TRENDnet TEW-633GR: 802.11n Draft 2.0 to the Rescue
by Kris Boughton on October 26, 2007 12:00 PM EST- Posted in
- Networking
Test Setup
Even though there are industry standard benchmarks such as IXIA IxChariot available for use, we decided to try something different for the consumer market equipment. Our primary testing will be comprised of a simple two-way data transfer, once from a wireless N enabled host machine to a D-Link DNS-323 network storage enclosure connected to one of the TEW-633GR's Gigabit LAN ports, and then a second transfer (again initiated by the host) but this time from the network drive to the host system in order to complete the circle. The host machine consists of an Intel Dothan-based notebook with Windows XP Pro SP2 and TRENDnet's matching TEW-621PC Wireless N PC Card installed in order to provide Draft-N wireless functionality.
Since these products are typically designed and marketed for the home or small office environment, our tests are run with appliances, cordless phones, and other devices in the 2.4GHz spectrum operating as they would in any real-world environment. While this test scenario is not perfect, nor void of any and all RF interference, it does place the product in a real-world setting where obstructions, interference, and other items can and will affect the capability of the product. You will experience better results in spacious open-wall office settings or large open-air buildings with a direct line of sight to the router which may be somewhat representative of some small business environments. However, home users are virtually guaranteed to be in a much less ideal environment.
During the course of testing the TEW-633GR router, with attached NAS drive, remains in the same test room while our host notebook location is varied in order to establish the appropriate testing distance from the router location. Benchmarking was performed by recording the time required to transfer a 100MB file from a distance of 10 feet, 25 feet, and 50 feet. Our ten foot distance path was performed with a clear line of sight between our two test points. The 25 foot distance transfer path requires the signal to span two inner walls. The longer 50 foot path, initiated from a second floor office, involves up to four walls and a floor (depending on the exact signal propagation path). We will expand our test conditions and file transfer sizes in the next article, but for now we are focused on the home setting with an average data transfer result.
Transfers were conducted unsecured and again with the use of WPA2-PSK/AES encryption to determine if encrypting the bit stream had any appreciable effect on transfer rate. All resultant rates are reported as an average of a minimum of five timed transfers in order to account for inconsistent, non-representative data points collected during the course of testing. Finally, the TEW-633GR was tested in a pure 802.11n mode and mixed 802.11g/n mode to measure the effect, if any, of supporting slower wireless G connections to the router.
Of note, we had to install this HotFix in order to properly utilize the WPA2-PSK security protocol under Windows XP SP2.
Even though there are industry standard benchmarks such as IXIA IxChariot available for use, we decided to try something different for the consumer market equipment. Our primary testing will be comprised of a simple two-way data transfer, once from a wireless N enabled host machine to a D-Link DNS-323 network storage enclosure connected to one of the TEW-633GR's Gigabit LAN ports, and then a second transfer (again initiated by the host) but this time from the network drive to the host system in order to complete the circle. The host machine consists of an Intel Dothan-based notebook with Windows XP Pro SP2 and TRENDnet's matching TEW-621PC Wireless N PC Card installed in order to provide Draft-N wireless functionality.
Since these products are typically designed and marketed for the home or small office environment, our tests are run with appliances, cordless phones, and other devices in the 2.4GHz spectrum operating as they would in any real-world environment. While this test scenario is not perfect, nor void of any and all RF interference, it does place the product in a real-world setting where obstructions, interference, and other items can and will affect the capability of the product. You will experience better results in spacious open-wall office settings or large open-air buildings with a direct line of sight to the router which may be somewhat representative of some small business environments. However, home users are virtually guaranteed to be in a much less ideal environment.
During the course of testing the TEW-633GR router, with attached NAS drive, remains in the same test room while our host notebook location is varied in order to establish the appropriate testing distance from the router location. Benchmarking was performed by recording the time required to transfer a 100MB file from a distance of 10 feet, 25 feet, and 50 feet. Our ten foot distance path was performed with a clear line of sight between our two test points. The 25 foot distance transfer path requires the signal to span two inner walls. The longer 50 foot path, initiated from a second floor office, involves up to four walls and a floor (depending on the exact signal propagation path). We will expand our test conditions and file transfer sizes in the next article, but for now we are focused on the home setting with an average data transfer result.
Transfers were conducted unsecured and again with the use of WPA2-PSK/AES encryption to determine if encrypting the bit stream had any appreciable effect on transfer rate. All resultant rates are reported as an average of a minimum of five timed transfers in order to account for inconsistent, non-representative data points collected during the course of testing. Finally, the TEW-633GR was tested in a pure 802.11n mode and mixed 802.11g/n mode to measure the effect, if any, of supporting slower wireless G connections to the router.
Of note, we had to install this HotFix in order to properly utilize the WPA2-PSK security protocol under Windows XP SP2.
Facebook
Twitter
RSS
13 Comments
View All Comments
smn198 - Monday, October 29, 2007 - link
With the backlash against HDD manufacturers quoting a GB as 1,000,000,000 bytes instead of 1,073,741,824 bytes or maybe more relevantly, broadband speeds not being up to their advertised rates, how long will it be before we see suits against WiFi equipment manufacturers (unless it has already happened and I missed it)?
Jedi2155 - Sunday, October 28, 2007 - link
Did someone from the D-link Gaming Router (DGL-4x00) design team get hired by Trendnet? The interface looks extremely similiar as well as the options.rslayer - Saturday, October 27, 2007 - link
It's great to have throughput numbers, however it would be nice to also get latency information. When using chatty protocols such as SMB/CIFS, the latency of a network has a huge impact on the actual bandwidth. This also brings up the fact that while you describe the direction the files were transferred, you didn't describe what protocol was used to make that transfer. If you are using SMB/CIFS, then you might want to try using HTTP to get real bandwidth numbers.legoman666 - Saturday, October 27, 2007 - link
While I agree that latency is important over wireless, why would you want to use HTTP to transfer large files over the network? When you want to copy 4gb from 1 computer to another computer on the network, do you fire up Apache and HTTP it over or do you use windows file sharing (Samba?). Or maybe you'd use FTP, but either way, why on earth would you use HTTP?eek2121 - Saturday, October 27, 2007 - link
SMB is not an efficient protocol. It was designed for 10 Mbps networks. SMB 2.0 is out now with vista, but you have to be transferring to other SMB 2.0 computers. Therefore copying files over LAN is not a good way to measure throughput, hence why they should use HTTP or FTP.siberus - Friday, October 26, 2007 - link
I wonder how much abuse this router can handle. I go through routers so fast. My general outlook on routers is terribly pessimistic. I'm on a 3 person network. I don't do anything too fancy just surf,game and BT the other two users just surf. I've tried multiple Brands and even bought some expensive models hoping I would "get what I paid for" but they've all been disappointing. My current linksys is already starting to bite the bullet but it's lasted almost a year which is better then the previous 4 routers. I really don't care about performance anymore reliability is way more important. My experience with wireless is even worse. Each new wireless router I get seems to be getting less and less range. I thought the Belkin N1 would be able to get at least some acceptable speed all the way up in my room but I couldn't even get a connection so I switched back to the linksys that im using now. Best range I've had so far was with D-Links Gamerlounge and after a few months that router just stopped being able to hold a connection. (the msn reconnecting sound drove me insane >.>)I've disabled wireless all together and opted to just use an Asoka plug to get stable connection in my room.notposting - Saturday, October 27, 2007 - link
The best solution is to roll your own router, I use a P3-450 with 128MB booting off CF card (CF->IDE adapter). A friend uses a headless P3 laptop. Even a Pentium 100 w/16 MB would be sufficient though, honestly.Then you just use the wireless device as an AP, which they should be able to handle--no NAT translation or firewall duties at that point. I have a crappy Motorola router--range sucks but I live in an apartment and it's stable now that it's just in AP mode.
Check out http://www.brazilfw.com.br">http://www.brazilfw.com.br -- it's basically a descendant of Coyote firewall. 2.30.1 is the last version that you can cram onto a floppy and boot.
bob4432 - Saturday, October 27, 2007 - link
i am still using a linksys wrt54g ver 2 running v4.30.1, HyperWRT 2.1b1 +tofu13c firmware. i have had this router nearly 2yrs and its longest uptime was ~450days (current uptime is 49days - i had to change my computer room set up and it was on the battery too long as i forgot about it), it is on a ups and runs 4-5 computers w/ 1 being on 24/7 w/ ftp, http w/ never a hiccup. w/ the 3rd party firmware you can up the power on this particular unit allowing for good coverage of our condo, both inside and outside w/ decent speeds (for a 'g' router).atm if this one died i would grab another one and see how it worked out, that sucks you are having such issues w/ your equipment :(
Foxy1 - Friday, October 26, 2007 - link
Anandtech has an exciting opportunity for a Proofreader to join our dynamic team.Key Responsibilities:
* Read copy for grammar, spelling, and style
* Edit copy as needed
* Ensure accuracy (spelling, grammar, punctuation, tone, and style)
* Improve readability of all copy
Requirements:
* Extremely detail-oriented
* Strong organizational, communication, and interpersonal skills
* Ability to work accurately in a fast-paced environment
* BA in English or related field of study
* 3+ years experience as Proofreader/Copy Editor
Anandtech has a comprehensive, competitive benefits package that includes Medical, Dental, and Vision insurance, 401K, Employee Stock Purchase Plan, a state-of-the-art Fitness Center, and much more! Forward resume to Anand Shimpi: anand@anandtech.com
JarredWalton - Saturday, October 27, 2007 - link
Sorry - "Karen" (that's me now, I guess) was out of town attending some meetings for a few days. I don't know that the article was any less valid, but I did make a few grammar/typo/whatever changes. I can state that personally, running one of the earlier Draft-N products, I will be very interested to hear which WiFi 802.11.n routers are best.Right now, all I can say for sure is that I would *not* recommend the NETGEAR RangeMax Next WNR834M. When it works, it's quite decent. However, I get periodic crashes, sometimes the WiFi network "disappears" (requiring a router reboot), and other oddities. I have a lot of (too many!) wireless devices (using different chipsets) and the drop-outs are irritating to say the least. Running a high-traffic BitTorrent client usually crashes the NETGEAR within a couple hours. And it doesn't have Gigabit Ethernet either, so I have to have a separate switch. At least the wired network doesn't crash.