WNDR3700 vs. linksys wrt610/e3000- varied review results?

[teknojnky]

this is not a response to anything in particular, maybe its just me, but I find it humorous some of the expectations of usb drives attached to routers, when dedicated nas devices have a hard time getting it right.

 

[j_nolesfan]

this is not a response to anything in particular, maybe its just me, but I find it humorous some of the expectations of usb drives attached to routers, when dedicated nas devices have a hard time getting it right.

That's a fair poke. I'd probably find it funny if you wanted to use 4X100mm wheels for inline skating competitions when 4X110mm is the fastest and most widely accepted configuration. It's probably not funny to you if you're not a serious inline skating competitor for the same reason that your networking comment isn't particularly funny to me because I'm not highly experienced in networking. Therefore, I first take the advertised features at face value, that the product does what it says. That's funny to you just like 100mm wheels are to me.

Jeff

 

[teknojnky]

but see, the products don't advertise your going to get NAS or even direct attached USB peformance, but it seems this is what alot expect.

Heck, usb drive attached directly to a computer is not all that great of a performance.

Attaching that same usb drive to a low power, non dedicated device, going through a network, is not going to magically make it perform better, and in reality adds several layers of performance draining to the chain.

Besides, since when is usb drives and media streaming become a core function of a router? /boggle

I'd rather the manufacturers remove the function and dedicate the silicon to improving the real functions of a router/gateway... routing and WLAN performance.

 

[bbear]

That's a fair poke. I'd probably find it funny if you wanted to use 4X100mm wheels for inline skating competitions when 4X110mm is the fastest and most widely accepted configuration. It's probably not funny to you if you're not a serious inline skating competitor for the same reason that your networking comment isn't particularly funny to me because I'm not highly experienced in networking. Therefore, I first take the advertised features at face value, that the product does what it says. That's funny to you just like 100mm wheels are to me.

Jeff

Very well said. It speaks to the old phrase about how ignorance is bliss. And that's exactly what some of these marketeers count on. They put a 2-bit feature that looks really cool on paper and stop short of saying it's a NAS; all the while, they've built the perception. So while it legitimately might provide some basic LAN storage functionality, it would be prudent for the customer to really educate themselves on exactly what USB storage hung off a router really is so that they can determine the value of it for themselves. Hence why forums like this are so valuable, so that we all can stay a step ahead of the marketeers, pros and novices alike.

 

[jemz0r]

I would take most reviews of Wireless products on the Internet as "reference" only.

Because most of them don't understand how wireless works.
I've seen the following "tests"

See ping time in Counter Strike over wireless
Transfer files over wireless
Download files from Internet over wireless
Download BitTorrent from the Internet over wireless
Benchmark using Speedguide.net


Those are all tests that do not give an unbiased, relevant, or accurate indication of actual performance.

Not to mention that wireless also depends on the environment the devices are tested. There are some variables these testers may encounter.

building materials - concrete walls, wood walls, glass, tin roof etc.
If they have triple concrete walls, then of course they would complain about poor range.

orientation/location they place the devices, can you be certain they place the devices exactly the same spot? maybe one device is on the floor, the other is higher up on a shelf. That makes a difference.

Other wireless signals in the area. For example, the Tomshardware review, the tester mentioned seeing over 10 wireless networks in his neighbourhood. Most likely there is no clean wireless channel as there is technically only 3 that is clean. So there will definitely be signal collisions happening. If a neighbour using an overlapping channel also started transfer files while their benchmark was happening, then bad luck to that product's vendor.

Test using Windows File transfer is technically not a fair test. Can you be certain their windows is totally clean and not affected by driver conflicts etc?
We've all seen Windows File transfer going slow even over wired connection.

Test that connect to the Internet is also not a fair test. Can you be certain their ISP, or any point between them and the server is not suffering a slowdown at that point in time?

There are plenty more examples of unfair testing of wireless products.
But that being said, I must commend SNB for being the most informed and is as unbiased as possible in a real world environment. Good work Tim.

 

[thiggins]

But that being said, I must commend SNB for being the most informed and is as unbiased as possible in a real world environment. Good work Tim.

Thanks for the kind words.

Doing wireless testing in a consistent way is difficult, due to the nature of the technology. But, as you point out, there are some things that reviewers can and should control.

Every test environment is different. So it's impossible to compare results from reviewer to reviewer. But with the large database of products we have, you can do product-to-product comparisons due to our consistent test methods.

 

[jemz0r]

I've seen lots of wireless equipment being tested due to the nature of my work.

There are many test methods, some practical, some not so practical.

For example, to test the actual electronics they test the devices inside a shielding chamber so all external interference is blocked. They also don't use antennas, they connect the wireless devices together using RF cables through an attenuation device.

Using this method, the signal goes from one device straight to another inside a shielding chamber. You'll get crazy high throughput figures using this method. It's testing the maximum performance in the perfect environment.


A more practical test would of course be over the air. But you'll need to find a completely clean location, such as an underground car park where there are no wireless signals at all. Then you can put the devices exactly 10 metres or 30 metres etc apart. Also try to use the exact same spots as the signal is also affected by surrounding walls, remember MIMO uses reflections to improve signal reception.

Of course, reviewers will find it hard to get the environments mentioned above, however they should still make the effort to understand and control their environment.
They can start by not testing windows file transfer method. The industry standard is using ixChariot, which is what SNB uses.

 

[tipstir]

I've seen lots of wireless equipment being tested due to the nature of my work.

There are many test methods, some practical, some not so practical.

For example, to test the actual electronics they test the devices inside a shielding chamber so all external interference is blocked. They also don't use antennas, they connect the wireless devices together using RF cables through an attenuation device.

Using this method, the signal goes from one device straight to another inside a shielding chamber. You'll get crazy high throughput figures using this method. It's testing the maximum performance in the perfect environment.


A more practical test would of course be over the air. But you'll need to find a completely clean location, such as an underground car park where there are no wireless signals at all. Then you can put the devices exactly 10 metres or 30 metres etc apart. Also try to use the exact same spots as the signal is also affected by surrounding walls, remember MIMO uses reflections to improve signal reception.

Testing should not be done using file transfer method. The industry standard is using ixChariot, which is what SNB uses.

Nothing wrong using the file transfer methodology for testing. You need to know how well it performs LAN to LAN. I've worked at many clients sites and 10 gig data over the wired and 802.11x protocols were tested to see how well transfers rates were. Main client wants their end users to be able to access data without delay as that's the scope here.

Most home users are going to care more about the wireless signal and secondary the data transfer rate. I am doing both. SNB test helps the user decide how well the router or wireless router performs wireless and wired WAN to LAN and LAN to WAN.

 

[jemz0r]

Nothing wrong using the file transfer methodology for testing. You need to know how well it performs LAN to LAN. I've worked at many clients sites and 10 gig data over the wired and 802.11x protocols were tested to see how well transfers rates were. Main client wants their end users to be able to access data without delay as that's the scope here.

Most home users are going to care more about the wireless signal and secondary the data transfer rate. I am doing both. SNB test helps the user decide how well the router or wireless router performs wireless and wired WAN to LAN and LAN to WAN.

ixChariot is a more accurate indication of throughput. I'm sure Tim can explain why as he uses it for SNB reviews.

 

[tipstir]

ixChariot is a more accurate indication of throughput. I'm sure Tim can explain why as he uses it for SNB reviews.

Yes I am aware of that. SNB main site explains what testing methods are used. I bring my own tools with me when I go to clients sites to test their wireless and data transfers. Might be not by client policies but like I tell them whatever it takes to get the job done. Never had issues with them with that statement.

 

[thiggins]

The main reason I use IxChariot is the high resolution it provides for looking at throughput variation. Closest inexpensive tool I have found is iperf /iperf. But they have 1 second minimum sampling period.

Protocol overhead for IxChariot is also minimal.

But IxChariot cost is out of reach for most people. I have access to it only through a special program for media reviewers.

File transfer or any other throughput measurement method can be fine depending on what you are doing with the data.

 

[I am Penguin]

so what is that review site? I'm just suprised that toms hardware gave the wndr3700 such a poor review when compared to the wrt610
(http://www.tomshardware.com/reviews/802.11n-wireless-router-access-point,2605.html)

I do trust smallnetbuilder's reviews more- you guys go very in depth in your review and really try to be as scientific as possible- but I also like toms hardware reviews in general so I was just suprised at the discrepancy

Looking at comments like these below i can see Tom's have not really put much effort in it. Sorry to say but you cannot compare that review with the SMB review. More consideration should have been put into the effort. I am disgusted in how the "twisted" review of Tom's can sway people to buy the wrong hardware. Luckilly I had my trust in Netgear to make and "informed" decision based on what I read and thereafter came across this site to confirm my choice. Yeah there are issues but as always Netgear will fix them.

Because of the restrictions in our test setup, we had to pair both of these Netgear routers with Netgear’s own WNDA3100 USB adapter. To make a long story short, this was disastrous, we went through three of these adapters before finding one that yielded even moderately decent results. As soon as we switched to our notebook’s internal 11n adapter, performance bounded forward. Unfortunately, we have to stick with our same-brand results. The WNDA3700 is a good router. Just be sure to use it with someone else’s client adapter.

we used a little tool called Bandwidth Monitor to grab a few seconds of activity from several file transfer tests. Check out these two comparisons, one showing a transfer from the router (red bars) and the other to the router (green bars).

Could Netgear give Linksys a run for its money? With a different client adapter, we suspect so. Some of our data sure hints at it

 

[stevech]

I see too often now, PC magazines (print/on-line) cannot afford true product testers for most reviews. So some joe-blow free-lance sends in a report on a product.

No vetting or peer-review at all.

Most of these mail-ins are superficial, usually wrong if I have direct experience, and are a sad comment on tech journalism today.