Network Router and Adapter Testing Methodology
Before we dive right into the router and adapter performance, we need to discuss a testing methodology. Being that this is one of the first reviews I’ve done for networking gear, I needed to set up my own ‘test bed’ for routers and network adapters. Thanks to the great people over at Small Net Builder and their excellent “How to Test a Wireless Router” article, I had some good foundations to build from so I could get accurate and consistent results across multiple tests.
To take as many variables out of the equation and maintain consistency across multiple tests I picked up two identical Dell Optiplex machines off of eBay to use as my “WAN” and “LAN” networks. The machine specs for the two machines are:
- Core 2 Duo E6600 CPU (2.4 GHz)
- 2 GB (2 x 1 GB) Kingston HyperX DDR2 800 (PC2 6400) SDRAM
- Western Digital WD1600AAJS 160 GB HD (8 MB Cache/7200 RPM)
- Trendnet TEG-PCITXRL Low Profile Gigabit PCI Adapter
These are both hooked up to a basic Dell keyboard, mouse and 20” LCD using a Syba 2-Port USB KVM Switch. When needed, I used 2 foot CAT 6 network cables from monoprice.com.
The Optiplex machines certainly aren’t the most powerful machines around, but to test network connectivity you really don’t need the Dream System from the PCPer Leaderboard. I just wanted a pair of machines that were identical to ensure consistent results from test to test.
To give us a baseline for both wired and wireless connectivity, we’re using one of the better routers out there, the Apple Airport Extreme. This little box from Apple has Gigabit WAN/LAN ports, 802.11 a/b/g/n wireless and supposedly ‘easy’ setup. I say supposedly because I had to fight with the Airport software to do some of the ‘advanced’ configuration I wanted to do such as setting up different subnets on the LAN/WAN ports, but that’s a story for another day.
For network adapter testing, there’s an open PCI Express slot and some USB ports in each of the Optiplex machines that are available. Since most USB adapters can also easily be used on a laptop, I’ll do testing of USB adapters using a laptop. The added benefit of using a laptop for USB adapter testing is that I’ll be able to test from a few other locations to simulate some real use test cases. For this testing I’ve drummed my old MSI laptop back into service.
- Core 2 Duo P7350 (2.0 GHz) CPU
- 4 GB (2 x 2 GB) Corsair DDR2/800 SODIMM RAM
- Kingston SSDNow V+ 100 96 GB SSD (SVP100S2/96G)
- ATI Mobility Radeon HD4670 (512 MB VRam)
- Intel Centrino Ultimate N 802.11n (draft) Wi-Fi Adapter (633ANHMW)
The MSI EX625 originally came with a general built in 802.11 b/g/n adapter, but I upgraded the internal Wi-Fi adapter to the Centrino Ultimate N as it’s one of the better like one of the better mini PCI Express adapter on the market. With ‘3×3’ triple streams and advertised speeds up to 450 Mbps it’ll be perfect to use as a baseline to test router Wi-Fi speeds as well as to compare to other Wi-Fi adapters.
Though hardware is only half the equation. We still need a way to actually measure network throughput and other stats of the routers and adapters so we can compare and contrast them against each other with some sort of consistency. There are simple online speed, but there’s just too much variance if we use that. For consistency sake, the machines will not be connected to the Internet at all and only networked amongst themselves. We’ll then use three pieces of software wrapped up in a script to give us a good picture of network performance.
First up is a simple DOS based tool contributed by Matrix21 over on Smallnetbuilder.com, aptly named, “Max_session_tool.” The Max Session Tool will open up UDP connections between two hosts until the receiving end starts to drop the packets giving us the maximum number of network sessions that the endpoints and network hardware can handle. Since the router is the primary keeper of the sessions, we’ll only run this test for routers.
Next up is the simple ping test. Most people will be familiar with using the ping command to check connections between two machines. With this simple tool we can check the minimum, maximum and average time needed to send a packet of data between two machines, as well as give us a total number of dropped packets. The ping test will give us a good way to check for latency or for any other network issues.
Finally are a pair of great tools from Pete over at Totusoft. We’re using LAN Speed Test (v3) paired with the LST Server software. While there is a free ‘Lite’ version of the LAN Speed Test available, at only $5 each, both LAN Speed Test and LST server are worth every penny and then some. LAN Speed Test will run a battery of file upload/downloads against a file share on another machine to give you Average, Minimum and Maximum Upload/Download speeds. Adding in the LST server let’s you upload/download the data into the target system’s memory as opposed to hard disk to remove any impact hard disk speeds/cache would have on the actual speed values. I can’t thank Pete enough as he happily answered some of my questions and even took some suggestions I had and added them into the applications. For anyone looking to do any speed testing, I highly suggest checking out Totusoft’s offerings.
With these three tools we can get a pretty good snapshot of network connectivity and throughput. When I first started working through the tests, it was a lot of manual work, so make life easier, I strung together a batch file to automate each set of tests.
The batch file does the following:
- For routers, run the Max Sessions Tool 7 separate times and document the maximum sessions allowed.
- Run 7 tests of 100 pings from client to target computer and document the Minimum, Maximum, Average ping time as well as any dropped packets.
- Run 7 upload and download tests using LAN Speed Test with 10 MB files, 10 times for each file. Document Average, Minimum and Maximum Upload and Download speeds in Mbps.
Once we have all the data, it’s added to a master spreadsheet. To try to take any wild values out of the mix, we drop the highest and lowest values from the test results and then average the remaining ones for our ‘final’ test value for each metric.
To add one final piece to the puzzle, we will test Wi-Fi connectivity from 3 locations that we will call for simplicity’s sake, ‘Lab Test’, ‘Real World Test’, and ‘Torture Test’. In the ‘Lab Test’ I’ve set up both the router and wi-fi adapter in my office and approximately 3 feet from each other. For the second ‘Real World Test’ the signal will have to travel half way across the house through two walls and one floor into the Living room about 45 feet away. Finally, we’ll wrap things up with a ‘Torture Test’ where the wireless signal has to travel 70 feet into my Laundry Room and punch through four walls, a floor, and everything in between. I don’t know what the builders lined the walls of the Laundry room with, but not only does it lay the smack down on wireless signals, but I think I can use it as a nuclear bunker.
So, without further ado, let’s dive into some actual results from our wired and wireless tests.
Chris, nice review. Looks
Chris, nice review. Looks like a pretty nice router. I may have to pick one up to play with.
I had a few questions you might be able to answer:
– Can you lock down the admin page to only work over wired? I see a place to open WAN access to it but not WIRED/WIRELESS.
– Were you able to try DYNDNS?
– Any tests on parental control / keyword filters? I only ask because I know a number of routers in the past flat out didn’t work. I know a lot of parents that struggle with this specific point 🙂
The Network Services Filter looks like an interesting feature but looks like you have to custom create a rule if you want to do it by machine (IP).
I did kind of laugh at the “Famous Game List” on the port forwarding. Kind of surprised the Famous Server List didn’t include IMAP or HTTPS. Not a huge deal.
I did notice on their demo that the QOS is system-wide so you can’t select QOS rules but device which kind of stinks. I guess a bandwidth limit tool like on Tomato would be the better way to deal with that possibly.
Nice job. Look forward to more networking stuff on PCPer.
Hey thanks! Glad you enjoyed
Hey thanks! Glad you enjoyed it. To answer your questions, I don't recall seeing any way to lock the router admin down by Wi-Fi, just by WAN.
As for DynDNS and parental controls, I didn't actually get to dive too deep into them. Due to the huge amount of tests and time constraints we couldn't dive into every available firmware option. I do plan in future reviews that are for single adapters or routers I'll try to dig a little bit deeper into software/firmware options.
I use Tomato at home as well and love their QoS and bandwidth monitoring capabilities, hopefully more vendors start expanding out on those.
I have the ASUS RT-N66U
It
I have the ASUS RT-N66U
It does support DYNDNS & DNS-O-Matic which is fantastic. With DNS-O-Matic you can have the RT-N66U update OpenDNS and any other DynDNS provider. DNS-O-Matic acts as a DynDNS forwarder in a sense.
You can see the interface here: http://event.asus.com/2012/nw/dummy_ui/en/Advanced_ASUSDDNS_Content.html
Whаt’s up, this weekend is
Whаt’s up, this weekend is good designed for me, since this occasion i am reading this great informative paragraph here at my residence.
My web page … meditation classes Vancouver
this will make a great
this will make a great addition to the site. Please test the WNDR4700 and if possible contact netgear and tell them to add idle spindown to the drives, I really want get one but am not sure since many storage routers in the past (cough… apple), have been so full of fail that it has given storage routers a bad image.
Also if possible cover info on 3rd party firmware (would like to avoid routers that have no hope of 3rd party firmware).
I think we have a Netgear or
I think we have a Netgear or two waiting to be testing, I just don't remember what model. I'll check when I get back next week and if we don't have that I'll see if we can get one from Netgear.
Good idea on the 3rd party firmware. I don't know if time would permit us to test a bunch of 3rd party firmware on routers, but I'll try to mention if it's an option on each router going forward. Doing a bit of googling, it looks like the RT-N66U can take some versions of Tomato, but I'm not sure it'd be worth it since you'd probably lose all the other functionality that ASUS has built in. Will really depend on your use case.
Nice review! And welcome
Nice review! And welcome aboard Chris! I really like the ASUS RT-N66U’s style and performance, definitely considering replacing my current router.
Still disappointed with the data throughput on wifi compared to the advertised speeds, unfortunately it seems like common practice these days. I’m curious to see what the test environment and setup they use to get these benchmarks in the first place.
But as you said, gotta love that marketing. Big numbers = £££ apparently
That has pretty much always
That has pretty much always been a problem with wifi and it gets worst each year, with the latest standard 802.11c they will claim 1.2 Gbit/s to 1.6 Gbit/s, but in reality, benchmark at around 170 mbit/s to 220 mbit/s.
Since 802.11b, the delta between advertised speed and actual speed has steadly grown, in the past you could get around 60% advertised speed as a best case, and now it is down to around 25% advertised speeds
Yeah, this is kinda a pet
Yeah, this is kinda a pet peeve of mine honestly. These advertised speeds remind me of the claims from the old broadband providers back in the day. I mean, my car can travel 1000 mph too, assuming I drop it from low earth orbit, but I'll never see those speeds.
I'm not really sure how they come up with the 300/450 Mbps 'up to' speeds when I couldn't even get close to half that with the router and adapter sitting right next to each other. I think I'm going to see that with every vendor though.
I have to admit some
I have to admit some responsibility here… I was the Sr. Product Manager for the USR x2 consumer product line back in the 90’s and the whole “56k” thing kind of started this.. 🙂
As you said though, if you happened to live at the head end, and plugged your modem directly into the head end modem, you might be able to get 56k so who was I to say you didn’t have that setup? 😉
I do think for this crowd, focusing on the Tomato, DD-WRT and other “enthusiast” class stuff is a good fit. I could really care less about the latest EA4500/6500 from Cisco. I can read 10 reviews about those one from the candy-store reviewers out there.
I’ve been running Tomato for at least 5-6+ years. Originally on the WRT-54G then the 54GL. I have it on a Netgear 3500L right now. I’ve also been playing around with M0n0wall (on a WatchGuard Firewall), Untangle (on a PC), DD-WRT, OpenWRT, PFSense and a few others.. Still searching for a solution that fits my needs (3 kids, 10 computers, 2 servers, 2 rokus, 2 apple TVs, 2 google TVs, and more).. QOS, Bandwidth limits, parental control (that work), bandwidth monitoring, networking monitoring are all things I’m interested in.
Network Monitoring in particular is becoming more and more of an issue. With every device online and wanting to “phone home”, I want a better idea of what is going on and who is talking to what. That may not be something average users want but done correctly, a powerful router with the right features would be a huge seller. Most of the stuff on the market has just sucked for the past 10 years with the only improvements really being new WiFi standards.
With the home network being the backbone for a huge portion of our life now, its crazy that more innovation hasn’t come out in this space. If you let Comacast (intentionally spelled), AT&T and other ISPs have their way, they’ll reach their greasy fingers all the way into the home and charge and monitor everything. No thanks. You can plug in your data pipe to my outside wiring and go away..
I think when they make claims
I think when they make claims such as 450mbit/s they mean, provided you can harness the power of magic while transferring your data, then you will reach 450mbit/s
I don’t even know how they get away with it, many companies (eg small net builder) have tried to set up best possible case scenarios for a wifi speed test (including disabling as many things as possibly that can lead to a bandwidth or processing overhead, and still never manage to come even close to the advertised speeds). Isn’t it false advertising if advertised results are impossible to to achieve is real world or even laboratory conditions?
Imagine if wifi style claims were accepted in other ways of life.
Buy our large 18 inch pizza!for $12:(opens box and it only has one 4 inch slice of pizza)
Buy our overpriced high speed internet service: (ends up being comcast, which is Latin for authentic 9600 baud experience)
How would you like a 70 inch OLED 4K TV. (buys it and wonders how they were able to fit a 70 inch tv in a 20 inch box, you then open it to find another box containing a 10 inch 100ms response time TFT LCD display)
Yes Razor it is however
Yes Razor it is however companies have done it for a long time. WAY back when, the whole monitor industry was taken to court over “Viewable” area. Claims that a monitor was 15″ when it had 14 inches of viewable glass caused that.
At USR, we had to put an asterix on the box and tell people that 53.3 was the highest you could get because of FCC regs. In our case we were claiming connection speed and not throughput. Yeah, splitting hairs but there was a difference.
I think in the case of WiFi there is no possible way to put out a claim that would apply to everyone. Every situation would have different interference etc..
The numbers used for WiFi,
The numbers used for WiFi, like 450mbps and for LAN, 1000mbps are the maximum symbol rate for the interface, not the maximum throughput. You have to add in protocol overhead, packet loss and retransmission, and then you end up with a much lower maximum throughput. You also have to take into consideration that on WiFi the clients must check every once in awhile to allow others to talk so they don’t step on each other. Also if the router is not in WiFi-n only mode, it must broadcast periodically at WiFi-b or WiFi-g speeds to let potential clients know that it exists, which will bring down the maximum even further.
A network standard can have a maximum, even if it is unreachable in the real world, but there is no way for the standard makers to know what typical performance will end up being, and it will vary for situation to situation.
It would be nice if router makers gave maximum and typical real world performance numbers on the box, but that seems like a mixed bag.
I’ve owned the N66U for about
I’ve owned the N66U for about 6 months now and it rocks. Data throughput on the wired computers on our home network (all are Gb LAN equipped)when transferring say a 1.5Gb movie are a sustained 105mbps. The wireless computers (2 story cape, 3 wireless computers in the upstairs bedrooms) stay at constant 150-200mbps (they have different adapters). With my daughter &son-in-law currently staying with us there are usually 3 computers (2 wired- my gaming rig & the HTPC, a wireless N 150 in a bedroom), 2 iPad 2’s, 4 Android 4G phones and 2 printers all running thru the RT-N66U and it hasn’t hiccuped once.
This forum on Tim Higgin’s site has lots of info and alternative firmware–http://forums.smallnetbuilder.com/forumdisplay.php?f=37 .
Sam
Great review Chris… I have
Great review Chris… I have a Rt-n66u up & running & cannot complain one iota. Well… maybe one. How in God’s name did you ever get the stand clipped into the back wall mount holes. ?? For the love of me i tried, but am scared if you apply alot of pressure it will break. It will not stand securely with the stand. I rather have it standing so it will cool better. Any suggestions…? TIA Rick, Montreal Canada
I was able to snap the stand
I was able to snap the stand into the base. I inserted the tabs for the stand into the holes, then pressed up (towards the end with the connections) firmly on the stand at the point where the tabs were located. I held the router cradled in my hands with my thumbs resting on the stand right at the tabs. By gradually increasing pressure, the stand did snap into place without breaking. 🙂
Yup, did the same thing!
I
Yup, did the same thing!
I feared breaking the bracket by applying so much pressure, but it did snap into place eventually. Very sturdy.
Asus PCE-N15
I can confirm
Asus PCE-N15
I can confirm it’s not just that you got a bad unit. I’ve owned this adapter for months and it seems to have an issue when used in an x16 slot.
It will work, but it sometimes “craps out”.
Found this on the Asus site…
P8Z77-V PRO (FAQ)
Due to hardware limition of the Realtek chip on PCE-N15 wireless card, PCE-N15 wireless card will function in PCIe x1 slot only. Therefore, if you would like to use PCE-N15 wireless card in your system, please plug the card into PCIe 1x slot.
I just bought this router and
I just bought this router and personally like it except tha it can share locally the USB storage to a maximum 5 clients pc. Or am I missing something? please help. thanks.
I just got the RT-N66U v.b1.
I just got the RT-N66U v.b1. This is an awesome router. The print server was the easiest set up I have ever had. My previous router started to have glitches through the switch so I replaced it with this one. I have never paid anything in the $200 range for a router before….but boy am I glad I did.
The web interface is fantastic! The Quick Set up was easy to follow and intuitive.
I did some product review before purchasing and decided to stay away from any of the routers that were AC (RT-AC66U)due to the lack of standards and available devices. This router was, in my opinion, the “best of the best” and affordable. I’m a little concerned about the heat management of the device as it does seem to run very warm. It does come with a 2 year warranty, so I would hope heat management was considered. I really think some type of small internal fan would have been a good idea.
Anyway, If anyone were to ask me about which wireless router to buy, this would be my first recommendation. Thumbs up to ASUS.
hi can you test the ASUS
hi can you test the ASUS RT-N66U agents de new apple airport extreme from june 10:th???
please do I would like to know wish is the best for a big house with 4 stories and every floor is 10 by 9 meters.
Is there any way to setup a
Is there any way to setup a USB drive connected to the ASUS RT-N66U so that it appears to computers on the same network as a mounted drive with a drive letter? The idea is to use the USB drive as a network file share and also as network backup device. It doesn’t look like the ftp or the media server options do that, but it’s tantalizingly close. Any ideas you have are appreciated.
Please ignore my question
Please ignore my question about mapping a drive letter for backup purposes. Got it going now.
Which DDNS-services does the
Which DDNS-services does the router support? I can not find any picture of an opened popup “Server”, where i could see the different entries.
Is there also one offering to use a custom DDNS-service?