HD Streaming Smackdown: Draft 11n vs. Powerline

[thiggins]

You might think that the winner between these two "alternative" networking technologies is obvious. But think again. [article link]

 

[VikingCrown]

For your HD Streaming Smackdown, did you test any MoCA products? Such as the NIM100 that was used by Verizon in their FIOS installs (until they incorporated it into the router)

they can be found on eBay regularly and they create a mesh network over your coaxial cable in your house. I am considering picking up a pair of them for my SageTV HD Extender I have

 

[thiggins]

I am waiting for the upcoming round of MoCA products from D-Link and Netgear that can be purchased at retail.

I think coax-based networking will be better than powerline, given the superior electrical characteristics of coax. But I won't know until I test them.

I did test a Corinex Coax-based product (not MoCA compliant) a few years back and got about 80 Mbps max.

 

[wmurwin]

Similiar Experience with Powerline

Excellent Article!

I too recently experiment with Powerline for Stream HD, and experience similar results. I too found my self trying to decide between 802.11n Kit and a Powerline Kit, I decided to try the Powerline Kit because it was $50 cheaper then the 802.11n kit...

I also found Powerline was affected by the noise generated by TVs, Refrigerators, etc. I ended up installing a Power Strip with AC Filter for the TV, AV Receiver, and Audio Amplifiers; which greatly improved the through put...but only enough to sustain DVD quality stream video.

When attempting to stream HD video (1080P @ 15Mbps) the Audio & Video could not keep in sync & there was significant delay so that the video was unwatchable. Probably with more AC Filters installed on (Refrigerators, etc) I could improve throughput, but I could spec what would be cost and how many I would need.

The other thing I learned was the fact that might need a Phase Coupler to bridge between the different phase of 110 Volts runs in my house.

For all these reason above I returned Powerline devices.

I am encouraged to see your results with 802.11n.

For myself, I believe its time to bite the bullet and bring in an electrician to pull CAT6.

Thanks.

 

[thiggins]

The other thing I learned was the fact that might need a Phase Coupler to bridge between the different phase of 110 Volts runs in my house.

Phase coupling should not be required for any current-generation powerline networking devices. They can pick up the required signal from line to line coupling.

For myself, I believe its time to bite the bullet and bring in an electrician to pull CAT6.

It's really the only way that you are going to get trouble-free HD streaming.

 

[VikingCrown]

I am waiting for the upcoming round of MoCA products from D-Link and Netgear that can be purchased at retail.

I think coax-based networking will be better than powerline, given the superior electrical characteristics of coax. But I won't know until I test them.

I did test a Corinex Coax-based product (not MoCA compliant) a few years back and got about 80 Mbps max.

I can't wait for them to come out! looks like the MoCa 1.1 spec ups the throughput to 270 with an actual output of 175. I'll be very interested to see how close their numbers are.

Any idea when the Netgear or D-Link products are coming out? Both pres releases say 2008, so they should be out soon I would think.

 

[thiggins]

D-Link has actually announced product (DXN-221 kit). Netgear has only "previewed" something at January's CES.

I am pinging D-Link on availability.

BTW, Actiontec has supposedly had a MoCA adapter for over a year. But they won't sent it for testing.

I would take any throughput numbers with a truckload of salt. Remember that both draft 11n and Homeplug talk about 100's of Mbps of throughput, but deliver 60-80 best case in most cases.

 

[VikingCrown]

I did see that DXN-221 kit, I hope that D-Link gets back to you so you can give it some real world testing so we can see how close they come to their proposed throughput numbers.

I agree to take them with a grain of salt as far as numbers are concerned, but if the previous throughput of 80-90 Mbps that i've seen from the NIM100's and the 1.0 spec (on various forums) are true and not too far off their claimed throughput of 110 Mbps, then the 1.1 spec that claims 175 Mbps may not be too far off the mark.

Please let us know if you get in touch with D-Link and get your hands on a pair of their devices! I can't wait!

~ Brian

 

[thiggins]

Are the throughput numbers you quoted TCP or UDP?

 

[VikingCrown]

If you mean the 1.0 Spec, I can just point you to the posts i got it from on the SageTv forum, they don't specify. Here is what i got from the MoCA FAQ page about the 1.1 spec:
"MoCA 1.1 extension has been ratified and offers additional benefits. These include parameterized quality of service (PQoS) for provisioning and bandwidth management of real time data requests for video applications.
MoCA 1.1 with packet aggregation also increases the net throughput performance from 110 Mpbs to 175 Mpbs.
MoCA is the only organization that actively promotes its net throughput metrics. While other may claim higher performance, these are generally theoretical bit rates. This is the difference between what is promised and what is received.
This new extension also increases the network from the current eight nodes to 16 nodes. "

 

[corndog]

I agree to take them with a grain of salt as far as numbers are concerned, but if the previous throughput of 80-90 Mbps that i've seen from the NIM100's and the 1.0 spec (on various forums) are true and not too far off their claimed throughput of 110 Mbps, then the 1.1 spec that claims 175 Mbps may not be too far off the mark.

~ Brian

I recently got a set of 4 NIM100's from ebay, and have to say that they really work well. If I do an IPERF test from an ethernet connected host on one side, through the NIM100, over CoAX, through another NIM100, and to a destination ethernet connected host on the other side, I get sustained 96megabit throughput. This seems to be pushing the limit of what is possible because the ethernet port on the NIM100 is only 100megabit. If they come out with a 1.1 it will need gigabit ports.

I have been using ReadyLinks CEB-402 HPNA Coax bridges for almost 2 years (not sure why people around here keep saying you can't purchase CoAX bridges - I simply bought these off their website ages ago) but they need to be rebooted every couple weeks because they lock up. My ReadyLinks give me about 60megabit on a similar test.

Based on this, I'd say that MoCA is definitely a more promising technology then HPNA over Coax. Their claims of being more honest about actual real world performance numbers are good, from where I'm sitting.

 

[thiggins]

Thanks for the performance data, corndog. Was that TCP or UDP?
Also how many splitters between the two end units and were there other signals active on the coax?

 

[Unregistered]

Thanks for the performance data, corndog. Was that TCP or UDP?
Also how many splitters between the two end units and were there other signals active on the coax?

That was TCP. The only splitter was the single one at one end that combines all of the runs together. This setup was isolated from my TV cable source during this test.

 

[sxr71]

I use 2 WNHDE111s to watch 720p video off my home server but it is something like 8mbps and so it works fine. However the real test will be to see if it can handle Blu-Ray bitrates. I've actually seen it get pretty high and enough to max out a 100mbps port so I guess with buffering there is a chance it may work, but I suspect the slightest things like someone walking around the apartment could upset the balance and cause stuttering.

I guess there's nothing like running ethernet cables under carpet.

 

[thiggins]

I guess there's nothing like running ethernet cables under carpet.

For Blu-ray streaming, you got that right.

 

[saturation]

For Blu-ray streaming, you got that right.

I use 2 WNHDE111s to watch 720p video off my home server but it is something like 8mbps and so it works fine. However the real test will be to see if it can handle Blu-Ray bitrates. I've actually seen it get pretty high and enough to max out a 100mbps port so I guess with buffering there is a chance it may work, but I suspect the slightest things like someone walking around the apartment could upset the balance and cause stuttering.

I guess there's nothing like running ethernet cables under carpet.

A number of tech centered folks have been streaming 1080p since 2008. Frame rates are usually <= 30 fps.


http://en.wikipedia.org/wiki/1080p#Internet_content

I've been using it problem free for over a month, including while still using 802.11g.

Bluray per se is problematic not because of the bit rate required, but because of DRM. How individuals achieve 1080p streaming are discussed on video centered websites, such as AVforum.

Regardless, unprotected 1080p30 video can be streamed by a sustained >= 40 Mbps through out the play period, not just a 1 minute measurement as seen in SNB reviews. A large player buffer helps tremendously. Often its much less, and my Resource Monitor typically reports a sustained pull of 5-10 Mbps while watching a movie, far under what I have the network designed for.

I find a throughput measured for at 15 minutes or longer, is a better indicator of how the network performs for 2-3 sustained hours. I just send a 1-10GB file down the network and insure it passes down consistently.

Further many folks fail to critically consider the distance between A and B units, and as SNB points out plainly, distance and obstruction severely affects the actual throughput from A to B.

Although a detailed technical discussion is possible the easiest thing to do is do it, and see what needs to be done to insure Mbps speed is consistent. At this time, streaming 1080p is not a plug and play event and because of the varied nature of wireless environments, have to be tailored to your need.

http://www.networkedmediatank.com/showthread.php?tid=24655

 

[thiggins]

If you are streaming 1080p uncompressed video over 11g, your player probably does a good amount of buffering.

There is no 11n solution I've seen yet that provides throughput consistent enough over even the 1 minute period used in my testing, let alone over multiple hours. For successful HD streaming, the player has to have a large buffer and the smarts to keep it full, even with highly variable network throughput.

 

[saturation]

I agree fully with you. A key issue is aesthetic, and 1080p will play under 10 hiccups / 2 hour period. At 11g it was doable with more hiccups [but cleanest at 720p], which prompted me to switch to 11n, given the bandwidth increase.

So, rather than discuss the technology, proof is in getting it to work.

Anyone interested can try using these tools. These folks make a great highly technical grade player that can be tuned for network inefficiencies and has multiple compiles for different OS:

http://www.mplayerhq.hu/design7/news.html

Some test material:

http://www.microsoft.com/windows/windowsmedia/musicandvideo/hdvideo/contentshowcase.aspx

I'd be interested to read experiences of anyone trying these tools.

If you are streaming 1080p uncompressed video over 11g, your player probably does a good amount of buffering.

There is no 11n solution I've seen yet that provides throughput consistent enough over even the 1 minute period used in my testing, let alone over multiple hours. For successful HD streaming, the player has to have a large buffer and the smarts to keep it full, even with highly variable network throughput.

 

[mjw]

Perhaps the tables will be turned with the Netgear XAVB5001

I'm looking forward to the next generation HomePlug AV500 device from Netgear. If the current generation of products is nearly providing acceptable streaming performance as indicated in Tim's review then perhaps the next generation will finally provide a realistic alternative to ethernet.

 

[thiggins]

I'm looking forward to the next generation HomePlug AV500 device from Netgear. If the current generation of products is nearly providing acceptable streaming performance as indicated in Tim's review then perhaps the next generation will finally provide a realistic alternative to ethernet.

We'll just need to wait and see. Belkin's "Gigabit" powerline adapter based on Gigle's chipset was disappointing. We'll have to see what Atheros' chipset can do.