Tuesday, January 26, 2010

Motorola, Clearwire, and the #WiMAX demo at #CES

As promised, this post will provide a follow up to my earlier article on Testing Clearwire #WiMAX at #CES in Las Vegas. While in Las Vegas for the Consumer Electronics show earlier this month, I rented a Motorola USB modem from Cheetah Wireless to connect to Clearwire's WiMAX network. Along with providing me with assurance that I could connect to the internet while I was at the show, (a good call since AT&T's 3G network was just about useless), I was interested to perform a little hands-on test as I roamed around to various venues in Las Vegas.

You can check my earlier post for the details, but to summarize - I achieved a peak download (DL) speed of 11.45 MBPS and peak upload (UL) of 0.88 MBPS. Average speeds were 5.14 MBPS DL and 0.58 MBPS UL. This is well within the advertised speeds from Clearwire:

I was quite surprised then, when I discovered the Motorola WiMAX demonstration in their exhibit at the Las Vegas Convention Center. It was puzzling how they were getting more than 3 MBPS UL, when I could only get 0.48 MBPS at the same time. How could they exceed the speeds that Clearwire specified for their network?

I asked Motorola if they were actually using the publicly available Clearwire network or some sort of demo setup, and this was their response:

First off, to answer your question, yes this was a public connection to Clear's WiMAX network in Las Vegas.

Motorola's WiMAX USB devices used at CES were standard USBw 100 devices provisioned to and using Clear's public WiMAX service. These devices have the "Unlimited Mobile Plan" from Clear which has download speeds up to 6.0 Mbps (with bursts over 10 Mbps) and upload speeds up to 1.0 Mbps. According to Clear's website, actual performance may vary.

During the show, Motorola was seeing speeds anywhere from 2 to 8 Mbps on the downlink to 0.5 to 3 Mbps on the uplink. Speed differences were noticed on our devices based on the time of the day for speed tests, which speed test service being used, and the location of the actual speed test servers. For example, Motorola noticed that speed tests done with the default Las Vegas servers typically had lower overall data rates while servers located on the west coast had much better results.
I also checked with Clearwire, and their initial response was that the Motorola results must have been "theoretical". In earlier discussions at the show, they had told me that they chose to emphasize DL speeds in allocating their TDD spectrum, since it is generally what impacts consumer experiences the most. Clearwire added:

In theory you can see faster UL speeds but today, commercially, your experience with the CLEAR modem was representative of the customer experience.
Something didn't add up. The 4.25(DL) : 3.20(UL) Motorola showed just didn't match the allocation ratio that Clearwire said they were using. So, still unsatisfied with the answers I received I turned to Corey O'Hanlon, the Sales Manager at Cheetah Wireless to see if he could help.

Corey diligently checked with his techs, and they came up with what turned out to be the real answer:
Motorola had an uncapped device. One of 2 ways that occurs is either, since they make the devices they can make an uncapable device, or through an account they have, Clear uncapped the device. The downloads would be very similiar since the download speeds are uncapped, but on the Clear end the upload speeds are throttled and capped at 1mb.
The end-result is that Clearwire's network is capable of much higher upload speeds than I experienced, but even their "unlimited" plan uses software to place a limit on what a user can achieve. No harm there, as long as the experience matches what is advertised. However, the demo that Motorola put on was definitely false advertising. I still don't know if they disabled the cap on their own, or if this was done in collaboration with Clearwire.

As a footnote, it is interesting to compare these WiMAX test results to a report that came out this week on LTE tests in Sweden. According to the report from Network World:
download speed offered by Swedish operator TeliaSonera's LTE (Long-Term Evolution) network never exceeded 12M bps when tested by market research company Northstream, a far cry from the "up to 50Mbps" promised on the provider's Web site.
The article goes on to say that upload speeds of 5 MBPS were achieved. So, after all the endless hype of WiMAX vs. LTE speeds, it's good to finally get comparable real-world results. As Verizon continues on their path to rollout LTE in the U.S. this year, and Clearwire continues on their ongoing WiMAX rollout plan, the good news is that many more of us will be able to experience peaks of 10 MBPS download with mobility soon. As for the upload speeds... 1-5 MBPS depending on your software hacking skills.


No comments: