MOCA gigabit?

[stevech]

The brief bits I've seen here, it looks like ~450Mbps is best possible with a direct connect over probably short coax, gotta figure error correction overhead, etc. in to everything. At least one user seems to be getting high 300Mbps range with a direct connection over I assume at least a few dozen feet of coax.

I'd imagine with a good coax network, you'd manage at least high 100Mbps range (IE close to 200Mbps). A lot better than MoCA 1/1.1. On a small coax network it might push 300Mbps.

450Mbps... is that net IP layer throughput or the marketing-hype burst frame bit rate in the MoCA modems?

 

[talllogic]

The brief bits I've seen here, it looks like ~450Mbps is best possible with a direct connect over probably short coax, gotta figure error correction overhead, etc. in to everything. At least one user seems to be getting high 300Mbps range with a direct connection over I assume at least a few dozen feet of coax.

I'd imagine with a good coax network, you'd manage at least high 100Mbps range (IE close to 200Mbps). A lot better than MoCA 1/1.1. On a small coax network it might push 300Mbps.

Currently I am using a pair of Actiontec MoCA 1.1 ethernet bridges over a direct connection. It's about 50 to 60 ft of total cabling comprised of three different cables. I consistently get up to 170mbps - very near the maximum of the standard. We'll see how the new devices perform.

 

[azazel1024]

450Mbps... is that net IP layer throughput or the marketing-hype burst frame bit rate in the MoCA modems?

Net yield. The standard is something like 670Mbps and net yield according to at least one manufacturer is around 450Mbps in real use. Granted, their "real use" is probably a 1 meter RG-6, and not an actual coax install in a home, but it is something.

 

[stevech]

How do you measure your net yield?

PS: layer 2 standards like 802.3 and 802.11 always talk about burst frame bit rate, not IP layer net yield. It's not easy to measure net yield in a LAN. Special methods.

Burst frame bit rate means very little, due to overhead in duplexing, coding, layer 3 overhead, etc.
But WiFi marketeers love to talk about high megabits and fool the consumers. So too, MoCA and HomePlug.

 

[azazel1024]

The way I do it is actual transfer rates using large files. I don't have MoCA 2.0 bridges to test with, so I can't tell you what it is. The ~450Mbps real world usable is apparently what some of the MoCA 2.0 bridge manufacturers are saying in non-marketing materials, with coding rate of 670Mbps per the standard.

I call net yield what I can actually use for the protocol I happen to care about, generally payload data rate. For example, net yield for me on GbE is about 96% or 960Mbps with 9k jumbo frames, because that is my yield for SMB file transfers. For my Archer C8 to my laptop on 2.4GHz, it is about 76%, or 228Mbps when close to my router. On 5GHz it is about 57%, or 494Mbps close to my router, again for SMB file transfers.

Not burst rates, steady state averaged data rates. Robocopy. Divide file size by transfer time, there is the net yield.

 

[stevech]

makes sense. But people too often toss out Mbps without saying it's layer 3/IP versus B.S. displayed by the WiFi or other ethernet-to-medium-X (wired or wireless).
Jumbo frame sizes - kind of misleading as few of use do that.

My gen-1 MoCA, used 24/7 for years, yields just 70Mbps w/TCP. Good enough by far. I'd pay something - not a lot, for more MoCA speed. Rarely needed luxury.

 

[azazel1024]

Mine (1.1) can hit about 92Mbps Layer 7 (SMB) throughput. Eh, slow down on GbE isn't too much without jumbo frame, but my networks works perfectly with 9k jumbo enabled, so I'll take the improvement from ~940 to ~960 Mbps.

 

[NGEN]

Not likely. Not needed is the main issue. Technology is the second issue.

96Mbps - what Internet Host will give you that speed?

Comcast. I average 255 Mbps down via DOCSIS 3.0