel 1, 6 or 11. Pick the channel that has the least amount of interference from other WiFi networks (fewest and weakest other networks). Also consider chatting to your neighbours to segregate the various networks onto non-overlapping channels. Here are some useful articles on the subject: http://www.smallnetbuilder.com/wireless ... d-wireless http://www.smallnetbuilder.com/wireless ... less-crowd
Not sure why you can't get wifi working with the vista laptop.
(a) Are you using 5 ghz or 2.4 ghz wifi? 5 ghz has less interference, but doesn't travel as far. Not sure which is best for your setup. (b) wifi interference and multipath are factors, and you probably should try changing channels.
Though I think you should look into getting some cat 6 cable installed in your home.
#rovingmedic & #StephenB Thanks for the info guys - Unfortunately I have been living with the 'crap' WiFi for over 10 years, and YES it would be nice to have some cat 6 cable installed. Difficulty with walls and floors etc. I have been through 5 or 6 different Routers, as technology improves, changing the AP hardware as well. The major problem is the sheer number of networks out there in this area (maybe everywhere), not all on 1, 6 and 11. Every 6 months or so, I investigate and have already changed channels many many times.
The links provided are similar to what I have read before (but thanks) , but I was prompted to think about one thing. Use an old Router as well as the current and split the network into 2 (11g and 11n). My other thought with an old router was to use it as an AP instead of my current AP hardware on the assumption that the signal would be stronger. I don't think I have seen anything on signal strengths of AP and Router. So still some avenues to explore.
I already have a dual channel Netgear extender (to reach the loft room), so use 4 networks myself (usually 3 as I turn off 1 of the 5GHz). Hence I am reluctant to buy another extender, just for the one poorly served part of the house (no it is NOT a big house)
P.S. I have 4 or 5 'obsolete' Netgear routers and a Netgear printer server that didn't work to choose from to use as an AP.
I use powerline equipment to extend my "wired" network around my concrete home. I'm using Devolo DLAN gear but there are other providers as well (TP-Link, Netgear, D-Link, etc). Smallnetbuilder has a number of reviews and comparisons to help choosing a system. My experience has been that AV200 is quite stable. Reports are that AV500 is also stable and offers a substantial increase in bandwidth. I have had some stability issues with the new AV1200 gear. Reports point to overheating of the hardware being a factor.
I know you may not want to go out and buy more hardware, but this approach does get around the issue of a crowded WiFi spectrum and has the potential to offer increased bandwidth over 11n. Powerline gear comes in 2 main flavours: wired only and WiFi with wired. To control cost, you could go for wired only gear and reuse your AP capable hardware.
Good luck with improving your network performance. Cheers, Shaun
Your old router will not have more signal strength than the new one. The maximum transmit power is regulated by country (and in your case is 100 mw). The old router might have less power, but it will not have more.
Powerline is an option - it can be hit or miss (depending on your power circuits), and (like wifi) your throughput is always much less than the link speed. However, in many situations it works well. Get at least AV500. A WiFi extender is another possibility - especially if it enables 5 ghz, which is much less crowded.
On the read/write question - anything you write can be queued up at every step in the network path. That means in Windows, your USB adapter, the router,and the NAS. The NAS itself can do read-ahead when you are doing the read test, but it cannot send any data until the read request reaches it. So the queuing is asymmetric, and favors the write path in cases where the wifi network is misbehaving.
FWIW, as far as the NAS is concerned the writes take longer (since the raid parity blocks need to be rewritten).
#rovingmedic #StephenB Thanks for your suggestions and replies (again)
StephenB wrote:
Your old router will not have more signal strength than the new one. The maximum transmit power is regulated by country (and in your case is 100 mw). The old router might have less power, but it will not have more.
To be clear I was not suggesting replace Router with Router, but the WNDA3100v2 access point with a Router switched to be an AP only. If AP transmitted power is less than 100mw this might be an option.
Re the use of powerlines, I considered these years ago, but I have many socket extensions and multiple fuse boxes (5) in the house. I also have at least 5 suppression extension sockets for IT equipment. I am led to believe that all of these MAY 'dilute' and/or 'interfere' with any signal on the mains. So I currently think it is not a goer.
StephenB wrote:
FWIW, as far as the NAS is concerned the writes take longer (since the raid parity blocks need to be rewritten).
Are there raid parity blocks on RAID10 as I thought it was mirrored with both mirrors being written in parallel?
To be clear I was not suggesting replace Router with Router, but the WNDA3100v2 access point with a Router switched to be an AP only.
In other words, a wifi extender. Most new netgear routers include an extender/bridging mode, you can also buy devices that only work as extenders. Either way you will have better antennas and probably more power than your dongle.
BaJohn wrote:
Re the use of powerlines, I considered these years ago, but I have many socket extensions and multiple fuse boxes (5) in the house. I also have at least 5 suppression extension sockets for IT equipment. I am led to believe that all of these MAY 'dilute' and/or 'interfere' with any signal on the mains. So I currently think it is not a goer.
I think you need to try something, though it might have some risk of failure. Multiple fuse boxes are an issue, ideally the two adapters would be connected to the same fusebox. Then connect the AP to the in-house adapter. The IT sockets won't be a problem if you aren't connecting the powerline adapters into them. Powerline noise can be a problem, so you should look for an adapter pair that has built-in powerline filters. Maybe try this: http://www.netgear.co.uk/home/products/ ... P1200.aspx
BaJohn wrote:
Are there raid parity blocks on RAID10 as I thought it was mirrored with both mirrors being written in parallel?
It is mirrored, I forgot you were running raid-10.
I have been working on this problem since late Feb. I already have an extender elsewhere in the house and do not want another one cluttering up the airwaves. We have about 30 networks (on 2.4 GHz) that I can see from my front room, for basically 3 channels (1,6,11). The 5GHz is also used as well.
The P1200 looks great, but seems to not be available in the UK. Not certain if it sensible to buy US version for UK. I'm sort of hoping it will be available here soon 🙂
I have made some progress in that I now have the read slightly faster than the write, which has improved a little. This was achieved by changing channels on the 5GHz band (not quite so crowded) from 140 to 100. I checked and double checked and yes the effect is consistent, which is great news.
There is just two more things I would like the answer to:- 1. Does a 'g' connected device on the 2.4GHz channel degrade a 'n' connected device on the 5GHz channel on the same router, in this case a Netgear WNDR4500? 2. If it does, or can do, how can I monitor which of the 'g' or 'n' is each device using dynamically?
The 2.4 ghz and 5 ghz bands are completely independent, and can run different protocols. And associating a 802.11g client will not degrade performance for 802.11n devices even in the 2.4 ghz band.
For powerline, you want to get the UK version, so you'd want to wait until that is available in your region.
