Belkin Gigabit Powerline HD Starter Kit - my notes so far

sabregen

Fully [H]
Joined
Jun 6, 2005
Messages
19,501
This has also been posted in the original FP thread, here:http://www.hardforum.com/showthread.php?t=1429729&highlight=gigabit+powerline

Well I bit. I read the kotaku release on Monday, and was instant intrigued. The computer room is upstairs on the front end of the house, and the living room where we do all of our movie watching is on the lower level, at the back end of the house. The upstairs and the downstairs share no common walls. The house was bought new in October of 2007. I don't mind taking on jobs around the house, and am comfortable doing wiring work, but the easiest solution would have been to run external wiring. Sadly, our home owners association prevents this.

Finally sold (since wiring internal to the house structure is very likely above my skill level) on powerline networking as a possible solution (after trying both ICS on a spare laptop, hooked to the x360 downstairs, acting as a bridge, and a buffalo router with DD-WRT on it, and a high gain antenna...both of which were terrible solutions), I went ot BB today, with some RewardZone certs in hand. I went to the PowerLine networking section, and there was ONE box of these, amongst many other 85mbps and 200mbps products from other vendors. I picked these up, over the others rated slower, at the same price.

let me be clear. I have bought multiple Belkin networking products in the past, and was dissatisfied with all of them, and returned all of them. Maybe it's because I have no previous experience with powerline networking, no basis for comparison, and am just entirely fed up with the wireless bridge situation. That being said, I was able to stream (flawlessly, with less than 10seconds initial buffering) a 1080p .MKV and a 1080p .MP4 over the powerline kit, from my server upstairs to two laptops hooked into a GbE switch, simultaneously.

Status light on the unit are as follows:

Power: Blue
Powerline network: Orange
Ethernet: Blue

This indicates that although I've got good power, and Ethernet shows GbE speed, the powerline network is sync'ing at sub 200mb/sec speed.

Ping tests from the laptops showed near identical results (within margin of error):

www.google.com - 80-85ms
to upstairs pcs (various) - 1-3ms
to router upstairs - sub 2ms

network file transfer (only 1 done so far) was a 1.4gb .mkv file in 6 minutes.

I have not yet tested outlets in the same room, as of yet, or other rooms upstairs, or otherwise possibly on the same breaker, to test powerline sync speeds higher than 200mb/sec. I also have not done more than one file transfer. I plan to do more testing, but have not yet.

I have already watched Wall-E and am now watching Chronicles of Narnia: Prince Caspian, with the wife. Up until this powerline kit, XBoX Live!, and Netflix Watch it Now were always done upstairs in the computer room, not downstairs on the couch.

In the end, it may not be gigabit, ethernet would be better, and cheaper, as well. I would not be inclined to run critical NFS, NAS, or iSCSI over it, VMotion, or anything like that...but so far, I think it's worth it.

More tests to come.
 
Further tests:

All downstairs power outlets showed powerline sync below 200Mb/sec (orange LED), both other LEDs showed Blue

All upstairs power outlets, with the EXCEPTION of the computer room showed powerline sync below 200Mb/sec (orange LED), both other LEDs were Blue

In the computer room, all outlets but one sync at "GbE" speeds indicated by a Blue Powerline LED. Both other LEDs were Blue. The one outlet that showed Poweline sync LED orange, achieved the same transfer time on the same 1.4GB .mkv file as the upstairs to downstairs test done earlier. ETA was ~6minutes.

In looking at the transfer rates, and the file size, it would appear that the average transfer speed is about 40-50Mb/sec when the Powerline Sync LED is Orange.

When looking at the same file, with the PowerLine Sync LED Blue, the transfer times are ~2minutes, give or take 15 seconds (estimated by both Mac OS X and Vista Ultimate x64). This means that the transfer is about 100Mb/sec.

Ping times remain consistent, no matter what power outlet I use in the house, although I did find one outlet in my daughters room that would not get a Powerline sync LED, at all. Throughput is only close to GbE in the same room, and that is to say that it is actually 100Mb/sec, in the same room.

Is it close to advertised speeds, in my house, with two adapters? No.
Are there listed features that are stated to make the network better, which I cannot test? Yes
Is it better than a laptop using ICS, or a wireless bridged router? Yes
Is it easy to set up? You'd have to be a complete tard to fuck this up.
Is it priced right? Given the price of competing products, and the hope for firmware updates, yes. It's the same price as the lower rated products
Are my results indicative of everyone's results? No. My house is 18months old, I only have two units to test.
Is the throughput enough, on the slower links, to stream multiple HD streams? Yes, compressed .MKVs and .MP4s were tested, simultaneously, without issue.
Would I buy them again, or buy more? Yes, considering the time it takes to implement, total lack of complexity to the process, and that it is a large enough pipe for my uses (XBoX Live!, NetFlix Watch It Now, compressed HD streaming from file server)
 
Good to see these early results.

I suggest using iperf version 1.7 to simplify the throughput measurements.

http://www.noc.ucf.edu/Tools/Iperf/default.htm

E.g.

server: iperf -s
client: iperf -c server -l 64k -t 15 -i 5 -r

where server is the name or IP of the remote computer running iperf -s.

E.g., across my wireless bridge:

F:\tools\bench\iperf>iperf -c barton-vista -l 64k -t 15 -i 5 -r
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 8.00 KByte (default)
------------------------------------------------------------
------------------------------------------------------------
Client connecting to barton-vista, TCP port 5001
TCP window size: 8.00 KByte (default)
------------------------------------------------------------
[604] local 192.168.0.189 port 62483 connected with 192.168.0.148 port 5001
[ ID] Interval Transfer Bandwidth
[604] 0.0- 5.0 sec 49.9 MBytes 83.8 Mbits/sec
[604] 5.0-10.0 sec 47.7 MBytes 80.0 Mbits/sec
[604] 10.0-15.0 sec 46.2 MBytes 77.5 Mbits/sec
[604] 0.0-15.0 sec 144 MBytes 80.4 Mbits/sec
[604] local 192.168.0.189 port 5001 connected with 192.168.0.148 port 50068
[ ID] Interval Transfer Bandwidth
[604] 0.0- 5.0 sec 44.3 MBytes 74.3 Mbits/sec
[604] 5.0-10.0 sec 51.9 MBytes 87.1 Mbits/sec
[604] 0.0-15.0 sec 151 MBytes 84.5 Mbits/sec
 
iperf (same test as you ran) reports 210-640Kbits/sec on the first block of tests, and 7.88Mbits/sec - 10.2Mbits/sec on the second block. It would appear (if I am reading the test properly) that the first block is small file xfer, and the second block is larger files.

your test appears to be consistently large files. my first bock was a few hundred k for each test, in the first block. the second block only went up to 16MB
 
Last edited:
iperf (same test as you ran) reports 210-640Kbits/sec on the first block of tests, and 7.88Mbits/sec - 10.2Mbits/sec on the second block. It would appear (if I am reading the test properly) that the first block is small file xfer, and the second block is larger files.

No, the tests are for the same block size and duration -- the difference between the first set of results and the second is the direction of transfer. The first is from the 'client' to the 'server', and the second (triggered by the -r parameter) are the reverse, from the 'server' to the 'client'.

your test appears to be consistently large files. my first bock was a few hundred k for each test, in the first block. the second block only went up to 16MB

The size is not an input to this test, but is rather an output. The test specifies buffer size (-l 64k chosen as representative), duration (-t 15), and reporting interval (-i 5). The amount transferred in the given time is a result, and dividing that result by the time of the test or interval gives the transfer rate.

I'm puzzled by the discrepancy between your iperf results and actual file transfer results (assuming you used my parameters exactly). If the files were large enough that memory caching effects could not have been significant, then I guess I'd take them as representative and forget about whatever went wrong with the iperf tests in your configuration.
 
I used your parameters, exactly, so I don't know the answer to that question, either. I plan to test in the same room (devices reporting 1GbE speeds) and posting those results, as well.
 
secondary results are in, for adapters in the same room, all lights blue:

first test: 136MByes @ 76.0 Mbits/sec
second test: 150MBytes @ 83.9Mbits/sec
 
So, what kind of encryption do these things use? In theory, that little powerline network doesn't die at your electrical box. I'm sure some "leaks" out your electrical hookup. Would it be possible for your neighbor to plug in a unit into his outlet and jump on your network?

I'm sure its unlikely as I doubt the network can handle whatever noise and crap comes over the 200amp (or more) service you have coming in. However, something to consider.
 
128bit AES encryption. If your theory is true...that's be one badass way to LAN the entire neighborhood!
 
Good to know it has encryption on it. Pretty decent speeds, though, I would have expected better from a newly built house. I doubt it would perform as well in my 25 year old home.
 
I bought one of these as well. Worked pretty well once I found 2 sockets that gave me a pink connection (it was like amber and blue mixed, which isnt noted in the instructions at all). But, twice in the last week, the adapters have just stopped working requiring me to reset them, and they run extremely hot. With those 2 things, I am taking them back before my return policy runs out. I think they are very sensitive to power fluctations still as well, since my speeds would drop if my wife was using the home theater. They are definitely going to depend on your wiring and house set up.
 
Hi All,
Thanks for the thread I run a pair of the Linksys 200Mb/s adapters. I have also used the Netgear 85Mb/s as well. One thing that has always trouble me about these is the heat ouput. You can literally almost burn your fingers holding onto one these...(backside-wall) for a few seconds... ie they get really hot. I think the power usuage is only like 15W. Whats the power rating/usage of the Belkins and do they get hot as well? Can you measure the temp some people have those sweat laser or ir thermometers... just checking ; )
 
SmallNetBuilder seems to think MoCA is a much better technology then Powerline adapters, what are your thoughts?

http://www.smallnetbuilder.com/content/view/30743/51/

SmallNetBuilder.com said:
The reason for this quest is that both powerline and wireless LANs—even draft 802.11n—have proven inadequate for reliable HD video streaming and online gaming. While some people have had success with both technologies, there are plenty of others who have tried them and found them wanting.
MoCA's advantage has always been its coax-cable transmission medium. A coax cable is a shielded, controlled impedance electrical environment, which is exactly what you want for high-speed data transmission. And while it is not terminated like 10Base2 for optimal high-speed signal handling, RG6 coax is a hell of a better transmission medium than your home's power line wiring or the interference-laden 2.4 GHz radio band.



I currently use $20 (Ebay) Actiontec MI424WR routers in bridge mode to get Ethernet all over my house via MoCA. Multiple 1080p streams without a problem. And did I say how cheap they are? ;)
 
I wish that the MoCA adapters were an option for me. Sadly, they are not. All of the cable runs in my house feed directly to the satellite dish, outside. Moreover, all 4 cable runs (yes, there's only 4 in a 2500sq ft two story) have TVs connected to them, from the satellite. This would prevent me from being able to just tie them all together via a joiner, or something similar to that.
 
Back
Top